Control Theory

 Control Theory Essay

Aside from formatting, this report is identical to the paper posted in IEEE Journal on Selected Areas in Communications, Volume twenty three, Number doze, December 2006. It was retroactively submitted as being a Columbia Pc Science specialized report on May 10, 2007 and is given a technical report quantity and time reflecting the original publication date.

A Control Theory Foundation for Self-Managing Calculating Systems Yixin Diao, Member, IEEE, Paul L. Hellerstein, Senior Affiliate, IEEE, Sujay Parekh, College student Member, IEEE, Rean Griffith, Gail Electronic. Kaiser, Senior Member, IEEE, and Dan Phung

Abstract—The high cost of functioning large computer installations has motivated a diverse interest in reducing the need for human intervention by causing systems self-managing. This newspaper explores the extent where control theory can provide an architectural and analytic foundation for building self-managing devices. Control theory provides a abundant set of strategies for building automated self-diagnosis and self-repairing systems with properties such as stability, brief settling times, and exact regulation. Nevertheless , there are challenges in applying control theory to computer systems, such as developing powerful resource versions, handling messfuhler delays, and addressing business lead times in effector activities. We suggest a deployable testbed intended for autonomic calculating (DTAC) that individuals believe can reduce the obstacles to dealing with research problems in making use of control theory to calculating systems. The original DTAC architecture is defined along with several problems that it can be used to look at. Index Terms—Actuator, closed loop control, dynamics, useful resource management, messfuhler, testbed.


Fig. 1 . Structures for autonomic computing.

Big t

HE Very high cost ownership of computing devices has resulted in a number of industry initiatives to minimize the burden of operations and management. Examples include IBM's Autonomic Computing, HP's Adaptive Facilities, and Microsoft's Dynamic Systems Initiative. Many of these efforts seek to reduce businesses costs simply by increased software, ideally to have systems be self-managing with no human intervention (since owner error have been identified as being a major method to obtain system failures [1]). While the concept of automatic operations features existed for 2 decades (e. g., [2]) as a way to adapt to changing workloads, failures, and (more recently) attacks, the scope of automation is still limited. We feel this is in part due to the a shortage of a fundamental knowledge of how computerized actions affect system patterns, especially, program stability. Various other disciplines including mechanical, electric, and aeronautical engineering make use of control theory to design responses systems. This kind of paper uses control theory as a way to discover a number of requirements for and challenges in building self-managing systems.

Manuscript received Summer 30, 2006; revised This summer 20, june 2006. The work with the Programming Systems Laboratory is usually supported partly by the Nationwide Science Base under Scholarhip CNS-0426623, Offer CCR-0203876, and Grant EIA0202063, and in part by Microsoft Research. Con. Diao, T. L. Hellerstein, and T. Parekh will be with the IBM Thomas J. Watson Exploration Center, Hawthorne, NY 10532 USA (e-mail: [email protected] apple. com; [email protected] ibm. com; [email protected] apple. com). 3rd there’s r. Grifп¬Ѓth, G. E. Kaiser, and D. Phung are with the Pc Science Section, Columbia University, New York, BIG APPLE 10027-7003 UNITED STATES (e-mail: [email protected] columbia. edu; [email protected] columbia. edu; [email protected] columbia. edu).

The APPLE autonomic computer architecture [3] provides a platform in which to generate self-managing systems. We utilize this architecture because it is extensively consistent with additional approaches which were developed (e. g., [4]). Fig. you depicts the components and crucial interactions to get a single autonomic manager and a single resource. The source (sometimes known as managed resource) is what is being created more self-managing. This could be just one system...

Referrals: [1] A. Fox and D. Patterson, " Self-repairing computers, ” Sci. Mechant., pp. 54–61, May 2003. [2] K. Milliken, A. Cruise, Ur. Ennis, A. Finkel, M. Hellerstein, Deb. Loeb, Deb. Klein, M. Masullo, They would. V. Woerkom, and D. Waite, " Yes/mvs plus the autonomation of operations intended for large computer system complexes, ” IBM Syst. J., vol. 25, no . 2, pp. 159–180, 1986. [3] IBM-Corporation. An executive blueprint intended for autonomic calculating. [Online]. Available: [4] G. Kaiser, T. Parekh, S. Gross, and G. Valetto, " Kinesthetics extreme: Another infrastructure intended for monitoring given away legacy devices, ” in Proc. fifth Annu. Int. Active Middleware Workshop, 2003, pp. 22–30. [5] S. Keshav, " A control-theoretic approach to flow control, ” in Proc. ACM SIGCOMM, Sep. 1991, pp. 3–15. [6] E. Li, M. H. Shor, J. Walpole, C. Pu, and Deb. C. Steere, " Building the effect of short-term rate variations about TCP-friendly over-crowding control patterns, ” in Proc. Virulent. Control Conf., 2001, pp. 3006–3012. [7] E. Altman, T. Proteger, and 3rd there’s r. Srikant, " Congestion control as a stochastic control problem with action delays, ” Automatica, vol. 35, pp. 1936–1950, 1999. [8] C. Sixth is v. Hollot, V. Misra, M. Towsley, and W. N. Gong, " On building improved controllers for AQM routers supporting TCP flows, ” in Proc. IEEE INFOCOM, Chuck, AK, Monthly interest. 2001, pp. 1726–1734. [9]#@@#@!!, " A control theoretic analysis of RED, ” in Proc. IEEE INFOCOM, Anchorage, O VE, Apr. 2001, pp. 1510–1519. [10] Capital t. F. Abdelzaher and And. Bhatti, " Adaptive articles delivery to get Web server QoS, ” Comput. Netw., vol. 23, pp. 1563–1577, 1999. [11] Y. Lu, A. Saxena, and Big t. F. Abdelzaher, " Differentiated caching companies: A control-theoretic approach, ” in Proc. Int. Conf. Distrib. Comput. Syst., Apr. 2001, pp. 615–654. [12] S. Parekh, K. Went up, J. T. Hellerstein, T. Lightstone, M. Huras, and V. Chang, " Taking care of the overall performance impact of administrative programs, ” in Proc. IFIP Conf. Distrib. Syst. Oper. Manage., the year 2003, pp. 130–142. [13] D. Sha, Back button. Liu, Con. Lu, and T. Abdelzaher, " Queueing model based network server performance control, ” in Proc. IEEE Real-Time Syst. Symp., December. 2002, pp. 81–90. [14] S. Parekh, N. Gandhi, J. Hellerstein, D. Tilbury, J. Bigus, and To. S. Jayram, " Employing control theory to accomplish service level objectives in performance managing, ” Current Syst. T., vol. twenty-three, pp. 127–141, 2002. [15] Y. Diao, N. Gandhi, J. D. Hellerstein, S. Parekh, and D. Tilbury, " Using MIMO opinions control to enforce policies for interrelated metrics with application towards the Apache Net server, ” IEEE/IFIP Netw. Oper. Manage., pp. 219–234, Apr. 2002. [16] Sumado a. Diao, M. L. Hellerstein, and S i9000. Parekh, " Optimizing quality of service using unclear control, ” Distrib. Syst. Oper. Take care of., pp. 42–53, 2002. [17] Y. Diao, J. T. Hellerstein, A. Storm, M. Surendra, H. Lightstone, H. Parekh, and C. Garcia-Arellano, " Applying MIMO linear control intended for load managing in computer systems, ” in Proc. Amer. Control Conf., Jun. 2004, pp. 2045–2050.

[18] Y. Diao, J. D. Hellerstein, G. Kaiser, H. Parekh, and D. Phung, " Selfmanaging systems: A control theory foundation, ” Eng. Autonomic Syst., pp. 441–448, April. 2005. [19] J. D. Hellerstein, Y. Diao, S. Parekh, and D. M. Tilbury, Opinions Control of Computing Systems.. Nyc: Wiley, 2005. [20] E. Ogata, Modern day Control Engineering, 3rd impotence. Englewood Cliffs, NJ: Prentice-Hall, 1997. [21] IBM. (2004) Autonomic computing toolkit. Tech. Rep. [Online]. Obtainable: html code [22] L. Zhang, C. Lu, Big t. F. Abdelzaher, and L. A. Stankovic, " Controlware: A middleware architecture intended for feedback control of software performance, ” in Proc. Int. Conf. Distrib. Comput. Syst., 2002, pp. 301–310. [23] TPC. (2004) (TPC) Net. Transaction Control Council, Technical. Rep. [Online]. Offered:

Rean Griffith received the B. Sc. degree in computer system science and management in the University of the West Indies, Barbados, in 2000 and the M. Sc. degree in computer scientific research from Columbia University, Ny, in 2003. Currently, he's working towards the Ph. D. degree in the Programming Systems Laboratory (PSL), Columbia University or college. His research interests consist of adaptive systems, self-healing systems, and system manageability.

Yixin Diao (M'01) received the Ph. Deb. degree in electrical architectural from Ohio State College or university, Columbus, in 2000. He's a Research Employee at the APPLE Thomas M. Watson Study Center, Hawthorne, NY. This individual has printed more than 35 papers and coauthored Reviews Control of Calculating Systems (New York: Wiley, 2004). His research pursuits include devices management software, adaptive control over dynamic systems, autonomic source allocation, and modeling and optimization of distributed devices. Dr . Diao is the person receiving several honours, including the 2002 Best Conventional paper Award at the IEEE paid Network Businesses and Supervision Conference plus the 2002–2005 IFAC Theory Daily news Prize in Engineering Applying Artificial Cleverness.

Joseph L. Hellerstein (M'93–SM'98) received the Ph. Deb. degree in computer technology from the University or college of A bunch of states, Los Angeles. He could be a Research Staff Member and Supervisor of the Adaptable Systems Section, IBM Jones J. Watson Research Middle, Hawthorne, NEW YORK, and a great Adjunct Mentor at Columbia University, New york city. He offers authored or coauthored about 100 expert reviewed content articles, an Addison-Wesley book about expert systems, and an e book entitled Reviews Control of Processing Systems (New York, Wiley, 2004). His research has addressed various areas of service top quality in calculating systems, including predictive recognition, automated prognosis, expert devices, and the putting on control theory to processing systems.

Gail E. Chef (M'85–SM'90) received the South carolina. B. degree from the Massachusetts Institute of Technology, Cambridge, in 1979, as well as the M. T. and Ph. D. deg from Carnegie Mellon School, Pittsburgh, PENNSYLVANIA, in 80 and 85, respectively. She's a Teacher of Computer Science as well as the Director from the Programming Systems Laboratory, Pc Science Office, Columbia University, New York. This wounderful woman has consulted or perhaps worked high seasons for courseware authoring, software process and networking online companies, several protection contractors, the Software Engineering Start, Bell Labs, IBM, Siemens, Sun, and Telcordia. Her laboratory has become funded by the Defense Advanced Research Projects Organization (DARPA), Nationwide Science Foundation (NSF), Office of Naval Research (ONR), National Astronautics and Space Administration (NASA), NYS Technology and Technology Foundation, and lots of companies. Your woman served around the Committee of Examiners to get the Educational Tests Service's Laptop Science Advanced Test (the GRE CS test) for three years, and has chaired her department's doctoral system since 1997. She gets published over 100 refereed papers in a range of software areas. Her research interests incorporate self-managing systems (autonomic computing), publish/subscribe event systems, security, Web technologies, collaborative function, information managing, distributed systems, and computer software development environments and tools. Dr . Chef was known as an NSF Presidential Youthful Investigator in Software Architectural and Computer software Systems in 1988. She dished up on the Editorial Board of IEEE Net Computing for several years, and was obviously a founding Connect Editor of ACM Orders on Application Engineering. Your woman Chaired an ACM SIGSOFT Symposium in Foundations society Engineering, Vice Chaired 3 of the IEEE International Conference on Given away Computing Devices, and provides frequently on conference program committees.

Sujay Parekh (S'05) received the M. H. degree in computer science from the College or university of Wa, Seattle. He can an Advisory Software Professional at the IBM Thomas L. Watson Analysis Center, Hawthorne, NY. He is a coauthor of Responses Control of Computer Systems (New York, Wiley, 2004). He has printed several documents and written for autonomic features for APPLE software. His research passions center about automating equally simple and sophisticated computing systems, and have included work in AI planning, equipment learning, laptop architecture, arranging algorithms, and control systems.

Dan Phung received the B. S i9000. degree in molecular biology from the University or college of New Mexico, Albuquerque, in 2001 and the M. S i9000. degree from Columbia University or college, New York, in 2004. Presently, he is operating towards the Ph level. D. level at Columbia University. He has executed research and published on a diverse selection of topics including human brain image resolution, genomic variations, and distributed software devices. His current research interests pertain to the management, energetic resource allowance, and educated scheduling of high-performance systems.

 Impact of recent Business Dissertation

Impact of recent Business Dissertation

Introduction The objective of this examine is to total two features which incorporate a profile portion and individual expression piece. The portfolio includes material about contemporary organization, which…...

 Essay upon Wonder Young boys by: Eileen Chabon

Essay upon Wonder Young boys by: Eileen Chabon

Stuck: Michael jordan Chabon's " Wonder Boys” Where will i even start with this book? It truly is, quite actually, one of the best-written books I've…...