Law-governed interaction

Abstract

Software technology is undergoing a transition from monolithic systems, constructed according to a single overall design, into conglomerates of semi-autonomous, heterogeneous and independently designed subsystems, constructed and managed by different organizations, with little, if any, knowledge of each other. Among the problems inherent in such conglomerates none is more serious than the difficulty to control the activities of the disparate agents operating in it, and the difficulty for such agents to coordinate their activities with each other.We argue that the nature of coordination and control required for such systems calls for the following principles to be satisfied: (1) coordination policies need to be enforced; (2) the enforcement needs to be decentralized; and (3) coordination policies need to be formulated explicitly-rather than being implicit in the code of the agents involved-and they should be enforced by means of a generic, broad spectrum mechanism; and (4) it should be possible to deploy and enforce a policy incrementally, without exacting any cost from agents and activities not subject to it.We describe a mechanism called law-governed interaction (LGI), currently implemented by the Moses toolkit, which has been designed to satisfy these principles. We show that LGI is at least as general as a conventional centralized coordination mechanism (CCM), and that it is more scalable, and generally more efficient than CCM. * Appeared in ACM Transactions on Software Engineering and Methodology (TOSEM) †

Regularity-Based Trust in Cyberspace

“…A community under LGI is open in the following sense: (a) its membership can change dynamically, and can be very large; and (b) its members can be heterogeneous. For more details about LGI, and about its implementation, than provided by this overview the the reader is referred to [18], and to [23] for more recent results.…”

On Shouting “Fire!”: Regulating Decoupled Communication in Distributed Systems

“…A community under LGI is open in the following sense: (a) its membership can change dynamically, and can be very large; and (b) its members can be heterogeneous. For more details about LGI than provided by this overview, the reader is referred to [13,1,2].…”

“…Our measurement of the most recent implementation of the LGI controller shows that each regulated event is processed in about 0.6 ms, on a Sun Fire 280R, UltraSPARC-III server (900 MHz), under the Solaris 2.8 operating system and Java 1.3. (Note that this version still does not incorporate the improvement measures suggested previously in [13].) The general picture that emerges below is as follows: Our mechanism tends to decrease the load on the mediator (in some cases dramatically), reducing the probability of congestion, and to increase the mediator's throughput, but it involves modest increase in latency, when the mediator is not congested.…”

“…8, where T α,β stands for the time of communication between two processes α and β, while is used as the time for finding the matching subscriptions, that of the accesscontrol, and that of event-evaluation by the LGI controller. Letting T a and T b be the latency in cases (a) and (b), respectively, we consider the relative overhead (T b − T a )/T a , by following the general discussion in [13].…”

“…Therefore, we take here an alternative approach, which, in the context of the publish/subscribe paradigm, can be characterized as follows: Instead of having the P/S mediator define and enforce an access-control policy over all publications and subscriptions made through it, we have a substantially richer policy defined and enforced directly over the agents attempting to communicate via the P/S mediator. This is done by means of a distributed coordination and control mechanism called Law-Governed Interaction (LGI) [13], and a middleware called Moses that implements this mechanism -an overview of which is provided in Sect. 3.…”