The publish/subscribe paradigm has found wide acceptance in a broad variety of use cases that differ dramatically in the characteristics of their workloads. Many different systems have been developed both by academia as well as industry, but there is no definitive benchmark, which enables a fair comparison between the different systems.
In this demo, we present PSBench, a benchmark specification and suite for publish/subscribe systems that covers a broad variety of publish/subscribe workloads and scenarios. The benchmark suite is extensible and generic, but the specification targets social games. Social games are the ideal use case since they have a very broad range of requirements and produce a variety of publications and subscriptions. We draw from our experience in massive multiplayer online games to construct a highly realistic workload. In this demo, we present the toolchain, the workload and the graphical interfaces that enable an extensive performance evaluation of publish/subscribe systems.
[1]
Samuel Kounev,et al.
Workload Characterization of the SPECjms2007 Benchmark
,
2007,
EPEW.
[2]
David M. Eyers,et al.
Towards a common API for publish/subscribe
,
2007,
DEBS '07.
[3]
Jörg Kienzle,et al.
Publish/subscribe network designs for multiplayer games
,
2014,
Middleware.
[4]
Reza Sherafat Kazemzadeh,et al.
The PADRES Publish/Subscribe System
,
2010,
Principles and Applications of Distributed Event-Based Systems.
[5]
Kaiwen Zhang,et al.
Distributed event aggregation for content-based publish/subscribe systems
,
2014,
DEBS '14.
[6]
David S. Rosenblum,et al.
Design and evaluation of a wide-area event notification service
,
2001,
TOCS.
[7]
Hans-Arno Jacobsen,et al.
Efficient event processing through reconfigurable hardware for algorithmic trading
,
2010,
Proc. VLDB Endow..