Methods and systems for generating availability management framework (AMF) configurations

[0001] This application is a continuation-in-part of U.S. patent application Ser. No. 11/961,849, filed on Dec. 20, 2007, and entitled “Methods and Systems for Generating Availability Management Framework (AMF) Configurations”, the disclosure of which is expressly incorporated here by reference. This application is also related to, and claims priority from, U.S. Provisional Patent Application Ser. No. 61/085,583, filed on Aug. 1, 2008, entitled “Automatic Generation of AMF Compliant Configuration Top-Down Approach”, the disclosure of which is also expressly incorporated here by reference.

[0002] The present invention generally relates to high availability (HA) systems (hardware and software) and, more particularly, to HA systems managed by an Availability Management Framework (AMF).

[0003] High availability systems (also known as HA systems) are systems that are implemented primarily for the purpose of improving the availability of services which the systems provide. Availability can be expressed as a percentage of time during which a system or service is “up”. For example, a system designed for 99.999% availability (so called “five nines” availability) refers to a system or service which has a downtime of only about 0.44 minutes/month or 5.26 minutes/year.

[0004] High availability systems provide for a designed level of availability by employing redundant nodes, which are used to provide service when system components fail. For example, if a server running a particular application crashes, an HA system will detect the crash and restart the application on another, redundant node. Various redundancy models can be used in HA systems. For example, an N+1 redundancy model provides a single extra node (associated with a number of primary nodes) that is brought online to take over the role of a node which has failed. However, in situations where a single HA system is managing many services, a single dedicated node for handling failures may not provide sufficient redundancy. In such situations, an N+M redundancy model, for example, can be used wherein more than one (M) standby nodes are included and available.

[0005] As HA systems become more commonplace for the support of important services such as file sharing, internet customer portals, databases and the like, it has become desirable to provide standardized models and methodologies for the design of such systems. For example, the Service Availability Forum (SAF) has standardized application programming interfaces (APIs) to aid in the development of portable, highly available applications. As shown in the conceptual architecture stack of FIG. 1, the AIS 10 is intended to provide a standardized interface for the HA middleware 16 to support HA applications 14. As described below, each set of AIS functionality is associated with an operating system 20 and a hardware platform 22. The reader interested in more information relating to the AIS standard specification is referred to Application Interface Specifications (AIS), Release 5, which is available at www.saforum.org, the disclosure of which is incorporated here by reference.

[0006] Of particular interest for the present application is the Availability Management Framework (AMF), which is a software entity defined within the AIS specification. According to the AIS specification, the AMF is a standardized mechanism for providing service availability by coordinating redundant resources within a cluster to deliver a system with no single point of failure. The AMF provides a set of application program interface (API) functions which are used, among other things, to set the states of components within a cluster and to determine the health of those components. The components are also provided with the capability to query the AMF for information about their state. An application which is developed using the AMF APIs and following the AMF system model leaves the burden of managing the availability of its services to the AMF. Thus, such an application does not need to deal with dynamic reconfiguration issues related to component failures, maintenance, etc.

[0007] The AIS standard also defines the information model (IM) for applications to be controlled by AMF. According to this IM, a service unit (SU) as a logical entity that aggregates a set of components, thereby combining their individual functionalities to provide a higher level service. A service unit can contain any number of components, but a particular component can be configured in only one service unit. Since each component is always enclosed in a service unit, from the AMF's perspective, the service unit can be considered the incremental unit of redundancy in the sense that it is the smallest logical entity that can be organized into groups in a redundant manner to provide and protect the service instances.

[0008] A service group (SG) is defined within the AMF information model as a logical entity that groups one or more SUs in order to provide service availability for a particular set of service instances according to a particular redundancy model. The AMF information model also defines an application as being a logical entity that contains one or more SGs. An application combines the individual functionalities of the constituent service groups to provide a higher level of service. From a software administration point of view, this grouping into an application reflects the set of SUs and contained components that are delivered as a consistent set of software packages to the AMF environment. Within this environment, by introducing the software management framework (SMF), the AMF information model was extended with types that allow the characterization of software developed for SAF systems, e.g., software which is used to provide HA services managed by the AMF.

[0009] This software can typically be deployed in different configurations depending on, for example, the characteristics of the system it is deployed on, the number of nodes and their characteristics, the capabilities of the AMF implementation and the requirements put on the provided services. Thus, a software vendor can describe their software such that it would allow for the widest possible usage by describing various AMF related characteristics of the software. Based on this a system integrator has to produce a particular configuration for the system the software is deployed on. However, none of the current SAF specifications define how to generate the information model, i.e., the configuration to be used by an AMF implementation in a particular system for a particular application. Therefore, one option for creating an information model would be to create it manually. Considering the potential for an AMF configuration file to have hundreds if not thousands of lines of code to be properly defined, creating a configuration file in a manual manner will be time consuming and prone to errors.

[0010] Accordingly, it would be desirable to provide systems and methods automating the generation of configurations for an AMF system or process.

[0011] Systems and methods according to the present invention address this need and others by automating the generation of configurations for an AMF system or process.

[0012] According to an exemplary embodiment, a method for generating a configuration for an Availability Management Framework (AMF) includes the steps of selecting entity types from an entity types file which can provide a set of service instances by first evaluating entity types associated with an application type in the entity types file, generating entities associated with the selected entity types; and distributing at least some of the generated entities on nodes of a cluster.

[0013] According to another exemplary embodiment, a system includes a processor for generating a configuration for an Availability Management Framework (AMF), including performing the functions of: selecting entity types from an entity types file which can provide a set of service instances by first evaluating entity types associated with an application type in the entity types file, generating entities associated with the selected entity types; and distributing at least some of the generated entities on nodes of a cluster.

[0014] According to another exemplary embodiment a computer-readable medium contains instructions stored thereon which, when executed by a computer or a processor, perform the steps of: selecting entity types from an entity types file which can provide a set of service instances by first evaluating entity types associated with an application type in the entity types file, generating entities associated with the selected entity types; and distributing at least some of the generated entities on nodes of a cluster.

[0033] The following description of the exemplary embodiments of the present invention refers to the accompanying drawings. The same reference numbers in different drawings identify the same or similar elements. The following detailed description does not limit the invention instead, the scope of the invention is defined by the appended claims.

[0034] To provide some additional context for this discussion, consider an exemplary Availability Management Framework (AMF) controlled High Availability (HA) system as shown in FIG. 2. Therein, four nodes (A, B, C and D) are associated with two service groups (SG1 and SG2). A service group is a group of service units (SUs) which provide service availability for one or more service instances (SIs) of particular services. For example, SG1 includes SU1 and SU2 which, in this example, support an instance of an e-mail service (hardware and software) and SG2 includes SU3, SU4 and SU5, which support two instances of a fax service (hardware and software). At runtime, for the email service instance SIA supported by SG1, AMF 300 may select to assign SU1 the active state and SU2 the standby state.

[0035] Each of the exemplary service units in SG1 has two components associated therewith. A component is the smallest logical entity on which the AMF 300 performs error detection and isolation, recovery, and repair. Thus, a component typically includes all of the functions which cannot be clearly separated for error containment or isolation purposes. The component is also the logical entity that a portion of a workload is assigned within the provisioning of a service instance, e.g. the email service. Such a portion of the workload is called a component service instance (CSI). The email service instance SI A is composed of two component service instances CSI A1 and CSI A2. Each of these can be assigned to a component within a service unit. Components can further be grouped into protection groups which reflect the redundancy associated with provisioning of the component service instances. For example, components C1 and C3 can form a first protection group for CSI A1 and components C2 and C4 can form a second protection group for CSI A2, both of which are associated with the email service instance SI A. Thus if component C1 fails while being assigned active for CSI A1 and C3 assigned as standby, the AMF 300 could fail over CSI A1 to component C3 the active state and, similarly, if component C2 fails while being active for CSI A2, then the AMF 300 could fail over CSI A2 to component C4 the active state. At the time of the fail over of CSI A1, AMF 300 will switch over or will fail over CSI A2 as well so the entire service unit SU2 becomes assigned active for the entire service instance SI A.

[0036] Service group SG2 illustrates a slightly different configuration wherein two instances of a fax service are supported by three service units SU3, SU4 and SU5. For example, SU3 and SU4 could each be assigned by AMF 300 the active state such that each supports one instance of the fax service, while SU5 could be assigned the standby state and operate as their redundant backup. In this case, components C5 and C7 would form one protection group associated with one of the two fax service instances SI B and within that protect component service instance CSI B1. Components C6 and C7 could form a second protection group to protect CSI C1 associated with the other SI C of the two fax service instances.

[0037] In the above described examples, service groups, service units and components can all considered to be logical entities associated with physical entities that perform work. These logical entities are assigned workloads in order to render the service, e.g., email or fax service. A component service instance (CSI) represents the workload that the AMF 300 can dynamically assign to a single component. Component service instances are grouped into a logical entity called a service instance (SI). A service instance aggregates all component service instances to be assigned to the individual components of the service unit in order for the service unit to provide that particular service instance, e.g., email or fax service. A service instance represents a single workload assigned to the entire service unit. As shown in FIG. 2, service instances are shown as well as their relationship with the components, service units and service groups on the four nodes. More specifically, for SIA 302, SU1 is assigned the active HA state and SU2 is assigned the standby HA state (as shown by the dashed connecting line). For SIA 302, two CSIs (A1 and A2) are assigned to components C1 and C3, and to C2 and C4, respectively. Regarding SIB 304, SU3 is assigned the active HA state and SU5 is assigned the standby state (as shown by the dashed connecting line). Similar comments apply towards SIC 306.

[0038] For an AMF software entity 300 to operate in the above described exemplary manner, the AMF software entity 300 needs to receive from a configuration repository 310 an input, file or system model which describes what is to be managed. In a Service Availability Forum (SAF) cluster the Information Model Management Service (IMM) typically serves as the configuration repository 310. According to one exemplary embodiment, such configuration information can take the form of an XML (eXternal Markup Language) file, however it will be appreciated that the particular form or format of the configuration input is not critical. According to exemplary embodiments, a high level method for generating the configuration input an AMF application will now be described with respect to FIG. 3(a). Initially an AMF Configuration Generator 402 receives inputs of entity types 404, e.g., as entity types files (ETFs) and services set/user input 406 to perform a mapping function which creates a system model for future use. This system model is output by the AMF Configuration Generator 402 in the form, for example, an XML file according to the SAF Information Model Management (IMM) Service XML schema 412. This output code in an IMM XML file 412 is then stored in the configuration repository 310 in step 414. These various inputs and steps will be described in more detail below.

[0039] The first exemplary input is entity types 404 which describe, for example, various AMF related characteristics associated with application software to be controlled by the AMF system. These entity types 404 are AMF entity types described in XML entity types files, as described by the SAF Software Management Framework (SMF) specification SAI-AIS-SMF-A.01.01 which is available at www.saforum.org, the disclosure of which is incorporated here by reference. From the SAF specifications, there are currently two mandatory AMF entity types that need to accompany any application software intended for such systems. These two entity types are the component types (CompType) and their respective component service types (CST). Other entity types may also be provided as inputs 404, however, they do not have to be present in an entity types file if the software implementation has no restrictions related to them.

[0040] The second exemplary type of input is the services set/user input 406. The services set/user input 406 typically includes information related to the services the site provides for which the AMF configuration is being created. In the AMF environment the services set/user input 406 is a set of service instances in which each service instance is described by a set of component service instances as well as other desired user inputs, e.g., a number of nodes provided with the system providing the service or services and the desired redundancy model to protect them. The services set/user input 406 is typically site specific. This exemplary method involves using SI templates each of which describe a set of SIs that shall be protected by SGs of the same type. Within each SI template the CSIs are also defined via CSI templates. The SI templates are typically based on the service types (SvcType) provided as service set/user input 406 or by the entity types files. The service type determines the composing CS types (CST) and any limits for the number of component service instances that an SI of the service type may contain. Accordingly, the CSI templates are based on the CS types provided with the software and the user input needs to indicate which CS types are used from these and how many instances of them are desired for the composition of an SI of the particular SI template. An SI template also defines the number of SIs to be configured based on this template and their grouping to SGs. Additionally, the desired redundancy model is typically input by a user and included in the SI template. Another input typically received by the AMF Configuration Generator 402 from the services set/user input 406 is the number of nodes. The number of nodes input is used to assign the service units of the configuration to nodes according to different criteria (e.g. implement hardware redundancy, load balancing among cluster nodes, etc.) if necessary. This information may be provided as a user input to the AMF Configuration Generator 402, or collected from the target system by other SAF services.

[0041] These exemplary input entity types 404 and services set/user input 406 are received at the AMF Configuration Generator 402 where the inputs are processed to generate a system model for a specific AMF configuration. A general discussion of the configuration generation will now be provided followed by a detailed (but purely exemplary) pseudo-code implementation. References to the pseudo code implementation are priced in the text below in parentheses.

[0042] Initially the exemplary configuration generation method (generateConf method) begins by matching each received SI template with the entity types 404 provided. A suitable SU type is selected from among those in the received input set (findSUType method). This selection method checks the SU types for the capability of providing the desired service type that the SI template refers to. If there are such SU types available, then one is selected by matching the capabilities of the component types of the SU type for each of the component service types with the capabilities required for providing the specified component service type by the CSI templates (findCompType method) in the required redundancy mode required by the enclosing SI template. According to one exemplary embodiment, the selection method finds the SU type(s) that provide the CSI(s) with the minimum number of components, i.e., the selection criterion results in maximizing the utilization of each component in the configuration. However, other selection criterion can be used as desired.

[0043] For the case where the selection criterion results in no SU types found that provide the desired composition of CSIs, this exemplary configuration generator 402 can determine whether the available component types can be grouped in another manner to construct a new, desired SU type. This can occur as part of the findSUType and findCompType pseudo-code functions described below, i.e., the findCompType method adds an appropriate component type from the complete set of component types to the newly constructed SU type for each set of CSIs that needs to be protected. This determination as to whether new SU types may be constructed is typically based upon a user input, e.g., a yes or no input regarding whether the AMF Configuration Generator 402 is permitted to perform new SU type construction. If a no is submitted then only the submitted types will be used. If a yes is submitted, then the submitted SU types will preferably be used, however, if the submitted SU types are found to be insufficient, the AMF Configuration Generator 402 will have the option to attempt to create an appropriate SU type. Once an SU type is found or constructed, this exemplary method chooses an appropriate SG type from the available set of SG types based on the features defined by the user for the desired redundancy model in the SI template. If no such appropriate SG type is available, a new SG type is created using input parameters found within the SI template.

[0044] Upon completion of finding the appropriate entity types, the AMF Configuration Generator 402 generates the entities, i.e., instances of these types (generateConf method) that can provide the desired services (createSis method). The configuration is populated so that both the active and the standby assignments of each SI protected by each SG can be satisfied. This is done by creating one SU with the required number of components so that the SU provides the required active and standby capabilities for all of the SIs as necessary and within them for all CSIs. Once all of the components of the first SU have been added, other SUs required by the redundancy model are copied from this first SU as desired for the HA environment to create a first SG. This procedure is then repeated for each SG necessary to protect all the service instances generated from an SI template in the configuration that is currently being created. Additionally, one exemplary method for creating this configuration uses equal ranking and equal load between all SUs. Alternatively, other exemplary methods can be used.

[0045] For example according to one exemplary embodiment, instead of using equal ranking and load, different ranks can be assigned to the different SUs within an SG. A global rank can be assigned as well as a per SI rank. If only the global rank is specified then for SIs this is the rank that would be used. According to another exemplary embodiment, other factors can be used in assigning rank such as, for example, SI load based upon different resource needs, e.g., a database application would be storage/memory intensive. Also a collaboration of different applications, or the opposite, could be considered. According to one exemplary embodiment, populating this configuration can be complemented by using a method that generates, based on the previously generated SI and CSI templates, all (or a desired portion) of the SI and CSIs respective attributes.

[0046] The exemplary processing described above with respect to the AMF Configuration Generator 402 will now be summarized as shown in the exemplary flowchart of FIG. 3(b). Initially, each received SI template is matched with the received entity types 404 in step 420. Matching also occurs in the follow on steps as needed. Processing then continues by finding a component type by selecting the component type within an SU type or among all component types (provided it is permitted to construct a new SU type) such that the component type provides a highest utilization for a specified redundancy model in step 422. Calculating required active capabilities for an SU type within an SG with respect to an SI template, wherein an SU of the SU type provides a group of SIs from a single template occurs in step 424. Similarly, calculating required standby capabilities for an SU type with respect to an SI template, wherein an SU of the SU type protects a group of SIs from a single template occurs in step 426. The Configuration Generator 402 then searches for a first SU type that can provide SIs based on one of the templates in step 428. This leads to searching for a first SG type that can provide SIs using a requested. SU type according to a specified redundancy model in step 430. Once the necessary entity types have been identified the configuration generation proceeds with populating the configuration with the entities of these types. This begins with generating all of the SIs and their respective CSIs to be protected by an SG based upon the templates in step 432. A first SU is created for the SG and then others are copied to create the complete SG. A next set of SIs and CSIs can then be generated for the next SG. If the SIs belong to the same SI template, the SG is just copied, until all the SIs required by the template have been assigned to an SG. For the next SI template the process can be repeated starting with the creation of a first SU. The process ends when all the SGs have been created for all SI templates.

[0047] The configuration, e.g., a file, generated by this exemplary process can, for example, be in an IMM XML format for use by an AMF via IMM. As mentioned above, the elements of this configuration can be generated by a type matching process which loops on the SI templates: for each SI template the SG type, the SU type and the component types are selected. Exemplary pseudocode illustrating this process in more detail is provided below. For the instance generation described above, there are additional loops as multiple entities are created for each selected type. The configuration, e.g., IMM XML, contains both the entity types and all the generated entities. During configuration generation, SUs can be assigned to nodes by setting the appropriate attributes. This is an optional attribute, so the AMF can use this attribute when it is provided, otherwise the AMF distributes the SUs in an implementation specific way. Note also that it may not always be desirable to assign an SU to a node, e.g., when the SU is assigned to a node it cannot be instantiated by AMF anywhere else, in which case, if the node is faulty, that SU would be lost with the node. If no node is specified AMF will pick a node from among those available.

[0048] Using the above described exemplary embodiments, the AMF Configuration Generator 402 generates a configuration which becomes an output of code in, for example, an IMM XML file 412. This output of an IMM XML file 412 is preferably provided in an acceptable format for use by an AMF system as described in, for example, SAF Information Model Management (IMM) specification SAI-AIS-IMM-A.02.01. This exemplary code is then stored in a Configuration Repository 310 (or other desirable computer readable medium or memory storage unit) in step 414 for future use by an AMF.

[0049] An additional process that occurs at the end of the exemplary generation of an AMF configuration is the method for distribution of the SUs on the cluster nodes. The AMF specification does not currently describe performance characteristics for entities, therefore an exemplary method for distribution of SUs can assume that each SU of the configuration produces equal workloads with respect to each other and with or without assignment. It is at this point that the number of nodes to be used needs to be known by the system as either a received user input or by the system knowing what is available, e.g., information provided by another SAF service. Based on this assumption, this exemplary method distributes the SUs among the nodes of the cluster by selecting for each SU the node that has the smallest load and that has no (or the fewest number of) SUs of the same SG that the current SU belongs to. These conditions can be provided by using a global round robin distribution of the SUs as the SUs are being created during the configuration process. Alternatively, this exemplary distribution method can be modified by assuming, based on performance characteristics of the system, different capacities for the different nodes and different load for the different SUs. Also, another exemplary alternative embodiment for configuration could be based upon the assumption of different failure scenarios desiring to avoid or minimize overload situations when one or more SUs/nodes fail in the system.

[0050] To further illustrate the above described exemplary embodiments for generating an AMF configuration, a purely illustrative exemplary embodiment will now be described with respect to FIGS. 4-9. In this exemplary system, the goal is to support file transfer protocol (FTP) services, simple mail transfer protocol (SMTP) services and hyper text transfer protocol (HTTP) services in an HA environment by creating a configuration file for use by an AMF to manage these services in an HA environment. For the input entity types 404 there are both component types and component service types. As can be seen in FIG. 4, there are two component types labeled CompType A 502 named Surge FTP and CompType B 504 named Apache. CompType A 502 can provide up to three active or six standby assignments for CST A 506 for FTP and, CompType A 502 can also provide up to one active or one standby assignments for CST B 508 for SMTP. Regarding CompType B 504, it is capable of providing up to three active and six standby assignments for CST C 510 for HTTP.

[0051] These various CSTs can have a number of different attributes associated with them as shown in FIG. 5. For example, CST A 506 for FTP is shown with two attributes 602 and 604 associated therewith. Attribute 602 could, for example, deal with IP ranges and attribute 604 could, for example, be associated with security levels. More or fewer attributes can be associated with a CST (as well as the other CSTs) as shown by CST B 508 for SMTP having three attributes and CST C 510 having just one attribute. These attributes associated with the CSTs are typically submitted as part of the input entity types 404 as they are determined by the way the software is implemented.

[0052] CSTs are also associated with service types. For example, as shown in FIG. 6, two service types are shown, SvcType A 702 representing the file transfers and email service and SvcType B 704 which represents the web page display service. Accordingly, SvcType A 702 is associated with CST. B 508 and CST A 506 representing the FTP and SMTP component service types. SvcType B is solely associated with CST C 510, with HTTP. These service types can be submitted as either entity types inputs 404 or as part of the services set/user input 406. This then leads to the various templates submitted as part of the services set/user input 406.

[0053] According to exemplary embodiments, the services set/user input 406 includes templates. Continuing with the current example, SI templates, CSI templates and a node template are illustrated in FIG. 7. An SI template 802 includes the following: a name, e.g., SI-temp-a 804; a redundancy model, e.g., N+M 806; a service type, e.g., File Transfer and Mail 808; the preferred number of active and standby assignments per SI, e.g., 1 Active+1 Standby 810; the values for N and M SUs as appropriate for the redundancy model, e.g., 3+1 SUs 812; and the desired number of SIs, e.g., 3 SIs. Associated with each SI template 802 are one or more CSI Templates 816 and 818. CSI template 816 includes a name, e.g., CSI-temp-A 820, a CST, e.g., FTP 822, and the number of CSIs in this case one 824. The other template is a node template 826 which includes a name 828 and the number of nodes 830 in the cluster. They are to be used for the distribution of the generated service units.

[0054] At some point after filling in the desired templates, these desired templates as well as the entity types 404 are processed by the AMF Configuration Generator 402. For example, as shown in FIG. 8, SI template 802 is being processed by the AMF Configuration Generator 402. Using the received inputs as well as other known information, such as specified constraints for an AMF, outputs are generated by the AMF Configuration Generator 402. In this case, populated SIs 902, 904 and 906 are generated. Additionally, in this example, a first SU 908 is created of a (new) SU type to meet the desired needs associated with the SIs (902, 904 and 906) of the SI template 802. A new SU type was created if there was no acceptable available SU type and the orphan types allowed for the creation of new SU type as needed. Also, the Configuration Generator 402 determined that the best CompType for CST A and CST B is CompType A which allows the SU 908 to only use three components to support each of the SIs 902, 904 and 906, which have three CSIs associated with them. To meet another need as specified by the SI template 802, e.g., the provision of three active SUs and one standby SU 812, four SUs are created and deployed as shown in FIG. 9 as part of the new SG 1002.

[0055] According to exemplary embodiments, illustrative pseudo-code used for generating an AMF configuration is shown below. This illustrative pseudo-code is broken down into three sections: Inputs, Outputs and Configuration Generation, for ease of understanding. The Inputs section of pseudo-code corresponds to the various inputs that may be received by the AMF Configuration Generator 402. Some of them are mandatory (e.g. component type, component service type, SI templates), others may be generated by the configuration generator (e.g. SU type). The Outputs are building blocks of the final configuration created from the Inputs by the Configuration Generation. The Outputs together with some of the Inputs presented to an AMF implementation via IMM in an XML file providing this way the complete AMF information model.

Inputs

[0056] The first group of inputs described are from the entity types 404 subset. A component service type is a set as described by

• [0057] cst={att_i; 0<=i<|cst|, any number of attributes}.
  All component service types compose the set
• [0058] CST={cst_i; 0<=i<|CST|}.
  A component capability, which is used in describing a component type, is described by a tuple of
• [0059] csCapability=<cst, compCap, maxAct, maxStdb>, where cst ε CST;

[0060] $\mathrm{compCap}=\{\begin{array}{c}\mathrm{x\_active}\mathrm{\_and}\mathrm{\_y}\mathrm{\_standby}\\ \mathrm{x\_active}\mathrm{\_or}\mathrm{\_y}\mathrm{\_standby}\\ \mathrm{l\_active}\mathrm{\_or}\mathrm{\_y}\mathrm{\_standby}\\ 1\mathrm{\_active}\mathrm{\_or}\mathrm{\_l}\mathrm{\_standby}\\ \mathrm{x\_active}\\ \mathrm{l\_active}\end{array}\mathrm{maxAct}=\{\begin{array}{c}\mathrm{x\_active}\mathrm{\_and}\mathrm{\_y}\mathrm{\_standby}\Rightarrow x,x\ge 1\\ \mathrm{x\_active}\mathrm{\_or}\mathrm{\_y}\mathrm{\_standby}\Rightarrow x,x\ge 2\\ \mathrm{l\_active}\mathrm{\_or}\mathrm{\_y}\mathrm{\_standby}\Rightarrow 1\\ \mathrm{l\_active}\mathrm{\_or}\mathrm{\_l}\mathrm{\_standby}\Rightarrow 1\\ \mathrm{x\_active}\Rightarrow x,x\ge 2\\ \mathrm{l\_active}\Rightarrow 1\end{array}\mathrm{maxStdb}=\{\begin{array}{c}\mathrm{x\_active}\mathrm{\_and}\mathrm{\_y}\mathrm{\_standby}\Rightarrow y,y\ge 1\\ \mathrm{x\_active}\mathrm{\_or}\mathrm{\_y}\mathrm{\_standby}\Rightarrow y,y\ge 1\\ \mathrm{l\_active}\mathrm{\_or}\mathrm{\_y}\mathrm{\_standby}\Rightarrow y,y\ge 2\\ \mathrm{l\_active}\mathrm{\_or}\mathrm{\_l}\mathrm{\_standby}\Rightarrow 1\\ \mathrm{x\_active}\Rightarrow 0\\ \mathrm{l\_active}\Rightarrow 0\end{array}$
A component type can be described as a set as follows.

• [0061] ct={csCapability_i; 0<=i<|ct|, all the component service (CS) types provided by the component type ct}
  All component types compose the set as described by
• [0062] CT={ct_i; 0<=i<|CT|}.

[0063] A second type of inputs may belong either to the entity types 404 or to the services set/user input 406 subset.

A member CS type is a tuple as described by

• [0064] memCst=<cst, maxCsi>, where cst ε CST;
  maxCsi is the number of CSIs of type cst that a service instance of a particular type may contain.
  A service type is a set
• [0065] st={memCst_i; 0<=i<|st| component service instances that compose the service type st}.
  All the service types compose the set
• [0066] ST={st_i; 0<=i<|ST|}.

[0067] A third type of inputs may belong either to the entity types 404 or to the services set/user input 406 subset or completely omitted as an input. In this latter case the configuration generation is creating these types.

Service Unit Types

The services of a service unit type is a set as described by sutServices={st_i; 0<=i<|sutServices|, all the service types the service unit provides and it is a subset of ST}.

A member component type is a tuple

• [0068] memCt=<ct, maxComp>where ct ε CT;
  maxComp is the maximum number of instances of this component type an SU of this type may contain.
  The component types of an SU type is a set
• [0069] sutCt={memCt_i; 0<=i<|sutCt|, all the component types the service unit may contain and |sutCt|<=|CT|}.
  A service unit type is a tuple
• [0070] sut=<sutCt, sutServices>
  All the service unit type compose the set
• [0071] SUT={sut_i; 0<=i<|SUT|}
  This set may be empty (SUT={ }) at the beginning of the configuration generation.
  Service Group Types

[0072] The service unit types of a service group type is a set as described by sgtSut={sut_i; 0<=i<|sgtSut|, all the service unit types the service group may contain and it is a subset of SUT}

A service group type is a tuple

• [0073] sgt=<sgtSut, redMod>, where

[0074] $\mathrm{redMod}=\{\begin{array}{c}2n\\ \mathrm{nplusm}\\ \mathrm{nway}\\ \mathrm{nwayactive}\\ \mathrm{noredundancy}\end{array}$
All service group types compose the set

• [0075] SGT={sgt_i; 0<=i<|SGT|}
  This set may be empty (SGT={ }) at the beginning of the configuration generation.

[0076] A fourth type of inputs is the services set/user input 406 which, as described according to exemplary embodiments above, includes Service Instance templates for the service instances and Component Service Instance templates for their component service instances that need to be protected by Service Groups. Pseudo-code for a variety of services set/user input 406 is shown below.

CSI Templates

A CSI template is a tuple

• [0077] csiTemp=<cst, numCsi>, where cst ε CST;
  numCsi is the number of CSIs of type cst created based on this template.
  All CSI templates compose the set
• [0078] CSIT={csiTemp_i; 0<=i<|CSIT|}.
  SI Templates
  A SI template is a tuple
• [0079] siTemp=<st, sicsTemps, redMod, numSIs, numAct, numStdb, numSUs>,
  where
• [0080] st ε ST;
• [0081] sicsiTemps={csiTemp_i; 0<=i<|sicsiTemps| ∀csiTemp_i.cst≡st_i.cst and csiTemp_i.numCsi≦st_i.maxCsi};
  numSIs is the number of SI that is created using this template

[0082] $\mathrm{numAct}=\{\begin{array}{c}\mathrm{redMod}\ne \mathrm{nwayactive}\Rightarrow 1\\ x;x\ge 2,\end{array}$
is the number of HA active assignments per SI;

[0083] $\mathrm{numStdb}=\{\begin{array}{c}\mathrm{redMod}\equiv \mathrm{nway}\Rightarrow y;y\ge 1\\ \mathrm{redMod}\equiv \mathrm{nwayactive}\Rightarrow 0\\ \mathrm{redMod}\equiv \mathrm{noredundancy}\Rightarrow 0\\ 1,\end{array}$
is the number of HA standby assignments per SI;

• [0084] numSUs=<sus, susAct, susStdb, susSpare>, where
  sus the total number of SUs in the SG protecting the SIs generated from this template; sus>=susAct+susStdb+susSpare;

[0085] $\mathrm{numAct}=\{\begin{array}{c}\mathrm{redMod}\equiv \mathrm{nplusm}\Rightarrow z\\ \mathrm{redMod}\equiv \mathrm{nwayactive}\Rightarrow z;z>1\\ \mathrm{redMod}\equiv \mathrm{noredundancy}\Rightarrow \mathrm{numSIs}\\ \mathrm{redMod}\equiv 2n\Rightarrow 1\\ \mathrm{redMod}\equiv \mathrm{nway}\Rightarrow 0,\end{array}$
is the number of SUs that can have only active assignments;

[0086] $\mathrm{susStbd}\mathrm{numStdb}=\{\begin{array}{c}\mathrm{redMod}\equiv \mathrm{nplusm}\Rightarrow w\\ \mathrm{redMod}\equiv \mathrm{nwayactive}\Rightarrow 0\\ \mathrm{redMod}\equiv \mathrm{noredundancy}\Rightarrow 0\\ \mathrm{redMod}\equiv 2n\Rightarrow 1\\ \mathrm{redMod}\equiv \mathrm{nway}\Rightarrow 0\end{array}$
is the number of SUs that can have only standby assignments;
susSpare is the number of spare SUs.
All SI templates compose the set

• [0087] SIT={siTemp_i; 0<=i<|SIT|}.

[0088] The final type of input is the number of nodes. Number of nodes can be an optional separate input, or defined elsewhere, e.g., read from the system for which the configuration is being generated. The number of nodes can be described using pseudo-code as a node template as follows. Each node template characterizes a distinct set of nodes that share characteristics. Currently only one template is used that represents all the nodes of the cluster.

The node template is a tuple

• [0089] nodeTemp=<numNodes>, where
  numNodes is the number of nodes created based on the template.
  All node templates compose the set
• [0090] NodeTemp={nodeTemp_i; 0<=i<|NodeTemp|}.

[0091] Outputs are a result of processing by the AMF Configuration Generator 402 to be used in the final configuration file. The pseudo-code used to describe and generate these Outputs is described below.

Outputs

Components

A component has a name and belongs to component type. It is a tuple coop=<safComp, ct>, where ct ε CT;

• [0092] safComp is a tuple safComp=<safSu, rdn>, where
  safSu is a SAF compliant LDAP distinguished name of the SU the component belongs to and
  rdn is a relative distinguished name of the component, which is unique within the SU.
  Service Units
  A service unit is a tuple
• [0093] su=<safSu, sut, memComps>, where
• [0094] safSu is a tuple safSu=<safSg, rdn>, where
  safSg is a SAF compliant LDAP distinguished name of the SG the service unit belongs to and
  rdn is a relative distinguished name of the SU which is unique within the SG; sut ε SUT;
• [0095] memComps={memComp_i; 0<=i<|memComps|, ∀memComp.safComp.safSu≡safSu;}
  memComp={ } at the start of the configuration generation.
  Service Groups
  A service group is a tuple
• [0096] sg=<safSg, sgt, memSus, nodeGroup>, where
• [0097] safSg is a tuple safSg=<safApp, rdn>, where
  safApp is a SAF compliant LDAP distinguished name of the application the service group belongs to, and
  rdn is a relative distinguished name of the SG, which is unique within the application; sgt ε SGT;
• [0098] memSus={memSu_i; 0<=i<|memSus|, memSu_i.sut ε sgt.sgtSut and ∀ memSu.safSu.safSg≡safSg;}
• [0099] memSUs={ } at the start of the configuration generation
• [0100] nodeGroup={node_i; 0<=i<|nodeGroup|,}; the nodeGroup ⊂ Cluster on which the SG is distributed.
  Component Service Instances
  A component service instance is a tuple
• [0101] csi=<safCsi, cst, attrValues>, where
• [0102] safCsi is a tuple safCsi=<safSi, rdn>, where
  safSi is a SAF compliant LDAP distinguished name of the service instance the component service instance belongs to, and
  rdn is a relative distinguished name of the CSI, which is unique within the service instance;
  cst ε CST;
• [0103] attrValues={attrValue_j; 0<=i<|cst|}
• [0104] attrValue=<attr, value>, where
• [0105] ∀attrValue.attr ε cst.attr;
  attrValue.value is a value set for the attribute
• [0106] attrValues={ } the method does not currently include the generation of attribute values.
  Service Instances
  A service instance is a tuple
• [0107] si=<safSi, st, sg, memCsis>, where
  safSi is a tuple safSi=<safApp, rdn>, where
  safApp is a SAF compliant LDAP distinguished name of the application the service instance belongs to, and
  rdn is a relative distinguished name of the Si, which is unique within the application;
  st ε ST;
• [0108] memCsis={memCsi_i; 0<=i<|memCsis|, ∀memCsi.cstst.cst; and ∀memCsi.safCsi.safSi≡safSi;}
• [0109] memCsis={ } at the start of the configuration generation
• [0110] sgSis={si_i; 0<=i<|sgSis|, ∀si_i.sg≡sg} the set of SIs protected by the same service group, they specify the same service group.
  Application
  The application for which the configuration is generated is a tuple
• [0111] app=<safApp, memSgs, memSis>, where
  safApp is a distinguished name of the application;
• [0112] memSgs={memSg_i; 0<=i<|memSgs|, ∀memSg.safSg.safApp≡safApp;} memSgs={ } at the start of the configuration generation.
• [0113] memSis={memSi_i; 0<=i<|memSis|, ∀memSi.safSi.safApp≡safApp;}
• [0114] memSis={ } at the start of the configuration generation.
  Nodes
• [0115] node=<safAmfNode, hostedSus>, where
  safAmfNode is a distinguished name of the node;
• [0116] hostedSus={hostedSu_i; 0<=i<|hostedSus|}
• [0117] hostedSus={ } at the start of the configuration generation.
  All nodes compose the cluster:
• [0118] Cluster={node_i; 0<=i<|Cluster|}

[0119] findCompType Method

Input

• • [0120] sut, the service unit type within which the search is performed, it may be empty
  • [0121] cst, the component service type that needs to be provided
  • [0122] redMod, the redundancy model within which cst is provided
    Output
  • [0123] chosenCompType, the selected component type
    Functional Description

[0124] The find CompType method selects the component type within the service unit type or among all component types that can provide the highest utilization for the given redundancy model.

[0125] validCompTypes is the set of component types that can provide a required component service type cst within the service unit type sut or among all the component types:

[0126] $\mathrm{validCompTypes}=\{{\mathrm{vct}}_i;\{\begin{array}{c}\begin{array}{c}\mathrm{sut}\ne \left\{\right\}\Rightarrow \forall \mathrm{sut}.{\mathrm{sutCt}}_j.{\mathrm{csCapability}}_k.\mathrm{cst}\equiv \\ \mathrm{cst}\Rightarrow {\mathrm{vct}}_i\end{array}\\ \mathrm{sut}\equiv \left\{\right\}\forall {\mathrm{ct}}_j.\mathrm{csCapability}.\mathrm{cst}\equiv \mathrm{cst}\Rightarrow {\mathrm{vct}}_i,\end{array}{\mathrm{ct}}_i\in \mathrm{CT}.$

[0127]

	begin
	MaxMin = 0
	chosenCompType = { }
	i = 0
	while (i < |validCompTypes|) // for each valid component type
	vct
	begin
	j = 0
	while (j < |vcti.csCapability|) // for each CS type it provides
	begin
	if redMod == 2n OR redMod == nplusm
	Min = min(vcti.csCapabilityj.maxAct,
	vcti.csCapabilityj.maxStandby)
	if redMod == nway
	// x_active_and_y_stanby is the mandatory capability
	// for nway
	if vcti.csCapabilityj == x_active_and_y_standby
	Min=min(vcti.csCapabilityj.maxAct,
	vcti.csCapabilityj.maxStandby)
	else
	Min = 0 // cannot be used for nway
	if redMod == nwayactive OR redMod == noredundancy
	Min= vcti.csCapabilityj.maxAct
	if Min > 0 AND Min > MaxMin
	begin
	MaxMin = Min
	chosenCompType = cti
	end
	j = j + 1
	end
	i = i + 1
	end
	if chosenCompType == { }
	return cst cannot be provided
	else
	return chosenCompType
	end

[0128] suActCap Methods

Input

• • [0129] siTemp, an SI template
    Output
  • [0130] suActCap, the required active capability for the SU type or an SU with respect to the Si template
    Functional Description

[0131] The required SU active capabilities are calculated with respect to the SI template with the assumption that an SU provides a group of SIs from a single template. In the formula ceil( ) is used to round up to the nearest integer.

[0132] $\mathrm{return}$ $\mathrm{suActCap}=\{\begin{array}{c}\mathrm{redMod}\equiv \mathrm{nway}\Rightarrow \mathrm{ceil}\left(\frac{\mathrm{siTemp}.\mathrm{numSIs}}{\begin{array}{c}\mathrm{siTemp}.\mathrm{numSUs}.\mathrm{sus}-\\ \mathrm{siTemp}.\mathrm{numSUs}.\mathrm{susSpare}\end{array}}\right)\\ \mathrm{redMod}\equiv \mathrm{nwayactive}\Rightarrow \mathrm{ceil}\left(\frac{\begin{array}{c}\mathrm{siTemp}.\mathrm{numSIs}\times \\ \mathrm{siTemp}.\mathrm{numAct}\end{array}}{\mathrm{siTemp}.\mathrm{numSUs}.\mathrm{susAct}}\right)\\ \mathrm{redMod}\equiv \mathrm{noredundancy}\Rightarrow \mathrm{siTemp}.\mathrm{numSIs}\\ \mathrm{redMod}\equiv 2n\Rightarrow \mathrm{siTemp}.\mathrm{numSIs}\\ \mathrm{redMod}\equiv \mathrm{nplusm}\Rightarrow \mathrm{ceil}\left(\frac{\mathrm{siTemp}.\mathrm{numSIs}}{\mathrm{siTemp}.\mathrm{numSUs}.\mathrm{susAct}}\right)\end{array}$

[0133] suStdbCap Methods

Input

• • [0134] siTemp, an Si template
    Output
  • [0135] suStdbCap, the required standby capability for the SU type or an SU with respect to the SI template
    Function Description

[0136] The required SU standby capabilities are calculated with respect to the SI template with the assumption that an SU provides a group of SIs from a single template. In the formula ceil( ) is used to round up to the nearest integer.

[0137] $\mathrm{return}$ $\mathrm{suStdbCap}=\{\begin{array}{c}\mathrm{redMod}\equiv \mathrm{nway}\Rightarrow \mathrm{ceil}\left(\frac{\begin{array}{c}\mathrm{siTemp}.\mathrm{numSIs}\times \\ \mathrm{siTemp}.\mathrm{numStdb}\end{array}}{\begin{array}{c}\mathrm{siTemp}.\mathrm{numSUs}.\mathrm{sus}-\\ \mathrm{siTemp}.\mathrm{numSUs}.\mathrm{susSpare}\end{array}}\right)\\ \mathrm{redMod}\equiv \mathrm{nwayactive}\Rightarrow 0\\ \mathrm{redMod}\equiv \mathrm{noredundancy}\Rightarrow 0\\ \mathrm{redMod}\equiv 2n\Rightarrow \mathrm{siTemp}.\mathrm{numSIs}\\ \mathrm{redMod}\equiv \mathrm{nplusm}\Rightarrow \mathrm{ceil}\left(\frac{\mathrm{siTemp}.\mathrm{numSIs}}{\mathrm{siTemp}.\mathrm{numSUs}.\mathrm{susAct}}\right)\end{array}$

[0138] findSUType Method

Input

• • [0139] siTemp, an Si template
    Output
  • [0140] chosenSuType, the selected service unit type
    Functional Description

[0141] This function searches for the first SU type that can provide the SIs of the given template siTemp. Optimization may be added depending on other requirements.

[0142] validSuTypes is the set of service unit types that can provide the required service type siTemp.st among all the service unit types:

[0143] $\mathrm{validSuTypes}=\{{\mathrm{vsut}}_i;\mathrm{SUT}\ne \left\{\right\}\{\begin{array}{c}\begin{array}{c}\mathrm{siTemp}.\mathrm{redMod}\ne \mathrm{nway}\Rightarrow \forall {\mathrm{sut}}_j.\mathrm{st}\equiv \\ \mathrm{siTemp}.\mathrm{st}\Rightarrow {\mathrm{vsut}}_i\end{array}\\ \mathrm{siTemp}.\mathrm{redMod}\equiv \mathrm{nway}\Rightarrow \\ \begin{array}{c}\Rightarrow \forall {\mathrm{sut}}_j.\mathrm{memCt}.\mathrm{csCapability}.\mathrm{compCap}\equiv \\ \mathrm{x\_active}\mathrm{\_and}\mathrm{\_y}\mathrm{\_standby}\bigwedge \end{array}\\ \bigwedge \forall {\mathrm{sut}}_j.\mathrm{st}\equiv \mathrm{siTemp}.\mathrm{st}\Rightarrow {\mathrm{vsut}}_i\end{array}{\mathrm{sut}}_j\in \mathrm{SUT};$

[0144] It may be an empty, set in which case a new SU type is created if it is permitted by limited.

[0145]

	begin
	chosenSuType = { }
	if validSuTypes ≠{ }
	begin
	i = 0
	while(i < |validSUTypes|) // find an SU type that
	begin
	j = 0
	while(j < |siTemp.sicsiTemp|) // supports all CSI templates
	begin
	// use findCompType to find an appropriate component
	if
	NOT(
	∃k,l  siTemp.csiTempj.cst ≡ vsuti.memCtk.csCapabilityl.cst
	)
	j = j + |siTemp.sicsiTemp|
	// make it break by being too big
	else
	if ((vsuti.memCtk.csCapabilityl.maxAct x
	x vsuti.memCtk.csCapabilityl.maxComp) <
	< (siTemp.csiTempj.numCsi x
	suActCap(siTemp))) OR
	((vsuti.memCtk.csCapabilityl.maxStdb
	x vsuti.memCtk.csCapabilityl.maxComp) <
	< (siTemp.csiTempj.numCsi x
	suStdbCap(siTemp)))
	j = j + |siTemp.sicsiTemp|
	// make it break by being too big
	j = j + 1
	end
	// next condition is only true if break occurred,
	// if so go for next SU type
	if j > |siTemp.sicsiTemp|
	i = i + 1
	else
	begin
	chosenSuType = vsuti
	return chosenSuType
	// returns the first appropriate SU type
	end
	end
	end
	if chosenSuType == { }
	// if nothing is found or no valid SU types
	begin
	chosenSuType = <{ }, siTemp.st> // create a new SU type
	i = 0
	while(i < |siTemp.sicsiTemp|)
	// that supports all CSI templates
	begin
	vct = { }
	vct = findCompType(chosenSuType,
	siTemp.csiTempi.cst,siTemp.redMod)
	if vct == { }
	return { } // cst cannot be provided
	if vct ∉ sut.sutCt
	chosenSuType.sutCt =
	chosenSuType.sutCt ∪ {<vct, NoLimit>}
	i = i + 1
	end
	SUT = SUT ∪ { chosenSuType }
	return chosenSuType
	end
	end

[0146] findSGType Method

Input

• • [0147] sut, an SU type
  • [0148] redMod the required redundancy model
    Output
  • [0149] chosenSgType, the selected service group type
    Functional Description

[0150] This function searches for the first SG type that can protect the SIs using the requested service unit type sut according to the requested redundancy model redMod. Optimization may be added depending on other requirements.

[0151]

	begin
	chosenSgType = { }
	i = 0
	while(i < |SGT|)
	begin
	// the SG type allows for the SU type and the redundancy
	model
	if sgti.redMod == redMod AND ∃k  sgti.sutk ≡ sut
	begin
	chosenSgType = sgti
	return chosenSgType
	end
	i = i + 1
	end
	if chosenSgType == { }
	begin
	chosenSgType = <sut, redMod>
	SGT = SGT ∪ { chosenSgType }
	return chosenSgType
	end
	end

[0152] createSis Method

Input

• • [0153] siTemp, an SI template for which the SIs are being generated
  • [0154] sg, the sg that shall protect this set of SIs
    Output
  • [0155] sgSis, the set of service instances protected by a service group
    Functional Description

[0156] Based on the provided SI template, this method generates all the service instances and their component service instances that shall be protected by a service group Currently this function does not include the generation of CSI attributes.

[0157]

	begin
	sgSis = { }
	i = 0
	while(i < |siTemp.numSIs|)
	begin
	si = <”safSi=...”, siTemp.st, sg, { } >
	j = 0
	while (j < |siTemp.sicsiTemps|) // for each CS type in the SI
	begin
	k = 0
	while (k < |siTemp.csiTempj.numCsi|)
	// create the specified number of CSIs
	begin
	si.memCsis = si.memCsis U {<”safCsi=...”,
	siTemp.csiTempj.cst, { } >}
	k = k + 1
	end
	j = j + 1
	end
	sgSis = sgSis U {si}
	i = i + 1
	end
	return sgSis
	end

[0158] Final configuration generation will now be described using the pseudo-code as described below.

[0159] generateConf Method

Input

• • [0160] CT all component types
  • [0161] CST all component service types
  • [0162] ST all service types
  • [0163] SIT all service instance templates
  • [0164] CSIT all component service instance templates
  • [0165] SUT service unit types, may be empty
  • [0166] SGT service group types, may be empty
  • [0167] nodeTemp the node template
  • [0168] userPreference is a set of preferences that guides the configuration generation. They allow for different options and optimization and can be extended accordingly. The pseudo-code does not reflect them.
    Output
  • [0169] AMF entities for a single application distributed on the nodes of a cluster, i.e. app=<safApp, memSgs, memSis>. Since the application refers to the SG set and the SI set, and they all refer to their children in turn, only the application itself is defined in this example as an output. This “app” is thus a tree with all the instances according to the Output section.
    Functional Description

[0170] This is the main function which, after determining the required entity types using the findSUType and findSGType methods, populates the configuration with all the instances listed as outputs. It also distributes the generated service units on the nodes.

[0171]

begin
	app = < “safApp=. . . ”, { }, {} > // create an application instance
of the output
	i = 0
	m = 0
	while(i < |SIT|) // for each SI template
	begin
	sut = { }
	sut = findSUType(siTempi) // find appropriate SU type
	if sut == { }
	return // no configuration can be generated
	sgt = { }
	sgt = findSGType(sut, siTempi.redMod)
	// find appropriate SG type
	sg = <“safSg=. . . ”, sgt, { }, {},{} >
	// a first SG instance of the output set is created
	app.memSis = createSis(siTempi, sg)
	// create the protected SIs instances and assigned
	// them to that SG instance
	// create first SU in the SG
	su = < “safSu=. . . ”, sut, { } >
	sg.memSus = su.memSus U {su}
	sg;nodeGroup = sg.nodeGroup U {nodem}
	nodem.hostedSus = nodem.hostedSus U {su}
	m = (m + 1) mod |Cluster|
	j = 0
	ct_set = { } // component types in use
	su_act = suActCap(siTempi)
	su_stdb = suStdbCap(siTempi)
	while(j < |siTempj.sicsiTemps|) // for all CS types
	begin
	ct = findComType ⁡ ( sut , siTemp i . csiTemp j . cst , siTemp i . redMod ) N = max ⁡ ( ceil ⁡ ( su_act × siTemp i . csiTemp j . numCsi ct . csCapability . maxAct ) , ceil ⁡ ( su_act × siTemp i . csiTemp j . numCsi ct . csCapability . maxStdb ) )
	mem_ct = <ct, N>
	// number of instances of a component type
	if mem_ct ∉ ct_set
	begin
	ct_set = ct_set U {mem_ct}
	// remember the number of instances
	k = 0
	while (k < N)
	// create the needed number of components
	begin
	su.memComps =
	su.memComps U {<“safComp =. . . ”, ct >}
	//components are created and added to the first SU
	k = k + 1
	end
	end
	else
	begin
	// find how many components are now in the SU
	∃l  mem_ctl.ct ≡ ct, mem_ctl, ∈ ct_set
	If mem_ctl.N < N // if not enough
	begin
	k = 0
	while (k < N - mem_ctl.N)
	// create the needed number of components
	begin
	su.memComps = su.memComps U
	{<“safComp = . . . ”, ct >}
	k = k + 1
	end
	mem_ctl.N = N
	end
	end
	j = j + 1
	end
	k = 1
	while (k < siTempi.numSus.sus)
	// clone the needed number of SUs
	begin
	new_su = < “safSu= . . . ”, su.sut, su.memComps>
	sg.memSus = su.memSus U {new_su}
	//second, third etc. SUs are created and added to the SG
	k = k + 1
	sg.nodeGroup = sg.nodeGroup U {nodem}
	nodem.hostedSus = nodem.hostedSus U {su}
	m = (m + 1) mod |Cluster|
	end
	app.memSgs = app.memSgs U {sg}
	//the SG is added to the application
	i = i + 1
	end
end

[0172] The exemplary embodiments described above provide methods and systems for generating an output of code for an IMM XML file 412 by an AMF Configuration Generator 402 for use by an AMF via IMM, typically in a HA environment. An exemplary node 1100 which can act as an AMF Configuration Generator 402 will now be described with respect to FIG. 10. Node 1100 can contain a processor 1102 (or multiple processor cores), memory 1104, one or more secondary storage devices 1106 and an interface unit 1108 to facilitate communications between node 1100 and the rest of the network, i.e., sources of input and receiver of the output. Additionally, the node 1100 can contain control instructions for automating the generation of the system model based upon received inputs. The memory (or the secondary storage) can be used for storage of exemplary inputs such as user inputs or known required entity types 404 as defined by current AMF standards. Thus, a node 1100 according to an exemplary embodiment may include a processor for generating a system model according to the format of an IMM XML schema for future (or current) use by an AMF.

[0173] Utilizing the above-described exemplary systems according to exemplary embodiments, a method for generating a configuration file is shown in the flowchart of FIG. 11. Initially a node receives inputs, associated with at least one service to be managed by the AMF and software used to provide the at least one service in step 1202. The node then processes the inputs to support the at least one service in step 1204. The node then generates a configuration using the processed inputs in step 1206, if possible. It will be appreciated that it is possible that no configuration can be generated if, for example, the required services cannot be provided by the system or software with the desired level of protection (redundancy).

[0174] It will be appreciated that the foregoing exemplary embodiments are merely illustrative and that numerous variations and permutations thereof are within the scope of this document. For example, although the configuration generated by these exemplary embodiments is described as an IMM XML, it will be appreciated by those skilled in the art that this is the input format for IMM, and not necessarily that which is provided directly to an AMF implementation. An AMF implementation may instead have other, non-standard input formats. The configuration which is generated by these exemplary embodiments may be presented in such format(s).

Top Down Approach

[0175] The foregoing exemplary embodiments describe systems, methods, devices and software for generating an AMF configuration based on, for example, user requirements and entity types provided by the vendor. Some of these exemplary embodiments employ a “bottom up” approach for selecting an SU type and component type as part of the AMF configuration generation process. That is, some of these exemplary embodiments search for a suitable SU type and component type set in an entities type file 404 based on the inputs which are provided to the configuration process and then generate compatible higher level type(s), e.g., an SG type, as part of the AMF configuration generation process. However, according to another exemplary embodiment described below, it may first be useful to look at the constraints (if any) associated with the higher level types in the input entity types file 404 as part of the search for a suitable SU type and component type in order to avoid generating unnecessary higher level types and/or higher level types which are inconsistent with such constraints.

[0176] Consider, for example, the conceptual view of an entity types file 404 illustrated as FIG. 12 which has been input to an AMF configuration generation process according to this exemplary embodiment. Therein, the entity types file 404 from, e.g., a vendor, has specified one complete group of types, e.g., an app type, an SG type, an SU type and a component type represented by column 1300 (each of which may have several instances associated therewith). The remaining types, e.g., represented by columns 1302, 1304 and 1306 are incompletely specified in the entity types file 404. The circled types in these columns are referred to herein as “orphaned” types because they are not supported by higher level types in the entity types file 404.

[0177] According to this exemplary embodiment, searching for a service unit type and a component type to support a service can be performed in an order which involves first evaluating those groups of types which are more fully specified and then evaluating those groups of types which are less fully specified. This enables, for example, such a search to consider the constraints which may be provided by the vendor for the higher level types and/or to avoid creating higher level types when a service may be supported by a service unit type and component type which are already supported by higher level types. In this context, such restrictions or constraints may include, for example, one or more of: relations between types, limitations, capabilities dependencies, or other constraints. An example of such a constraint is a redundancy type associated with the SG type. An example of a search process for selecting a service unit type and component type in support of a service according to an exemplary embodiment is illustrated in the flowchart of FIG. 13.

[0178] For a given set of service instance template requirements 1400, the process first evaluates an application type in an entity types file at step 1402 to determine whether that application type has a constraint which conflicts with one of the inputs to the configuration generation process, e.g., a user input requirement associated with redundancy. Stated differently, step 1402 evaluates the existing application type to determine whether it satisfies the requirements associated with this service instance template. Using, as an exemplary reference, the entity types file shown in FIG. 12, the application type in column 1300 could, for example, be evaluated initially at step 1402. If a conflict exists, then the flow follows the “No” path from block 1402 to select another application type from the entity types file at step 1403. Alternatively, if the first selected application type satisfies the requirements for this service instance template, then the flow moves on to step 1404, wherein a service group type associated with the first selected application type in column 1300 is checked to see if it meets the requirements. If so, then the flow moves down to the next lower level at step 1406 wherein a service unit type associated with that service group type is checked to see if it satisfies the requirements for this service. If so, then that SU type, SG type and application type are selected for this service instance template at step 1408. If, however, a conflict is detected at the service group type level or the service unit type level, then another entity type at the same level can be evaluated (if one exists in the entity types file) by following the “No” path from the respective decision blocks 1404 or 1406, and looping back through blocks 1410 or 1412.

[0179] If the AMF configuration generator according to this exemplary embodiment cannot find a set of SU type, SG type and application type among those in the entity types file which satisfy the requirements for this service template, then the flow exits the “Yes” path from block 1403 to consider the orphaned service group type(s) (if any) in the entity types file. For example, if the types associated with the application type in the column 1300 did not provide a suitable set of types for the service instance template of interest, then the process could move on to column 1302 which includes an orphan service group type. This orphan service group type is then evaluated at step 1414 to determine if it has a conflict with the requirements, e.g., related to redundancy. If not, then the flow moves on to step 1416 wherein one or more service unit types associated with this service group type are checked to see if they can satisfy the requirements. If a set of SU type and an orphan SG type from the entity types file satisfy the requirements for this service instance template, then those entity types are selected and an application type can be created at step 1418 to support the SU type and SG type.

[0180] If the orphan service group type satisfies the requirements for this service instance template, but the first checked service unit type does not, then other service unit types associated with this service group type (if any exist) may be evaluated by looping back to step 1416 through step 1420. If none of the service unit types associated with this orphan service group type satisfy the requirements, then the flow can follow the “Yes” path from decision block 1420 wherein other orphan service group types can then be evaluated, if any exist in this entity types file. Otherwise, if this is the last orphan service group type in this entity type file, then the flow can follow the “Yes” path from block 1422 to block 1424 wherein orphan service unit types are evaluated, e.g., the orphan service unit type shown in column 1304 of FIG. 12. If this (or another via the loop including step 1428) orphan service unit type fulfills the requirements associated with the service instance template 1400, then that orphan service unit type is selected in step 1426 and an SG type and application type are created to support that SU type.

[0181] Otherwise, if no orphan service unit types satisfy the requirements imposed by the service instance template 1400, then the flow follows the “Yes” branch from step 1428 to step 1430. Therein, it can be checked to see whether a suitable service unit type can be constructed from orphan component types. If so, then a supporting SU type, SG type and application type are created at step 1432. If not, then the service instance template cannot be satisfied at step 1434 and the process terminates. Note that, in steps 1418, 1426 and 1432, new types may be created or existing types can be checked to see if they satisfy the SI requirements in which case they can “adopt” the orphaned type which is being selected.

[0182] It should be appreciated that the particular order of search is not critical to these exemplary embodiments and, therefore, the SU types found in an entity types file may be evaluated in any desired order as part of the AMF configuration process. However, according to some exemplary embodiments, the AMF configuration generator tool may not be permitted to generate new instances of higher level types, e.g., App types and SG types, unless the SU type under consideration is an orphan, i.e., is not supported by the higher level types in an entities file. Thus, using again the exemplary entity types file 404 conceptually illustrated in FIG. 12, an AMF configuration tool according to one exemplary embodiment would be permitted to create the types illustrated above the circled, orphaned types in FIG. 14, but would not be permitted to create, e.g., another SG type in column 1300.

[0183] This latter possibility is exemplified in FIG. 15 wherein an AMF configuration generator (or a type finder portion 1600 thereof) according to this exemplary embodiment first searches for an appropriate SU type among the type groups which have higher level types in the entities file, e.g., as represented by rows 1300 and 1302 and the overlaid arrows therein, in the manner discussed above with respect to FIG. 13. The search is, as described above, based on the calculated parameters returned from calculator portion 1602, e.g., in a manner similar to that described above and in the exemplary pseudocode below. Having not found a suitable SU type in these type groups, the type finder 1600 then evaluates SU type 1604 and finds it to be suitable for supporting the service of interest. To support a new SU type 1604, the type creator portion 1606 of the AMF configuration generator then creates an App type and SG type in column 1304, e.g., in the manner described above with respect to the “bottom up” embodiments. All of these processes are described in more detail below.

[0184] Thus, an overall method for generating an AMF configuration according to this exemplary embodiment is illustrated in the flow diagram of FIG. 16. Each of these steps will be described in some detail. Therein, at step 1700, the data to be used in the configuration process is collected. Generally speaking, such inputs include two different parts, e.g., the entity types describing the software applications, and the set of services that the particular system should provide using this software. Regarding the entity types, these are typically provided by the software vendor as an XML entity types file. The XML schema for an entity types file is specified by the SAF SMF specification as described above. Accordingly, there are two mandatory entity types that the entity types file must contain for the software, i.e., the component types and the component service types they provide. Other entity types may also be provided, however these types do not have to be present in an entity types file if there are no restrictions or constraints related to them. On the other hand, an AMF configuration contains all of the entity types defined by AMF, therefore these higher level types are also one of the expected outputs of these exemplary embodiments. The entity types in the entity types file can be classified to find those types that are not referenced by any other type, i.e., the orphan types described above. More specifically, the orphan component types are not referenced by any service unit type, the orphan service unit types are not referenced by any service group type, and the orphan service group types are not referenced by any application type. As mentioned earlier, according to some exemplary embodiments, orphaned types have no limitations on how they are used with respect to, other higher level types whereas unorphaned types may be not be permitted to have higher level types created therefor.

[0185] The second part of the inputs which are collected at step 1700 are the services that the site will provide for which the AMF configuration is being created. As mentioned above, the AMF specification defines the services as a set of service instances in which each service instance is described by a set of component service instances. This is site specific information and will be entered into the AMF configuration generation tool according to exemplary embodiments by a user, e.g., a site designer. To ease this task this exemplary embodiment provides for service instance (SI) templates that characterize sets of service instances which are similar and that shall be protected in a similar manner. Therefore each template specifies a set of SIs that are protected by service groups (SG) of the same type. Within each service instance template the participating component service instances (CSI) are also defined via templates. These CSIs are based on the component service types provided by the software, and indicate how many instances are necessary for the composition of a service instance. An SI template also defines, for example, to which service type it belongs, the desired redundancy model, the number of service instances based on this template and their grouping of SIs to service groups. Based on the set of the SI templates, the corresponding set of service instances and their component service instances is generated by this exemplary embodiment. This can be complemented, according to some exemplary embodiments, by a procedure which generates all of the service and component service instances with their exact attribute settings.

[0186] The configuration generator (e.g., the generateConf method in the pseudocode below) loops through the SI templates specified by the user. Returning to FIG. 16, for each SI template, the entity types that can provide the service instances are selected at step 1702, e.g., by finding them among the existing types and/or creating new types. The type selection step 1702 starts with matching an SI template with the entity types provided as input. According to this exemplary embodiment, first the service group types of the existing application types are checked, e.g., to determine if the redundancy model of the service group type matches the one required by the SI template. If it does, then the service unit types of this service group type are checked (e.g., using the findSut method in the pseudocode below) to determine if they can (a) provide the required service type and (b) support the load of SIs expected to be assigned to them. If both of these conditions are met then, according to this exemplary embodiment, the service unit type, the service group type and the parent application are chosen for the service instances of the SI template.

[0187] On the other hand, if none of service group types of any application type can provide the service, then the “orphan service group types” are checked in the same way as described above. However, since these service group types are not referenced by any application type, when a suitable service unit type is found, an application type needs to be created. To perform application type creation, the already created application types are checked to determine whether any of them has the service group type as a member or can be extended with that service group type (e.g., using the findAppt method in the pseudocode below). If not, a new application type is created (e.g., using the createAppType method in the pseudocode below) and added to the set of created application types (createdAPPTs). If none of the orphan service group types can provide the service, then the “orphan service unit types” are checked in a similar way. Again, when a service unit type is found, a service group type is either created (e.g., using the createSgType method in the pseudocode below) or found (e.g., using the findSgt method in the pseudocode below to search the set of createdSGTs) among the already existing created service group types to adopt the orphan service unit type. If a new service group type needs be created (it is added to the set of created service group types createdSGTs), then the appropriate application type also needs to be found or created. If none of the orphan service unit types can provide the service, then a service unit type is created (e.g., using the createSuType in the pseudocode below), and then adopted by a found/created service group type, that in turn will be adopted by a found/created application type.

[0188] Once all of the appropriate entity types have been found for an SI template, this exemplary embodiment then generates the entities of these types (e.g., within the generateConf in the pseudocode below) that would provide the service instances needed to support the requested services at step 1704. That is the AMF configuration is populated so it can provide for both the active and the standby assignments of each service instance protected by each service group. This can be accomplished by creating one service unit with the required number of components so that it provides the required active and standby capabilities for all the component service instances. Once all of the components of the first SU have been added, other SUs required by the redundancy model are copied from this first SU. If more than one service group is required for the SI template, the first service group is copied until all service instances can be served. This exemplary embodiment assumes equal ranking for all of the service units within each service group. However the present invention is not so limited and, for example, a ranking procedure can be added to dictate an ordered handling of service units within the service groups. Additionally, the entity types may be refined or optimized by tailoring the parameters of the type instances to their use based on the created entity instances.

[0189] Since the AMF specification does not currently cover performance characteristics for entities, this exemplary embodiment assumes that each service unit of the configuration produces equal load and that each cluster node provides the same capacity, although it will be appreciated by those skilled in the art that the present invention is not so limited. As a result this exemplary embodiment distributes the SUs among the nodes, at step 1706 in FIG. 16, of the cluster by selecting, for each SU, the node that has the smallest load in terms of number of service units assigned and that has no SU (or a smallest number of) associated with the same SG that the current SU belongs to. These conditions can be implemented by, for example, performing a global round robin distribution of the SUs as they are being created. Moreover, this exemplary distribution method can be further optimized or enhanced, e.g., based on performance characteristics of the system by assuming different capacities for the different nodes and different load for the different service units and or service instances and/or based on the assumption of different failure scenarios to avoid overload situations when one or more SUs/nodes fail in the system. These additional conditions will result in a non-round robin distribution of the SUs on the nodes. The process associated with steps 1702, 1704 and 1706 continues until all of the SI templates are processed and the configuration is populated with entities to provide their service instances in the SAF system.

[0190] As a result of applying the method illustrated in FIG. 16, and described above, to the inputs, an AMF configuration is generated by this exemplary embodiment that specifies all of the instances for all of the entity types and their entities. The result can then be stored or saved, e.g., in a memory device or recordable medium, in any desired format as indicated by step 1708. For example, this output can be presented and/or stored in accordance with the AIS specifications as an XML file created according to the IMM schema, which contains the description of the instances as specified in the UML model of AMF. In this way, an SAF compliant AMF implementation can be provided with the information model of all of the entities which it needs to manage as an input. Due to the potential complexity of AMF managed systems, this provides an automated approach to a complex and error prone task.

[0191] The following pseudo-code describes, at a high level, an exemplary implementation of the afore-described exemplary embodiment. Other implementations are also possible.

[0192] Component Service Types

A component service type is a set

• [0193] cst={att_i; 0<=i<|cst|, any number of attributes}.
  All component service types compose the set
• [0194] CST={cst_i; 0<=i<|CST|}.
  This set is a mandatory input and remains constant. It is provided typically by the software vendor.

[0195] Component Types

A component capability is a tuple of

• [0196] csCapability=<cst, compCap, maxAct, maxStdb>, where cst ε CST;

[0197] $\begin{array}{c}\mathrm{compCap}=\{\begin{array}{c}\mathrm{x\_active}\mathrm{\_and}\mathrm{\_y}\mathrm{\_standby}\\ \mathrm{x\_active}\mathrm{\_or}\mathrm{\_y}\mathrm{\_standby}\\ 1\mathrm{\_active}\mathrm{\_or}\mathrm{\_y}\mathrm{\_standby}\\ 1\mathrm{\_active}\mathrm{\_or}\_1\mathrm{\_standby}\\ \mathrm{x\_active}\\ 1\mathrm{\_active}\end{array}\\ \max \mathrm{Act}=\{\begin{array}{c}\mathrm{x\_active}\mathrm{\_and}\mathrm{\_y}\mathrm{\_standby}\Rightarrow x,x\ge 1\\ \mathrm{x\_active}\mathrm{\_or}\mathrm{\_y}\mathrm{\_standby}\Rightarrow x,x\ge 2\\ 1\mathrm{\_active}\mathrm{\_or}\mathrm{\_y}\mathrm{\_standby}\Rightarrow 1\\ 1\mathrm{\_active}\mathrm{\_or}\_1\mathrm{\_standby}\Rightarrow 1\\ \mathrm{x\_active}\Rightarrow x,x\ge 2\\ 1\mathrm{\_active}\Rightarrow 1\end{array}\\ \max \mathrm{Stdb}=\{\begin{array}{c}\mathrm{x\_active}\mathrm{\_and}\mathrm{\_y}\mathrm{\_standby}\Rightarrow y,y\ge 1\\ \mathrm{x\_active}\mathrm{\_or}\mathrm{\_y}\mathrm{\_standby}\Rightarrow y,y\ge 1\\ 1\mathrm{\_active}\mathrm{\_or}\mathrm{\_y}\mathrm{\_standby}\Rightarrow y,y\ge 2\\ 1\mathrm{\_active}\mathrm{\_or}\_1\mathrm{\_standby}\Rightarrow 1\\ \mathrm{x\_active}\Rightarrow 0\\ 1\mathrm{\_active}\Rightarrow 0\end{array}\end{array}$
A component type is a set

• [0198] ct={csCapability_i; 0<=i<|ct|, all the component service types provided by the component type ct}

[0199] Component Type Sets

All component types compose the set

• [0200] CT={ct_i; 0<=i<|CT|}.
  This set is a mandatory input and remains constant. It is provided typically by the software vendor.
  orphanCTs is the subset of component types that are not referenced by any SU type in the input set (orphanCTs ⊂ CT). Note that even when they are adopted by created service unit types, orphan component types remain in the set of orphanCTs

[0201] Service Types

A member CS type is a tuple

• [0202] memCst=<cst, maxCst>, where cst ε CST;
• [0203] maxCsi is the number of CSIs of type est that a service instance of a particular type may contain.
  A service type is a set
• [0204] st={memCst_i; 0<=i<|st| component service instances that compose the service type st}.
  All the service types compose the set
• [0205] ST={st_i; 0<=i<|ST|}.
  This set is a mandatory input and remains constant.

[0206] Service Unit Types

The services of a service unit type is a set

• [0207] sutServices={st_i; 0<=i<|sutServices|, all the service types the service unit provides and it is a subset of ST}
  A member component type is a tuple
• [0208] memCt=<ct, maxComp>where ct ε CT;
• [0209] maxComp is the maximum number of instances of this component type an SU of this type may contain.
  The component types of an SU type is a set
• [0210] sutCt={memCt_i; 0<=i<|sutCt|, all the component types the service unit may contain and |sutct|<=|CT|}.
  A service unit type is a tuple
• [0211] sut=<sutCt, sutServices>

[0212] Service Unit Type Sets

All the service unit type compose the set

• [0213] SUT={sut_i; 0<=i<|SUT|}
• [0214] This set may be empty (SUT={ }) at the beginning of the configuration generation.
  orphanSUTs is the subset of service unit types that are not referenced by any SG type in the input set (orphanSUTs ⊂ SUT). Note that even when they are adopted by created service group types, they remain in the set of orphanSUTs
  createdSUTs is initially an empty set (createdSUTs−{ }). It is the collection of service unit types that are created during the process of generating a configuration. It is also a subset of the service unit types set (createdSUTs ⊂ SUT).

[0215] Service Group Types

The service unit types of a service group type is a set

• [0216] sgtSut={sut_i; 0<=i<|sgtSut|, all the service unit types the service group may contain and it is a subset of SUT}
  A service group type is a tuple
• [0217] sgt=<sgtSut, redMod>, where

[0218] $\mathrm{redMod}=\{\begin{array}{c}2n\\ \mathrm{nplusm}\\ \mathrm{nway}\\ \mathrm{nwayactive}\\ \mathrm{noredundancy}\end{array}$

[0219] Service Group Type Sets

All service group types compose the set

• [0220] SGT={sgt_i; 0<=i<|SGT|}
  This set may be empty (SGT={ }) at the beginning of the configuration generation.
  orphanSGTs is a collection of service group types that are not referenced by any SU type in the input set (orphanSGTs ⊂ SGT). Note that even when they are adopted by created application types, they remain in the set of orphanSGTs.
  createdSGTs is initially an empty set (createdSGTs={ }). It is the collection of service group types that are created during the process of generating a configuration. It is also a subset of the service group types set (createdSGTs ⊂ SGT).

[0221] Application Types

An application type is a set described by

• [0222] appt—{sgt_i; 0<=i<|SGTs|; all the service group types the application type may contain}

[0223] Application Type Sets

All application types compose the set

• [0224] APPT={appt_i; 0<=i<|APPT|}.
  This set may be empty (APPT={ }) at the beginning of the configuration generation.
  createdAPPTs is initially an empty set (createdAPPTs={ }). It is the collection of application types that are created and added to this set during the process of the configuration generation. It is a subset of the application types set (createdAPPTs ⊂ APPT).

[0225] CSI Templates

A CSI template is a tuple

• [0226] csiTemp=<cst, numCsi>, where cst ε CST;
  numCsi is the number of CSIs of type cst created based on this template.
  All CSI templates compose the set
• [0227] CSIT={csiTemp_i; 0<=i<|CSIT|}.
  This set is a mandatory input and remains constant. It is provided by the site designer.

[0228] SI Templates

A SI template is a tuple

• [0229] siTemp=<st, sicsTemps, redMod, numSIs, numAct, numStdb, numSUs>, where st ε ST;
• [0230] sicsiTemps={esiTemp_i; 0<=i<|sicsiTemps| ∀csiTemp_i.cst≡sti_i.cst and csiTemp_i.numCsi≦st_i.maxCsi};
  numSIs is the number of SI that is created using this template

[0231] $\mathrm{numAct}=\{\begin{array}{c}\mathrm{redMod}\ne \mathrm{nwayactive}\Rightarrow 1\\ x;x\ge 2,\end{array}$
is the number of HA active assignments per SI;

[0232] $\mathrm{numStdb}=\{\begin{array}{c}\mathrm{redMod}\equiv \mathrm{nway}\Rightarrow y;y\ge 1\\ \mathrm{redMod}\equiv \mathrm{nwayactive}\Rightarrow 0\\ \mathrm{redMod}\equiv \mathrm{noredundancy}\Rightarrow 0\\ 1,\end{array}$
is the number of HA standby assignments per SI;

• [0233] numSUs=<sus, susAct, susStdt,susSpare>, where
  sus the total number of SUs in the SG protecting the SIs generated from this template; sus=susAct+susStdb+susSpare;

[0234] $\mathrm{numAct}=\{\begin{array}{c}\mathrm{redMod}\equiv \mathrm{nplusm}\Rightarrow z\\ \mathrm{redMod}\equiv \mathrm{nwayactive}\Rightarrow z;z>1\\ \mathrm{redMod}\equiv \mathrm{noredundancy}\Rightarrow \mathrm{numSIs}\\ \mathrm{redMod}\equiv 2n\Rightarrow 1\\ \mathrm{redMod}\equiv \mathrm{nway}\Rightarrow 0,\end{array}$
is the number of SUs that can have only active assignments;

[0235] $\mathrm{susStbd}\mathrm{numStdb}=\{\begin{array}{c}\mathrm{redMod}\equiv \mathrm{nplusm}\Rightarrow w\\ \mathrm{redMod}\equiv \mathrm{nwayactive}\Rightarrow 0\\ \mathrm{redMod}\equiv \mathrm{noredundancy}\Rightarrow 0\\ \mathrm{redMod}\equiv 2n\Rightarrow 1\\ \mathrm{redMod}\equiv \mathrm{nway}\Rightarrow 0\end{array}$
is the number of SUs that can have only standby assignments;

• [0236] susSpare is the number of spare SUs.
  All SI templates compose the set
• [0237] SIT={siTemp_i; 0<=i<|SIT|}.
  This set is a mandatory input and remains constant. It is provided by the site designer.

[0238] Node Templates

The node template is a tuple

• [0239] nodeTemp=<numNodes>, where
  numNodes is the number of nodes created based on the template.
  All node templates compose the set
• [0240] NodeTemp={nodeTemp_i; 0<=i<|NodeTemp|}.
  This set is a mandatory input and remains constant. It is provided by the site designer.

[0241] Outputs to Generate

[0242] Components

A component has a name and belongs to component type. It is a tuple comp=<safComp, ct>, where ct ε CT;

• [0243] safComp is a tuple safComp=<safSu, rdn>, where
• [0244] safSu is a SAF compliant LDAP distinguished name of the SU the component belongs to and rdn is a relative distinguished name of the component, which is unique within the SU.

[0245] Service Units

A service unit is a tuple

• [0246] su=<safSu, sut, memComps>, where
• [0247] safSu is a tuple safSu=<safSg, rdn>, where
• [0248] safSg is a SAF compliant LDAP distinguished name of the SG the service unit belongs to and rdn is a relative distinguished name of the SU which is unique within the SG;
• [0249] sut ε SUT;
• [0250] memComps={memComp_i; 0<=i<|memComps|, ∀memComp.safComp.safSu≡safSu;}
• [0251] memComp={ } at the start of the configuration generation.

[0252] Service Groups

A service group is a tuple

• [0253] sg=<safSg, sgt, memSus, nodeGroup>, where
• [0254] safSg is a tuple safSg=<safApp, rdn>, where
• [0255] safApp is a SAF compliant LDAP distinguished name of the application the service group belongs to and
• [0256] rdn is a relative distinguished name of the SG, which is unique within the application; sgt ε SGT;
• [0257] memSus={memSu_i; 0<=i<|memSus|, memSu_i.sut ε sgt.sgtSut and ∀memSu.safSu.safSg≡safSg;}
• [0258] memSUs—{ } at the start of the configuration generation
• [0259] nodeGroup={node_i; 0<=i<|nodeGroup|,}; the nodeGroup ┤ Cluster on which the SG is distributed

[0260] Component Service Instances

A component service instance is a tuple

• [0261] csi=<safCsi, cst, attrValues>, where
• [0262] safCsi is a tuple safCsi=<safSi, rdn>, where
• [0263] safSi is a SAF compliant LDAP distinguished name of the service instance the component service instance belongs to and
• [0264] rdn is a relative distinguished name of the CSI, which is unique within the service instance; cst ε CST;
• [0265] attrValues={attrValue_i; 0<=i<|cst|}
• [0266] attrValue=<attr, value>, where
• [0267] ∀attrValue.attr ε cst.attr;
• [0268] attrValue.value is a value set for the attribute
• [0269] attrValues={ } the method does not cover the generation of attribute values. This is a potential future enhancement.

[0270] Service Instances

A service instance is a tuple

• [0271] si=<safSi, st, sg, memCsis>, where
• [0272] safSi is a tuple safSi=<safApp, rdn>, where
• [0273] safApp is a SAF compliant LDAP distinguished name of the application the service instance belongs to and
• [0274] rdn is a relative distinguished name of the SI, which is unique within the application; st ε ST;
• [0275] memCsis={memCsi_i; 0<=i<|memCsis|, ∀memCsi.cst∃st.cst; and ∀memCsi.safCsi.safSi≡safSi;}
• [0276] memCsis={ } at the start of the configuration generation
• [0277] sgSis={si_i; 0<=i<|sgSis|, ∀si_i.sg≡sg} the set of SIs protected by the same service group, they specify the same service group.

[0278] Applications

The application for which the configuration is generated is a tuple

• [0279] app=<safApp, memSgs, memSis>, where
• [0280] safApp is a distinguished name of the application;
• [0281] memSgs={memSg_i; 0<=i<|memSgs|, ∀memSg.safSg.safApp≡safApp;} memSgs={ } at the start of the configuration generation.
• [0282] memSis={memSi_i; 0<=i<|memSis|, ∀memSi.sajSi.safApp≡safApp;}
• [0283] memSis={ } at the start of the configuration generation.
  All applications compose the set
• [0284] Apps={app_i; 0<=i<|Apps|}

[0285] Nodes

• [0286] node=<safAmfNode, hostedSus>, where
• [0287] safAmfNode is a distinguished name of the node;
• [0288] hostedSus={hostedSu_i; 0<=i<|hostedSus|}
• [0289] hostedSus={ } at the start of the configuration generation.
  All nodes compose the cluster:
• [0290] Cluster={node_i; 0<=i<|Cluster|}

[0291] Configuration Generation

[0292] calculateSuActLoad Method

Input

• • [0293] siTemp, an SI template
    Output
  • [0294] suActCap, the required active capability for the SU with respect to the SI
    Functional Description

[0295] The required SU active capabilities are calculated with respect to the SI template with the assumption that an SU provides a group of SIs from a single template. In the formula ceil( ) is used to round up to the nearest integer.

[0296] $\mathrm{return}$ $\mathrm{suActLoad}=\{\begin{array}{c}\mathrm{redMod}\equiv \mathrm{nway}\Rightarrow \mathrm{ceil}\left(\frac{\mathrm{siTemp}.\mathrm{numSIs}}{\begin{array}{c}\mathrm{siTemp}.\mathrm{numSUs}.\mathrm{sus}-\\ \{\begin{array}{c}\mathrm{siTemp}.\mathrm{numSUs}.\mathrm{susSpare}\equiv 0\Rightarrow 1\\ \Rightarrow \mathrm{siTemp}.\mathrm{numSUs}.\mathrm{susSpare}\end{array}\end{array}}\right)\\ \mathrm{redMod}\equiv \mathrm{nwayactive}\Rightarrow \mathrm{ceil}\left(\frac{\mathrm{siTemp}.\mathrm{numSIs}\times \mathrm{siTemp}.\mathrm{numAct}}{\mathrm{siTemp}.\mathrm{numSUs}.\mathrm{susAct}}\right)\\ \mathrm{redMod}\equiv \mathrm{noredundancy}\Rightarrow 1\\ \mathrm{redMod}\equiv 2n\Rightarrow \mathrm{siTemp}.\mathrm{numSIs}\\ \mathrm{redMod}\equiv \mathrm{nplusm}\Rightarrow \mathrm{ceil}\left(\frac{\mathrm{siTemp}.\mathrm{numSIs}}{\mathrm{siTemp}.\mathrm{numSUs}.\mathrm{susAct}}\right)\end{array}$

[0297] calculateSuStdbLoad Method

Input

• • [0298] siTemp, an SI template
    Output
  • [0299] suStdbCap, the required standby capability for the SU with respect to the SI
    Function Description

[0300] The required SU standby capabilities are calculated with respect to the SI template with the assumption that an SU provides a group of SIs from a single template. In the formula ceil( ) is used to round up to the nearest integer.

[0301] $\mathrm{return}$ $\mathrm{suStdbLoad}=\{\begin{array}{c}\mathrm{redMod}\equiv \mathrm{nway}\Rightarrow \mathrm{ceil}\left(\frac{\begin{array}{c}\mathrm{siTemp}.\mathrm{numSIs}\times \\ \mathrm{siTemp}.\mathrm{numStdb}\end{array}}{\begin{array}{c}\mathrm{siTemp}.\mathrm{numSUs}.\mathrm{sus}-\\ \{\begin{array}{c}\mathrm{siTemp}.\mathrm{numSUs}.\mathrm{susSpare}\equiv 0\Rightarrow 1\\ \Rightarrow \mathrm{siTemp}.\mathrm{numSUs}.\mathrm{susSpare}\end{array}\end{array}}\right)\\ \mathrm{redMod}\equiv \mathrm{nwayactive}\Rightarrow 0\\ \mathrm{redMod}\equiv \mathrm{noredundancy}\Rightarrow 0\\ \mathrm{redMod}\equiv 2n\Rightarrow \mathrm{siTemp}.\mathrm{numSIs}\\ \mathrm{redMod}\equiv \mathrm{nplusm}\Rightarrow \mathrm{ceil}\left(\frac{\mathrm{siTemp}.\mathrm{numSIs}}{\mathrm{siTemp}.\mathrm{numSUs}.\mathrm{susStdb}}\right)\end{array}$

[0302] createAppType Method

Input

• • [0303] sgt, an service group type
    Output
  • [0304] appt, an application type
    Function Description

[0305] This function creates an application type that supports a given service group type.

Begin

[0306] Create a new application type appt={sgt}

[0307] ADD appt TO createdAPPTs

End

[0308] createSgType Method

Input

• • [0309] sut, an service unit type
  • [0310] siTemp.redMod, the redundancy model required by the SI template
    Output
  • [0311] sgt, a service group type
    Function Description

[0312] This function creates a service group type that supports a given service unit type, and sets the redundancy model of this service group to same one of the SI template

[0313]

	BEGIN
	Create a new service group type sgt = <sgtSut, redMod>
	ADD sut TO sgt.sgtSut
	SET sgt.redMod TO siTemp.redMod
	ADD sgt TO createdSGTs
	END

[0314] createSuType Method

Input

• • [0315] siTemp, an SI template
    Output
  • [0316] sut, a service unit type
    Function Description

[0317] This function creates a service unit type that supports a given service type.

[0318]

	BEGIN
	Create a new service unit sut=<sutCt, sutServices>
	ADD siTemp.st TO sut.sutServices
	FOR every csiTemp IN siTemp. sicsTemps
	CALL findOrphanCT WITH siTemp.csiTemp, siTemp
	RETURNING foundCT //foundCT is the found
	component type
	IF foundCT IS Null THEN
	SET sut TO Null // Delete the created SU Type
	BREAK
	ELSE
	SET memCT.ct TO foundCT
	SET memCT.maxComp TO no limit
	ADD memCT TO sut.sutCt
	ADD sut TO createdSUTs
	ENDIF
	ENDFOR
	RETURN sut
	END

[0319] findAppt Method

Input

• • [0320] givenSgt, service group type
    Output
  • [0321] appt an application type
    Function Description

[0322] This function searches for the first created application type that can support the given SG type. This function is NOT used for finding application types that are not created Modifications to this function are possible based of preferences regarding the extensibility of created application types

[0323]

	BEGIN
	BOOLEAN found = false
	FOR every appt IN createdAPPT
	FOR every sgt IN appt
	IF sgt = givenSgt THEN
	RETURN appt
	SET found TO true
	BREAK out of createdAPPT
	ENDIF
	ENDFOR
	ENDFOR
	IF found = false THEN
	RETURN Null
	ENDIF
	END

[0324] findSgt Method

Input

• • [0325] givenSut service unit type
  • [0326] redMod, a redundancy model
    Output
  • [0327] sgt, a service group type
    Function Description

[0328] This function searches for the first created SG type that can support the given SU type. This function is NOT used for finding SG types that are not created.

[0329]

	BEGIN
	BOOLEAN found = false
	FOR every sgt IN createdSGTs
	IF sgt.redMod = redMod THEN
	FOR every sut IN sgt.sgtSut
	IF sut = givenSut THEN
	RETURN sgt
	SET found TO true
	BREAK out of createdSGTs
	ENDIF
	ENDFOR
	ENDIF
	ENDFOR
	if found = false THEN
	RETURN Null
	END

[0330] findSut Method

Input

• • [0331] siTemp, an SI template
  • [0332] sutSet, a set of SU types ⊂ SUT// we could be looking in orphanSUTs or sgt.sgtSut or createdSUTs
    Output
  • [0333] sut, a service unit type
    Function Description

[0334] This function searches for the first SU type that can support the SIs of the given template siTemp. Optimization may be added depending on other requirements.

[0335]

BEGIN
BOOLEAN found = false
FOR every sut IN sutSet
FOR every st IN sut.Services
IF st = siTemp.st THEN
FOR every csiTemp in siTemp.sicsTemps
SET found TO false
CALL findCT WITH siTemp,csiTemp,sut RETURNING foundCt
IF foundCt = Null
THEN //this sut is not valid
CONTINUE and GO TO next sut in the sutSet
ELSE
SET found TO true
ENDIF
ENDFOR
IF found = true THEN
RETURN sut //the sut in the current
sutSet iteration
BREAK from the sutSet loop
ENDIF
ENDIF
ENDFOR
ENDFOR
IF found = false THEN//all the su types are not valid
RETURN Null
ENDIF
END

[0336] findCt Method

Input

• • [0337] siTemp, an SI template
  • [0338] csiTemp, a CSI template
  • [0339] sut, a service unit type
    Output
  • [0340] ct, a component type
    Function Description

[0341] This function searches for the first component type that can support the CSIs of the given template csiTemp. Optimization may be added depending on other requirements.

[0342]

BEGIN
BOOLEAN found = false
FOR every memCt IN sutCt
IF memCt.ct.csCapability.cst = csiTemp.cst THEN
CASE siTemp.redMod OF
nway:
IF memCt.ct.csCapability.compCap = x_active_and_y_standby THEN
CALL calculateSuActLoad WITH siTemp RETURNING suActLoad
CALL calculateSuStandbyLoad WITH siTemp RETURNING
suStandby
IF memCt.maxComp * memCt.ct.csCapability.maxActive >
csiTemp.numCsi * suActLoad
AND
memCt.maxComp * memCt.ct.csCapability.maxStandby >
csiTemp.numCsi * suStandbyLoad THEN
SET found TO true
RETURN memCt.ct
BREAK from the sutCt loop
ELSE
GO to next memCt
ENDIF
ENDIF
OTHERS: //in case of other redundancy models
CALL calculateSuActLoad WITH siTemp RETURNING suActLoad
CALL calculateSuStandbyLoad WITH siTemp RETURNING suStdbLoad
IF memCt.maxComp * memCt.ct.csCapability.maxActive >
csiTemp.numCsi * suActLoad
AND
memCt.maxComp * memCt.ct.csCapability.maxStdby >
csiTemp.numCsi * suStdbLoad THEN
found = true
RETURN memCt.ct
BREAK from the sutCt loop
ELSE
GO to next memCt
ENDIF
ENDCASE
ENDIF
ENDFOR
IF found = false THEN
RETURN Null
ENDIF
END

[0343] findOrphanCt Method

Input

• • [0344] siTemp, an SI template
  • [0345] csiTemp, a CSI template
    Output
  • [0346] ct, a component type
    Function Description

[0347] This function searches for the first component type in orphanCTS that can support the CSIs of the given template csiTemp. Optimization may be added depending on other requirements.

[0348]

BEGIN
BOOLEAN found = false
For every ct IN orphanCTS
CASE siTemp.redMod OF
nway:
IF ct.csCapability.compCap IS NOT x_active_and_y_standby THEN
CONTUNUE and GO TO next ct in orphanCTs
ELSEIF ct.csCapability.cst = csiTemp.cst THEN
RETURN ct
SET found TO true
BREAK out of orphanCTs
ENDIF
OTHERS:
IF ct.csCapability.cst = csiTemp.cst THEN
RETURN ct
SET found TO true
BREAK out of orphanCTs
ENDCASE
ENDFOR
IF found = false
RETURN Null
ENDIF
END

[0349] calculateNumOfComp Method

Input.

• • [0350] ct, a component type
  • [0351] siTemp, an SI template
  • [0352] csiTemp, an CSI template
    Output
  • [0353] numOfComp, an integer reflecting the number of component
    Function Description

[0354] This function calculates the number of components of a given component type that need to be created inside a particular service unit

[0355]

BEGIN
FOR every csCapability IN ct
IF csCapability.cst = csiTemp.cst THEN
CALL calculateSuActLoad WITH siTemp
RETURNING suActLoad
CALL calculateSuStandbyLoad WITH siTemp
RETURNING suStdbLoad
numOfComp = CEIL ( MAX ( (csiTemp.numCsi *
suActLoad )/csCapability.maxAct,
(csiTemp.numCsi *
suStdbLoad )/
csCapability.maxAct))
RETURN numOfComp
BREAK out of ct loop
ENDIF
ENDFOR
END

[0356] assigntoNode Method

Input

• • [0357] su, an service unit
  • [0358] nodeGroup, a node group
    Output
  • [0359] No output
    Function Description

[0360] This function will assign the given service unit to a node that have the smallest number of service units assigned to it.

[0361]

BEGIN
BOOLEAN assigned = false
FOR every node IN the cluster
IF |node.hostedSus| < |firstnode.hostedSus|
THEN // first node is the first node in the cluster
ADD su TO node.Sus
ADD node TO nodeGroup
SET assigned TO true
BREAK out of cluster
ENDIF
ENDFOR
IF assigned = false THEN
ADD su TO firstnode.Sus
ADD first node TO nodeGroup
ENDIF

[0362] generateConf Method

Input

• • [0363] CT all component types
  • [0364] CST all component service types
  • [0365] ST all service types
  • [0366] SIT all service instance templates
  • [0367] CSIT all component service instance templates
  • [0368] SUT service unit types
  • [0369] SGT service group types
  • [0370] APPT application types
  • [0371] nodeTemp the node template
  • [0372] userPerference is a set of preferences that guides the configuration generation. They allow for different options and optimization and can be extended accordingly. The pseudo-code does not reflect them.
    Output
  • [0373] Apps, i.e. AMF entities for all applications distributed on the nodes of a cluster. Note that the sets of types provided as input are also updated with the created entity types.
    Function Description

[0374] This is the main function that is responsible for creating all the entities in AMF configuration, and populate their attributes.

[0375]

BEGIN
BOOLEAN foundSut = false
FOR every siTemp IN SIT
foundSut = false
FOR every appt IN APPT
FOR every sgt IN appt.sgt
IF sgt.redMod = siTemp.redMod THEN
CALL findSut WITH siTemp, sgt.sgtSut RETURNING sut
IF sut IS NOT Null THEN
SET foundSut TO true
Create new application app = <”safApp=...”, appt, memSgs, memSis>
Create new service group sg=< ”safSg=... ”, sgt, memSus,
nodeGroup>
ADD sg TO app.memSgs
Create new service unit su = <”safSu=... ”, sut, memComps>
ADD su TO sg.memSus
FOR every siTemp.csiTemp in siTemp.sicsTemps
CALL findCt WITH sut, siTemp, csiTemp RETURNING ct
CALL calculateNumOfComp WITH ct, siTemp, csiTemp
RETURNING numOfComp
FOR numOfComp
Create component comp = < ”safComp=... ”, ct >
ADD comp TO su.memComps
ENDFOR
ENDFOR
CALL assigntoNode WITH su, sg.nodeGroup
FOR siTemp.numSUs−1//cloning the created su.
Create new service unit newSu = <”safSu=... ”, sut,
memComps>
FOR every comp IN su.memComps
create component newComp
< ”safComp=... ”,comp.ct>
ADD newComp TO newSu.memComps
ENDFOR
CALL assigntoNode WITH newSu, sg.nodeGroup
ADD newSu TO sg.memSus
ENDFOR
ADD app TO Apps
CONTINUE and GO to next SI template
ENDIF
ENDIF
ENDFOR
ENDFOR
IF foundSu = false THEN //no application type was found
FOR every sgt IN orphanSGT
IF sgt.redMod = siTemp.redMod THEN
CALL findSut WITH siTemp, sgt.sgtSut RETURNING sut
IF sut IS NOT Null THEN
SET foundSut TO true
CALL findAppt WITH sgt RETURNING appt
IF appt = Null THEN
CALL createApptype WITH sgt RETURNING appt
ENDIF
Create new application app = <”safApp=...”, appt, memSgs, memSis>
Create new service group sg=< ”safSg=... ”, sgt, memSus,
nodeGroup>
ADD sg TO app.memSgs
Create new service unit su = <”safSu=... ”, sut, memComps>
ADD su TO sg.memSus
FOR every siTemp.csiTemp in siTemp.sicsTemps
CALL findCt WITH sut, siTemp, csiTemp RETURNING ct
calculateNumOfComp WITH ct, siTemp, csiTemp
RETURNING numOfComp
FOR numOfComp
Create component comp = < ”safComp=... ”, ct >
ADD comp TO su.memComps
ENDFOR
ENDFOR
CALL assigntoNode WITH su, sg.nodeGroup
FOR siTemp.numSUs−1//cloning the created su.
Create new service unit newSu = <”safSu=... ”, sut,
memComps>
FOR every comp IN su.memComps
create component newComp
< ”safComp=... ”,comp.ct>
ADD newComp TO newSu.memComps
ENDFOR
CALL assigntoNode WITH newSu, sg.nodeGroup
ADD newSu TO sg.memSus
ENDFOR
ADD app TO Apps
CONTINUE and GO to next SI template
ENDIF
ENDIF
ENDFOR
ENDIF
IF foundSu = false THEN //no sg type was found
CALL findSut WITH siTemp, orphanSUTs RETURNING sut
IF sut IS NOT Null THEN
SET foundSut TO true
CALL findSgt WITH sut, siTemp.redMod RETURNING sgt
IF sgt = Null THEN
CALL createSgtype WITH sut, siTemp.redMod RETURNING sgt
ENDIF
CALL findAppt WITH sgt RETURNING appt
IF appt = Null THEN
CALL createApptype WITH sgt RETURNING appt
ENDIF
Create new application app = <”safApp=...”, appt, memSgs, memSis>
Create new service group sg=< ”safSg=... ”, sgt, memSus,
nodeGroup>
ADD sg TO app.memSgs
Create new service unit su = <”safSu=... ”, sut, memComps>
ADD su TO sg.memSus
FOR every siTemp.csiTemp in siTemp.sicsTemps
CALL findCt WITH sut, siTemp, csiTemp RETURNING ct
CALL calculateNumOfComp WITH ct, siTemp, csiTemp
RETURNING numOfComp
FOR numOfComp
Create component comp = < ”safComp=... ”, ct >
ADD comp TO su.memComps
ENDFOR
ENDFOR
CALL assigntoNode WITH su, sg.nodeGroup
FOR siTemp.numSUs−1//cloning the created su.
Create new service unit newSu = <”safSu=... ”, sut, memComps>
FOR every comp IN su.memComps
create component newComp < ”safComp=... ”,comp.ct>
ADD newComp TO newSu.memComps
ENDFOR
CALL assigntoNode WITH newSu, sg.nodeGroup
ADD newSu TO sg.memSus
ENDFOR
ADD app TO Apps
CONTINUE and GO to next SI template
ENDIF
ENDIF
IF foundSu = false THEN //no su type was found
CALL findSut WITH siTemp, createdSUTs RETURNING sut
IF sut = Null THEN
CALL createSutype WITH siTemp RETURNING sut
ENDIF
IF sut IS NOT Null THEN
SET foundSut TO true
CALL findSgt WITH sut, siTemp.redMod RETURNING sgt
IF sgt = Null THEN
CALL createSgtype WITH sut, siTemp.redMod RETURNING sgt
ENDIF
CALL findAppt WITH sgt RETURNING appt
IF appt = Null THEN
CALL createApptype WITH sgt RETURNING appt
ENDIF
Create new application app = <”safApp=...”, appt, memSgs, memSis>
Create new service group sg=< ”safSg=... ”, sgt, memSus, nodeGroup>
ADD sg TO app.memSgs
Create new service unit su = <”safSu=... ”, sut, memComps>
ADD su TO sg.memSus
FOR every siTemp.csiTemp in siTemp.sicsTemps
CALL findCt WITH sut, siTemp, csiTemp RETURNING ct
CALL calculateNumOfComp   WITH ct, siTemp, csiTemp
RETURNING numOfComp
FOR numOfComp
Create component comp = < ”safComp=... ”, ct >
ADD comp TO su.memComps
ENDFOR
ENDFOR
CALL assigntoNode WITH su, sg.nodeGroup
FOR siTemp.numSUs−1//cloning the created su.
Create new service unit newSu = <”safSu=... ”, sut, memComps>
FOR every comp IN su.memComps
create component newComp < ”safComp=... ”,comp.ct>
ADD newComp TO newSu.memComps
ENDFOR
CALL assigntoNode WITH newSu, sg.nodeGroup
ADD newSu TO sg.memSus
ENDFOR
ADD app TO Apps
CONTINUE and GO to next SI template
ENDIF
ENDIF
IF foundSu = false THEN// no types were found to support the si template
MESSAGEBOX (“Sorry no types were found to support the current siTemplate”)
GO TO next SI template
ENDIF
ENDFOR
ADD createdAPPTs TO APPT
ADD createdSGTs TO SGT
ADD createdSUTs TO SUT
END

[0376] Note that although the foregoing, exemplary pseudo-code performs evaluations between SI requirements and type capabilities in a strict sense, i.e., checking to see if each value in type's capabilities tuple or set is equal to a corresponding value in the SI requirements tuple or set to determine if a type supports those requirements, other evaluation metrics can be used to determine the suitability of a type. For example, for created application types an AMF configuration generator according to exemplary embodiments may decide that it is sufficient that the required set of SG Type is present as a subset. This allows for the extension of existing created types. As a result the configuration will have different numbers of application types and therefore different number of applications, which reflects different strategies of grouping, e.g., grouping by functionality. Strict equivalence results in a 1:1 relation between created SU types, created SG types and created application types. Allowing subsets without other strategies results in a single APPT as it was done in the bottom-up approach.

[0377] The foregoing description of exemplary embodiments of the present invention provides illustration and description, but it is not intended to be exhaustive or to limit the invention to the precise form disclosed. Modifications and variations are possible in light of the above teachings or may be acquired from practice of the invention. The following claims and their equivalents define the scope of the invention.