Извините, этот техт доступен только в “Американский Английский”. For the sake of viewer convenience, the content is shown below in the alternative language. You may click the link to switch the active language.
MATHEMATICAL MODELING OF REQUESTS FLOW TO CLOUD COMPUTE CLUSTER
DOI: 10.36724/2072-8735-2020-14-10-39-44
Dmitriy O. Kupriyanov,
Moscow Technical University of Communications and Informatics, Moscow, Russia,
dim.kupro@gmail.com
Abstract
Quality of service parameters estimation becomes even more valuable when using cloud compute services. Mathematical model of described cloud-deployed web application in terms of Processor Sharing (PS) policy for mono-service traffic type proposed in this research. This model has Poisson distribution of the incoming requests flow with intensity ?. Requests awaiting in queue, queue length is considered to be unlimited and all requests should be served. Request in this system is HTTP request with a special payload in JSON format. The size of this payload is different for each request but it lies in a narrow band of values (bytes or decades of bytes). A model of cloud compute cluster was built. Characteristics of relative serving efficiency and relative bandwidth of a single requests flow was calculated using this model for different amount of resource provided for processing of a single request. The dependency of these characteristics from cluster load coefficient is demonstrated in charts. Some conclusions on cloud cluster QoS parameters behavior after the change of input requests flow size. Proposed model helps estimating quality of service parameters and adopting the infrastructure to increased or decreased number of requests from customers and could be used for architecting, deploying and administrating web services.
Keywords: cloud computing, web applications infrastructure, Markovian processes, Processor Sharing, queuing theory, QoS
References
1. RFC 7231, HTTP/1.1 Semantics and Content, June 2014
2. Kubernetes Documentation pages. Managing Compute Resources for Containers. Resource units in Kubernetes. April, 13 (https://kubernetes.io/docs/concepts/configuration/manage-compute-resources-container/)
3. Google Kubernetes Engine documentation pages. April, 18 (https://cloud.google.com/kubernetes-engine/docs)
4. Limoncelli T.A., Chalup S.R., Hogan C.J. The Practice of Cloud System Administration — Addison-Wesley. 2015. 524 p.
5. Stepanov S.N. Teletraffic theory: concepts, models, annexes / Series «Theory and practice of infocommunications». Moscow: Hot Line — Telecom, 2015. 868 p.
6. Iversen V.B., Stepanov S.N. Methods of computation part decrease for calculating telecommunication models with losses, based on ignoring states with low probability. Information transfer problems. 2001. Vol. 37. No.3. P. 82-95. (in Russian)
7. Iversen V.B., Stepanov S.N., Kostrov A.V. Dimensioning of Multiservice Links Taking Account of Soft Blocking. Lecture Notes in Computer Science. 2006. Vol. 4003 LNCS. P. 3-10.
8. Stepanov S.N., Stepanov M.S. Resource planning for joint multiservice real time traffic and elastic data traffic. Automatics and telemechanics. 2017. No. 11. P. 79-93. (in Russian)
9. Iversen V.B., Stepanov S.N. The optimal dimensioning of multi-service links. Proceedings from COST285 Mid-term Symposium, Munchen, September 8-10, 2005. Chapter 7 (pp. 151-178) in A. Nejat Ince & Ercan Topuz (editors): «Modeling and simulation tools for emerging telecommunication networks». Springer 2006. 510 p.
10. Bonald T., Roberts J. Internet and the Erlang formula. ACM SIGCOMM Computer Communication Review. 2012. Vol. 42. No. 1. P. 22-30.
11. Bonald T., Feuillet T. Network Performance Analysis — ISTE Ltd 2011, DOI:10.1002/9781118602911. 253 p.
12. Stepanov S.N. Basics of multiservice networks teletraffic. Moscow: Eco- Trends. 2010. 392 p. (in Russian)
13. Stepanov S.N., Savenkov Y.Y Informational flows clustering for increasing efficiency of channel resource usage in multiservice networks. Telecommunications. 2010. No. 11. P. 22-25. (in Russian)
Information about author:
Dmitry O. Kupriyanov, postgraduate student, Moscow Technical University of Communications and Informatics, Moscow, Russia