Two-tier resource allocation for slowdown differentiation on server clusters

Slowdown, defined as the ratio of a request's queueing delay to its service time, is accepted as an important quality of service metric of Internet servers. In this paper, we investigate the problem of providing proportional slowdown differentiation (PSD) services to various applications and clients on cluster-based Internet servers. We extend a closed-form expression of the expected slowdown of a popular Internet workload model with a typical heavy-tailed service time distribution from a single server mode to a server cluster mode. Based on the closed-form expression, we design a two-tier resource allocation approach, which integrates a dispatcher-based node partitioning scheme and a server-based dynamic process allocation scheme. We evaluate the two-tier resource allocation approach via extensive simulations and compare it with an one-tier node partitioning approach. Simulation results show that the two-tier approach can provide fine-grained PSD services on cluster-based Internet servers. We implement the two-tier approach on a cluster testbed. Experimental results further demonstrate the feasibility of the approach in practice.

[1]  Parameswaran Ramanathan,et al.  Proportional differentiated services: delay differentiation and packet scheduling , 2002, TNET.

[2]  Daniel A. Reed,et al.  NCSA's World Wide Web Server: Design and Performance , 1995, Computer.

[3]  Roch Guérin,et al.  A simple FIFO-based scheme for differentiated loss guarantees , 2007, Comput. Networks.

[4]  Van Jacobson,et al.  A Two-bit Differentiated Services Architecture for the Internet , 1999, RFC.

[5]  Cheng-Zhong Xu,et al.  A robust packet scheduling algorithm for proportional delay differentiation services , 2004, IEEE Global Telecommunications Conference, 2004. GLOBECOM '04..

[6]  Jerome A. Rolia,et al.  Characterizing the scalability of a large web-based shopping system , 2001, ACM Trans. Internet Techn..

[7]  Cheng-Zhong Xu,et al.  Processing Rate Allocation for Proportional Slowdown Differentiation on Internet Servers , 2004, IPDPS.

[8]  Gianfranco Ciardo,et al.  Profit-driven service differentiation in transient environments , 2003, 11th IEEE/ACM International Symposium on Modeling, Analysis and Simulation of Computer Telecommunications Systems, 2003. MASCOTS 2003..

[9]  Tao Yang,et al.  Proceedings of the 5th Symposium on Operating Systems Design and Implementation Integrated Resource Management for Cluster-based Internet Services , 2022 .

[10]  Prasant Mohapatra,et al.  Performance Evaluation of Service Differentiating Internet Servers , 2002, IEEE Trans. Computers.

[11]  Marcos José Santana,et al.  Using adaptive priority scheduling for service differentiation QoS-aware Web servers , 2004, IEEE International Conference on Performance, Computing, and Communications, 2004.

[12]  Edward D. Lazowska,et al.  Adaptive load sharing in homogeneous distributed systems , 1986, IEEE Transactions on Software Engineering.

[13]  Erich M. Nahum,et al.  Locality-aware request distribution in cluster-based network servers , 1998, ASPLOS VIII.

[14]  Ashish Goel,et al.  Multi-processor scheduling to minimize flow time with ε resource augmentation , 2004, STOC '04.

[15]  David K. Y. Yau,et al.  A proportional-delay DiffServ-enabled Web server: admission control and dynamic adaptation , 2004, IEEE Transactions on Parallel and Distributed Systems.

[16]  Mor Harchol-Balter.  Task assignment with unknown duration , 2002, JACM.

[17]  Tao Yang,et al.  Demand-driven service differentiation in cluster-based network servers , 2001, Proceedings IEEE INFOCOM 2001. Conference on Computer Communications. Twentieth Annual Joint Conference of the IEEE Computer and Communications Society (Cat. No.01CH37213).

[18]  Cheng-Zhong Xu,et al.  Harmonic proportional bandwidth allocation and scheduling for service differentiation on streaming servers , 2004, IEEE Transactions on Parallel and Distributed Systems.