Sizing multi-tier systems with temporal dependence: benchmarks and analytic models
Tóm tắt
Temporal dependence, as a synonym for burstiness, is often found in workloads (i.e., arrival flows and/or service times) in enterprise systems that use the multi-tier paradigm. Despite the fact that burstiness has deleterious effects on performance, existing modeling and benchmarking techniques do not provide an effective capacity planning for multi-tier systems with temporal dependence. In this paper, we first present strong evidence that existing models cannot capture bursty conditions and accurately predict performance. Therefore, we propose a simple and effective sizing methodology to integrate workload burstiness into models and benchmarking tools used in system sizing. This modeling methodology is based on the index of dispersion which jointly captures variability and burstiness of the service process in a single number. We report experimentation on a real testbed that validates the accuracy of our modeling technique by showing that experimental and model prediction results are in excellent agreement under both bursty and non-bursty workloads. To further support the capacity planning process under burstiness, we propose an enhanced benchmarking technique that can emulate workload burstiness in systems. We find that most existing benchmarks, like the standard TPC-W benchmark, are designed to assess system performance only under non-bursty conditions. In this work, we rectify this deficiency by introducing a new module into existing benchmarks, which allows to inject burstiness into the arrival stream in a controllable and reproducible manner by using the index of dispersion as a single turnable knob. This approach enables a better understanding of system performance degradation due to burstiness and makes a strong case for the usefulness of the proposed benchmark enhancement for capacity planning of enterprise systems.
Tài liệu tham khảo
Almeida V, Bestavros A, Crovella M, de Oliveira A (1996) Characterizing reference locality in the WWW. In: IEEE conference on parallel and distributed information systems, Dec 1996
Almeida V, Arlitt M, Rolia J (2002) Analyzing a web-based system’s performance measures at multiple timescales. ACM Perform Eval Rev 30(2):3–9
Andersen AT, Nielsen BF (1998) A Markovian approach for modeling packet traffic with long-range dependence. IEEE J Sel Areas Commun 16(5):719–732
Andersen AT, Nielsen BF (2002) On the use of second-order descriptors to predict queueing behavior of MAPs. Nav Res Logist 49(4):391–409
Arlitt M, Jin T (1999) Workload characterization of the 1998 World Cup website. Technical Report HPL-1999-35R1
Arlitt M, Williamson C (1996) Web server workload characterization: the search for invariants. In: Proc of ACM SIGMETRICS, pp 126–137
Arlitt M, Friedrich R, Jin T (1999) Workload characterization of a web proxy in a cable environment. ACM Perform Eval Rev 27(2):25–36
Balbo G, Serazzi G (1996) Asymptotic analysis of multiclass closed queueing networks: common bottlenecks. Perform Eval 26(1):51–72
Banga G, Druschel P (1999) Measuring the capacity of a web server under realistic loads. WWW 2(1–2):69–83
Barford P, Crovella M (1998) Generating representative web workloads for network and server performance evaluation. ACM Perform Eval Rev 26(1):151–160
Casale G, Zhang E, Smirni E (2007) Characterization of moments and autocorrelation in MAPs. ACM Perform Eval Rev 35(1):27–29. Special issue on MAMA workshop
Casale G, Mi N, Smirni E (2008) Bound analysis of closed queueing networks with workload burstiness. In: Proc of SIGMETRICS, pp 13–24
Casale G, Mi N, Cherkasova L, Smirni E (2010) Dealing with burstiness in multi-tier applications: new models and their parameterization (under submission)
Casale G, Mi N, Smirni E (2010) Model-driven system capacity planning under workload burstiness. IEEE Trans Comput 59(1):66–80
Cox DR, Lewis PAW (1966) The statistical analysis of series of events. Methuen, London
Crovella M, Bestravos A (1996) Self-similarity in Word Wide Web traffic: evidence and possible causes. In: Proc of SIGMETRICS
Fonseca R, Almeida V, Crovella M, Abrahao B (2003) On the intrinsic locality properties of web reference streams. Proc IEEE INFOCOM
Garcia D, Garcia J (2003) TPC-W e-commerce benchmark evaluation. IEEE Comput 36:42–48
Gusella R (1991) Characterizing the variability of arrival processes with indexes of dispersion. IEEE J Sel Areas Commun 19(2):203–211
Jung J, Krishnamurthy B, Rabinovich M (2002) Flash crowds and denial of service attacks: characterization and implications for CDNs and websites. In: Proc of WWW, pp 293–304
Kant K, Tewary V, Iyer R (2001) An internet traffic generator for server architecture evaluation. In: Proc of workshop computer architecture evaluation using commercial workloads
Kobayashi H, Mark BL (2009) System modeling and analysis: foundations of system performance evaluation
Krishnamurthy D, Rolia J (1998) Predicting the QoS of an electronic commerce server: those mean percentiles. ACM Sigmetrics Perform Eva Rev 26(3):16–22
Li H (2010) Realistic workload modeling and its performance impacts in large-scale eScience grids. IEEE Trans Parallel Distrib Syst 21(4):1045–9219
Liu Z, Niclausse N, Jalpa-Villanueva C (2001) Traffic model and performance evaluation of web servers. Perform Eval 46(2–3)
Liu Z, Wynter L, Xia CH, Zhang F (2006) Parameter inference of queueing models for it systems using end-to-end measurements. Perform Eval 63(1):36–60
Menascé DA, Almeida VAF, Dowdy WL (1994) Capacity planning and performance modeling: from mainframes to client–server systems
Menascé DA, Almeida VAF, Reidi R, Pelegrinelli. R. Fonesca F, Meira W Jr. (2000) In search of invariants in e-business workloads. In: Proc of ACM conf electronic commerce, pp 56–65
Mi N, Zhang Q, Riska A, Smirni E, Riedel E (2007) Performance impacts of autocorrelated flows in multi-tiered systems. Perform Eval 64(9–12):1082–1101
Mi N, Casale G, Cherkasova L, Smirni E (2008) Burstiness in multi-tier applications: symptoms, causes, and new models. In: Proc of Middleware
Mi N, Casale G, Cherkasova L, Smirni E (2009) Injecting realistic burstiness into a traditional client–server benchmark. In: Proc of ICAC
Minh TN, Wolters L (2009) Modeling job arrival process with long range dependence and burstiness characteristics. In: Proc of int’l symp on cluster computing and the grid, pp 324–330
Mosberger D, Jin T (1998) httperf: a tool for measuring web server performance. In: Proc of workshop internet server performance
Neuts MF (1989) Structured stochastic matrices of M/G/1 type and their applications. Dekker, New York
Paxon V, Floyd S (1995) Wide area traffic: the failure of poisson modeling. IEEE/ACM Trans Netw 3(3):226–244
Ranjan S, Rolia J, Fu H, Knightly E (2002) QoS-driven server migration for internet data center. In: Proc of IWQoS, pp 3–12
Ranjan S, Rolia J, Fu H, Knightly F (2002) Qos-driven server migration for Internet data centers. In: Proc of IWQoS
Reiser M, Lavenberg S (1980) Mean-value analysis of closed multichain queueing networks. J ACM 27(2):312–322
Riedi RH, Crouse MS, Ribeiro VJ, Baraniuk RG (1999) A multifractal wavelet model with application to network traffic. IEEE Trans Inf Theory 45(4):992–1018
Rolia J, Vetland V (1998) Correlating resource demand information with arm data for application services. In: Proc of WOSP, pp 219–230
Slashdot effect, Wikipedia, Oct 13, 2008. http://en.wikipedia.org/wiki/Slashdot_effect
Sriram K, Whitt W (1986) Characterizing superposition arrival processes in packet multiplexers for voice and data. IEEE J Sel Areas Commun 4(6):833–846
Urgaonkar B, Pacifici G, Shenoy P, Spreitzer M, Tantawi A (2005) An analytical model for multi-tier internet services and its applications. In: Proc of ACM SIGMETRICS, pp 291–302
Urgaonkar B, Shenoy P, Chandra A, Goyal P (2005) Dynamic provisioning of multi-tier internet applications. In: Proc of ICAC
Villela D, Pradhan P, Rubenstein D (2002) Provisioning servers in the application tier for e-commerce systems. ACM Trans Internet Technol 7(1):7
Williams A, Arlitt M, Williamson C, Barker K (2005) Web workload characterization: ten years later. Springer, New York
www.mercury.com/us/products/diagnostics. HP (Mercury) diagnostics
Zhang Q, Cherkasova L, Mathews G, Greene W, Smirni E (2007) R-capriccio: a capacity planning and anomaly detection tool for enterprise services with live workloads. In: Proc of Middleware, pp 244–265
Zhang Q, Cherkasova L, Smirni E (2007) A regression-based analytic model for dynamic resource provisioning of multi-tier applications. In: Proc of ICAC