System for Combining a Global Object Identifier With a Local Object Address in a Single Object Pointer

1. A method of moving objects from one process to another within a computer network having one or more computer systems and distributed memory in which data objects are stored, the processes each being associated with at least one of a plurality of virtual address spaces in the distributed memory and creating a plurality of objects, wherein each created object is identified by an object identifier that uniquely identifies the object across the virtual address spaces, and is identified by an object address that uniquely identifies the object within any one of the virtual address spaces, the method comprising the steps of: (a) combining a first object identifier, comprising the concatenation of two parts I and J, and a first object address in a first virtual address space comprising the concatenation of two parts A.sub.0 and B.sub.0, that identify an object to be moved from the first virtual address space into a second virtual address space, to thereby produce a base object pointer comprising the concatenation of a first pointer part, represented as P.sub.0, that is equal to the first identifier part I of the first object identifier (P.sub.0 =I), a second pointer part, represented as Q.sub.0, that is equal to the first address part A.sub.0 of the first object address (Q.sub.0 =A.sub.0), and a third pointer part, represented as R.sub.0, that is equal to the second identifier part of the first object identifier (R.sub.0 =J=B.sub.0); (b) associating the base object pointer (P.sub.0 Q.sub.0 R.sub.0) with the object to be moved from the first virtual address space in the distributed memory; (c) selecting a second object address in the second virtual address space such that the second object address is logically divided into a first part represented as A.sub.1, and a second part represented as B.sub.1, such that the second part of the second object address (B.sub.1) is equal to the second address part of the first object address (B.sub.1 =B.sub.0 =J); (d) copying the object to be moved into the second virtual address space of the distributed memory at the second object address (A.sub.1 B.sub.1), if the object is not already present in the second virtual address space; and (e) associating a second object pointer (P.sub.1 Q.sub.1 R.sub.1) comprising the concatenation of a first part, represented as P.sub.1, a second part, represented as Q.sub.1, and a third part, represented as R.sub.1, with the copied object in the distributed memory such that the first part of the second object pointer P.sub.1 is equal to the first identifier part of the base object pointer (P.sub.1 =P.sub.0 =I), the second part of the second object pointer Q.sub.1 is equal to the first part of the second object address (Q.sub.1 =A.sub.1), and the third part of the second object pointer R.sub.1 is equal to the third part of the base object pointer (R.sub.1 =R.sub.0 =J).

2. A method as defined in claim 1, wherein each process further includes, in its respective associated virtual address space, an object mapping table having an entry comprising a first part (P.sub.i) and a second part (Q.sub.i) for each base object pointer created by the first process; and wherein the selecting step (c) comprises the steps of: (c.1) selecting a second object address (A.sub.1 B.sub.1) such that A.sub.1 =Q.sub.i if there exists an entry (P.sub.i Q.sub.i) in the object mapping table such that P.sub.i =P.sub.0 ; and otherwise (c.2) adding a new entry (P.sub.1 Q.sub.1) to the object mapping table such that P.sub.1 =P.sub.0 and Q.sub. =A.sub.1.

3. A method as defined in claim 2, wherein the distributed memory further includes a plurality of networked objects, each including one or more internal object pointers comprising the concatenation of four parts S, P, Q, and R referring to other networked objects; and wherein the step of copying further comprises: (d.1) replacing the fourth part (S) of each internal object pointer (SPQR) with a first predetermined value that denotes a pending relocation of the networked object to a new process; and the method further comprises the step of: (f) for each internal object pointer (SPQR), responsive to the fourth part (S) of each internal object pointer having the first predetermined value, performing the steps of (f.1) selecting a new object address, comprising two parts A and B, in the second virtual address space such that B=R, and (f.2) replacing each internal object pointer (SPQR) with a new internal object pointer comprising four parts S.sub.n, P.sub.n, Q.sub.n, and R.sub.n such that P.sub.n =P, Q.sub.n =A, R.sub.n =R, and the fourth part S.sub.n of the internal object pointer equals a second predetermined value that denotes completion of the relocation of each object.

4. A method as defined in claim 3, wherein the selecting step (f.1) comprises the steps of: (f.1.1) selecting the new object address (AB) such that A=Q.sub.i if there exists an entry (P.sub.i Q.sub.i) in the object mapping table such that P.sub.i =P; and otherwise (f.1.2) adding a new entry (P.sub.n Q.sub.n) to the mapping table such that P.sub.n =P and Q.sub.n =A.

5. A method as defined in claim 7, wherein the performing step (f) is executed asynchronously with the replacing step (e).

6. A method as defined in claim 3, wherein the performing step (f) is executed asynchronously with the replacing step (e).

7. A method as defined in claim 1, wherein the distributed memory further includes a plurality of networked objects each including one or more internal object pointers, represented as the concatenation of four parts S, P, Q, and R, referring to other networked objects, and the step of copying further comprises: (d.1) replacing the fourth part (S) of each internal object pointer (SPQR) with a first predetermined value that denotes a pending relocation of each object to a new process; and the method further comprises the steps of: (f) for each internal object pointer (SPQR), responsive to the fourth part (S) having the first predetermined value, performing the steps of (f.1) selecting a new object address (AB) in the second virtual address space such that B=R, and (f.2) replacing each internal object pointer (SPQR) with a new internal object pointer comprising four parts S.sub.n, P.sub.n, Q.sub.n, and R.sub.n such that P.sub.n =P, Q.sub.n =A, R.sub.n =R and the fourth part S.sub.n equals a second predetermined value that denotes completion of the relocation of each object.

8. A method as defined in claim 7, wherein the performing step (f) is executed asynchronously with the replacing step (e).

9. In a multi-tasking network of one or more computer systems with distributed memory in which are stored data and data structures, including a plurality of objects each created by one of a plurality of concurrent processes, each of the processes being associated with at least one of a plurality of virtual address spaces in the distributed memory, each object being stored in one of the associated virtual address spaces at a local virtual object address represented as the concatenation of a first part A and a second part B, each object including a global object identifier that uniquely identifies the object across the virtual address spaces, the global object identifier comprising the concatenation of a first part I and a second part J, a data structure for denoting the global object identifier (IJ) and the local virtual object address (AB) of each object, said distributed memory having a data structure comprising: a first part P representing the first part (I) of the global object identifier; a second part Q representing the first part (A) of the local virtual object address; and a third part R representing the second part (J) of the global object identifier, wherein the first and third parts (IJ) are held invariant over the network and J=B.

10. In a multi-tasking network of one or more computer systems with distributed memory in which are stored data and data structures, including a plurality of objects each created by one of a plurality of concurrent processes, each of the processes being associated with at least one of a plurality of virtual address spaces in the distributed memory, each object being stored upon creation in an associated virtual address space at a base object address comprising the concatenation of a first part A.sub.0 and a second part B.sub.0, a system for moving each object from a first process associated with a first virtual address space to a second process associated with a second virtual address space, the system comprising: a pointer generator coupled to the associated virtual address space in the distributed memory that combines an object identifier comprising two parts I and J with the base object address (A.sub.0 B.sub.0), and thereby produces a base object pointer (P.sub.0 Q.sub.0 R.sub.0) logically divided into a first part (P.sub.0 =I), a second part (Q.sub.0 =A.sub.0), and a third part (R.sub.0 =J=B.sub.0); first linking means, coupled to the pointer generator in the distributed memory, for associating the base object pointer (P.sub.0 Q.sub.0 R.sub.0) with the object to be moved; an address generator coupled to the second process in the distributed memory for selecting a first object address (A.sub.1 B.sub.1) in the second virtual address space such that the first object address can be logically divided into a first part (A.sub.1) and a second part (B.sub.1) such that B.sub.1 =B.sub.0 =J; copying means coupled to the second process in the distributed memory for copying each object to produce a copied object at the second object address (A.sub.1 B.sub.1) in the second virtual address space in the distributed memory; and second linking means, coupled to the copying means in the distributed memory, for associating a second object pointer (P.sub.1 Q.sub.1 R.sub.1) with the copied object in the second virtual address space such that P.sub.1 =P.sub.0 =I, Q.sub.1 =A.sub.1 and R.sub.1 =R.sub.0 =J.

11. A system as defined in claim 10, further comprising: an object mapping table coupled to each process in the associated virtual address spaces for storing an entry (P.sub.i Q.sub.i) representing the first part (P.sub.i) and the second part (Q.sub.i) for each object pointer created by each process; entry restoring means coupled to the object mapping table in each associated virtual address space for selecting the second object address (A.sub.1 B.sub.1) such that A.sub.1 =Q.sub.i if there exists an entry (P.sub.i Q.sub.i) in the object mapping table such that P.sub.i =P.sub.0 ; and entry creating means coupled to each object mapping table in each associated virtual address space for adding a new entry (P.sub.1 Q.sub.1) to each mapping table such that P.sub.1 =P.sub.0 and Q.sub.1 =A.sub.1.

12. A system as defined in claim 11, wherein the distributed memory further includes a plurality of networked objects each including one or more internal object pointers that refer to other network objects, each internal object pointer represented as the concatenation of four parts S, P, Q, and R, the system further comprising: flagging means, coupled to each process in the distributed memory, for replacing the fourth part (S) of each internal object pointer (SPQR) with a first predetermined value that denotes a pending relocation of each object to a new process; address selecting means, coupled to each process in the distributed memory, for selecting a new object address (AB) in the second virtual address space such that B=R for each internal object pointer (SPQR); and updating means, coupled to the address selecting means, for replacing each internal object pointer (SPQR) with a new internal object pointer (S.sub.n P.sub.n Q.sub.n R.sub.n) such that P.sub.n =P, Q.sub.n =A, R.sub.n =R and the fourth part S.sub.n equals a second predetermined value denoting completion of relocation of each object to the second virtual address space.

13. A system as defined in claim 10, wherein the distributed memory further includes a plurality of networked objects each including one or more internal object pointers that refer to other network objects, each internal object pointer comprising the concatenation of four parts S, P, Q, and R, the system further comprising: flagging means, coupled to each process in the distributed memory, for replacing the fourth part (S) of each internal object pointer (SPQR) with a first predetermined value that denotes a pending relocation of each object to a new process; address selecting means, coupled to each process in the distributed memory, for selecting a new object address (AB) in the second virtual address space such that B=R for each internal object pointer (SPQR); and updating means, coupled to the address selecting means, for replacing each internal object pointer (SPQR) with a new internal object pointer (S.sub.n P.sub.n Q.sub.n R.sub.n) such that P.sub.n =P, Q.sub.n =A, R.sub.n =R and the fourth part S.sub.n equals a second predetermined value denoting completion of relocation of each object to the second virtual address space.