Minimizing the mean delay of quorum-based mutual exclusion schemes

2009 28th IEEE International Symposium on Reliable Distributed Systems

Reiter

2009

Abstract-In wide-area settings, unpredictable events, such as flash crowds caused by nearly instantaneous popularity of services, can cause servers that are expected to respond quickly to instead suddenly respond slowly. This presents a problem for achieving consistently good performance in quorum-based distributed systems, in which clients must choose which quorums (sets of servers) to access. Typically, clients are motivated to choose quorums containing the servers that respond fastest. Often, these may be the closest servers, but when the closest servers are particularly slow to respond, e.g., because of a changed workload, servers that are farther may actually respond faster. In this paper, we show how clients can locally change their quorum selections efficiently such that the overall system performance rapidly converges to that of the best global strategy for the current conditions. Moreover, we discuss how to benefit even when changes in quorums must be accompanied by expensive state-transfer operations.

Section: Related Workmentioning

confidence: 99%

When and How to Change Quorums on Wide Area Networks

Merideth

2009 28th IEEE International Symposium on Reliable Distributed Systems

Reiter

2009

“…If we set α = 0, the response time for a client request becomes δ f (v, Q) = max w∈f (Q) d (v, w). This transforms the objective function into average network delay, a measure studied previously [9,14,15,19]. We stress that these definitions are merely tools for placing quorum systems; our interest is in seeing how well we can use these models to efficiently deploy quorum systems in realistic wide-area topologies.…”

Section: Algorithmsmentioning

confidence: 99%

“…That is, these techniques strive to adapt the efficiencies of content distribution networks (CDNs) like Akamai to more dynamic services, and in doing so they employ accesses to quorums of proxies in order to coordinate activities. Second, there has recently been significant theoretical progress on algorithms for deploying quorums in physical topologies, so that certain network measures are optimized or approximately optimized (e.g., [9,12,14,15,19,27]). These results have paved the way for empirical studies using them, which is what we seek to initiate here.…”

Section: Introductionmentioning

confidence: 99%

Minimizing Response Time for Quorum-System Protocols over Wide-Area Networks

37th Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN'07)

Reiter

2007

“…Kobayashi et al [11] looked at the problem of designing quorums Q to minimize Avg v∈V [minQ∈Q δ(v, Q)]. They gave a branch-and-bound algorithm for it, which could be evaluated only on topologies with up to 20 nodes due to its exponential running time, and they also conjectured that the problem is NP-hard.…”

Section: Related Workmentioning

confidence: 99%

“…Constraints (14) and (11) ensure that if xtQ = 1 then all the elements in Q are indeed placed on some subset of {v0, . .…”

Section: Vn−1} So That D(v0 V1) ≤ D(v0 V2) ≤ · · · ≤ D(v0 Vn−1mentioning

confidence: 99%

Quorum placement in networks to minimize access delays

Gupta

Maggs

Proceedings of the Twenty-Fourth Annual ACM Symposium on Principles of Distributed Computing

et al. 2005

A quorum system is a family of sets (themselves called quorums), each pair of which intersect. In many distributed algorithms, the basic unit accessed by a client is a quorum of nodes. Such algorithms are used for applications such as mutual exclusion, data replication, and dissemination of information. However, accessing spread-out quorums causes access delays that we would like to minimize. Furthermore, every member of the quorum incurs processing load to handle quorum accesses by clients.In this paper we study the problem of placing quorums in a physical network so as to minimize the delay that clients incur by accessing quorums, and while respecting each physical node's capacity (in terms of the load of client requests it can handle). We provide approximation algorithms for this problem for two natural measures of delay (the max-delay and total-delay). All our algorithms ensure that each node's load is within a constant factor of its capacity, and minimize delay to within a constant factor of the optimal delay for all capacity-respecting solutions. We also provide better approximations for several well-known quorum systems.