Course 16082

Performance engineering of distributed e-business systems

 

Department in charge: COMPUTER ARCHITECTURE
Nature: Optional
Number of credits: 3
Language: English
Academic year: 2006/07
Semester: Second semester
Vacancies limit for external students: No limits
Supervisor/s:  JORDI TORRES


Description:
Performance and scalability issues in modern e-business systems are gaining in importance as we move from hype and prototypes to real operational systems. System architects and deployers are often faced with questions such as the following:

 

·       Which deployment platform (hardware and software) would provide the best scalability and cost/performance ratio for a given project?

·       Are there potential hardware or software bottlenecks in the platform selected?

·       How much hardware resources (servers, CPUs, memory, etc.) are needed to meet the service-level agreements?

·       For a given deployment, what would be the average transaction response time, throughput and server utilization under the expected workload?

 

Answering the first two questions requires measuring the performance and scalability of e-business platforms which is typically done using benchmarks. Answering the second two questions requires predicting the performance of a given e-business deployment which is normally done through load testing or performance modeling. The course will provide an introduction to the state-of-the-art in the above mentioned areas concentrating on benchmarking and performance modeling.

 

Content:

1.       Introduction to performance engineering of e-business systems

1.1.  Life cycle of e-business systems

1.2.  Basic performance concepts

1.3.  Approaches to performance engineering

 

2.       Experimental performance evaluation

2.1.  Metrics of performance

2.2.  Average performance and variability

2.3.  Errors in experimental measurements

2.4.  Comparing alternatives

2.5.  Measurement tools and techniques

2.6.  Experimental design

 

3.       Benchmarking e-business platforms

3.1.  Benchmarking methodologies

3.2.  Overview of popular benchmarks

3.3.  Proprietary benchmarks

3.4.  Open-source benchmarks

3.5.  Industry-standard benchmarks

3.6.  Applications of benchmarks

4.       Performance modeling and prediction

4.1.  Operational analysis

4.2.  Characterizing e-business workloads

4.3.  Overview of performance models

4.4.  Measurement techniques

4.5.  Model validation and calibration

4.6.  Performance modeling methodology

 

5.       Case Studies

 

 Evaluation method:

The evaluation considers three type of activities:
- Attendance and participation.
- Talk and presentation related to some of the topics covered in the course.
- Paper (10-12 pages in a tutorial-like form) about one topic related to the course or a experimental evaluation of a particular product/technique.

Calendar:

 

(From 3pm to 6pm)

 

Wednesday June 7,  room C6-E101
Friday June 9,  room C6-E101
Monday June 12 ,  room C6-E106
Wednesday June 14 ,  room D6-114
Friday June 16 ,  room C6-E101
Monday June 19 ,  room C6-E106
Tuesday June 20 ,  room C6-E106
Wednesday June 21 ,  room 00 C6-E101

Tuesday July 4 ,  room C6-E106
Wednesday July 5 ,  room  C6-E101


Bibliography:

1.       "Performance Engineering of Distributed Component-Based Systems - Benchmarking, Modeling and Performance Prediction" by Samuel Kounev

Shaker Verlag, 2005, ISBN: 3832247130

http://www.amazon.de/exec/obidos/ASIN/3832247130/ref=pd_ts_b_10/303-9087391-3338644

 

2.       "Performance by Design: Computer Capacity Planning by Example" by Daniel A. Menasce, Virgilio A.F. Almeida and Lawrence W. Dowdy

Prentice Hall, 2004, ISBN 0-13-090673-5

 

3.       “Measuring Computer Performance - A Practitioner's Guide” by David J. Lilja

Cambridge University Press, 2000, ISBN 0-521-64105-5

      

4.      "Scaling for E-Business: Technologies, Models, Performance, and Capacity Planning" by Daniel A. Menasce and Virgilio A.F. Almeida

Prentice Hall, 2000, ISBN 0-13-086328-9

 

5.       “Capacity Planning for Web Performance: Metrics, Models, and Methods” by Daniel Menasce, Virgilio A. F. Almeida

Prentice Hall, 1998, ISBN 0136938221

 

6.      Conference and journal papers.

 

Instructors:


Dr.-Ing. Samuel Kounev

Affiliations

University of Cambridge - Computer Laboratory
OPERA Group, Systems Research Group
PostDoc Research Fellow

 

Darmstadt University of Technology
Databases and Distributed Systems Group
Research Associate

 

Email: skounev [at] dvs1.informatik.tu-darmstadt.de or Samuel.Kounev [at] cl.cam.ac.uk

Web Page: http://www.dvs1.informatik.tu-darmstadt.de/staff/skounev/

 

Office:   UPC Campus Nord C6-115
Telf. 93 401 7223

 

Dr. Jordi Torres

Affiliations

Technical University of Catalonia

Computer Architecture Department

Associate Professor

 

Autonomic Systems and eBusiness Platforms

Barcelona Supercomputing Center  (BSC-CNS)

Research Manager

 

Email:  torres [at] ac.upc.edu

Web Page: http://personals.ac.upc.edu/torres

 

Office:   UPC Campus Nord C6-217
Telf. 93 401 7223