The relevance of object groups and multicast shared distributed object systems

ABSTRACT

We argue that object groups and multicast invocations are very useful in the construction of an object based platform for building multi-user applications. We illustrate our argument with reference to three aspects of our design. The first aspect relates to the issue of informing users when other users have altered shared objects. The second aspect concerns the design of an optimistic form of concurrency control for replicas of shared objects. The third aspect is a suggested solution to the problem of distributing capabilities to groups of users. Our conclusion is that, although ordered multicast is essential for some purposes, reliable and even unreliable multicast are also very useful in contexts requiring object group invocations.

REFERENCES

• 1: J. Dollimore , E. Miranda , Wang Xu, The design of a system for distributing shared objects, The Computer Journal, v.34 n.6, p.514-521, Dec. 1991 [doi>10.1093/comjnl/34.6.514]
• 2: {Dollimore and Wang Xu, 1991} Jean Dollimore and Wang Xu. The Private Access Channel a mechanism for controlling access to shared distributed objects; January 1991, revised March 1992. QMW CSL Report No 537.
• 3: {Min and Dollimore 1992} Sung-Gi Min and Jean Dollimore. A conflict based Optimistic Concurrency Control Protocol. April 1992. QMW CSL Report No. 588.
• 4: Eliot Miranda, BrouHaHa- A portable Smalltalk interpreter, Conference proceedings on Object-oriented programming systems, languages and applications, p.354-365, October 04-08, 1987, Orlando, Florida, United States
• 5: W. E. Weihl, Local atomicity properties: modular concurrency control for abstract data types, ACM Transactions on Programming Languages and Systems (TOPLAS), v.11 n.2, p.249-282, April 1989 [doi>10.1145/63264.63518]