Quality of Service and Distributed Systems Management . University of Ottawa, of Western Ontario and

1
Quality of Service and Distributed Systems
Management.University of Ottawa, of Western
Ontario and UQAM, 2000-2001http//beethoven.site
.uottawa.ca/DSRG/citr-ec-QoS.htm
Project Overview

• Principal investigator Gregor v. Bochmann (U of
  Ottawa)
• Co-investigators Brigitte Kerhervé (UQAM)
• Hanan
  Lutfiyya (U of Western Ontario)

2
Background

• In the context of electronic commerce,
  standardization of the equipment and software
  packages is not possible. Furthermore, the
  Internet is evolving as a collection of
  heterogeneous networks with different
  capabilities. Therefore the infrastructure has to
  be able to deal with subsystems of different
  capabilities. This is particularly true for the
  quality of service (QoS) of communication between
  the different end-systems involved in the
  application. It is expected that the Internet
  will provide in the future, in addition to its
  traditional best effort'' service also some
  form of guaranteed service quality, probably at a
  cost, which is based on the reservation of
  appropriate system resources. Other limitations
  of the QoS are related to server congestion or
  limited network access bandwidth of the client
  workstation. Even though QoS negotiation and
  adaptation issues have been studied before, the
  global system management aspects related to the
  provisioning of negotiated QoS have not been
  studied extensively.
• In this project, we are mainly concerned with the
  negotiation of appropriate response time for
  queries in the context of electronic commerce.
  The management of the quality of real-time
  multimedia presentations was studied under a
  previous CITR project.

3
Objectives

• To develop policies, using the principle of
  different classes of quality, for the management
  of distributed applications in servers and over
  the Internet.
• To develop scalable models for distributed system
  management which are suitable for supporting a
  very large number of electronic commerce buyers
  and vendors, especially addressing QoS issues.
• To develop query optimization techniques for
  parallel shared-nothing database servers using
  QoS information provided by the underlying
  networks

4
Major technical challenges

• Formulation of policies at a high level of
  abstraction, close to the user's perception of
  quality and benefit
• distinction between several classes of service
  (depending on user profile)
• Acquiring knowledge about the system's current
  performance parameters in the distributed
  environment and ensuring that the policies are
  satisfied at run time
• monitoring of dynamic system parameters (server
  workload, network congestion status)
• Scalability
• large user population, large number of servers,
  large databases
• QoS considerations in distributed query
  processing
• impact of communication QoS on query optimization
• considering different optimization criteria

5
Problems addressed / novel aspects

• Distributed QoS management architecture
• QoS management functions specification, mapping,
  negotiation, resource reservation, adaptation and
  monitoring
• To obtain scalable QoS management, the QoS
  management functions may have to be distributed
  (user workstation, servers, networks, brokers)
• Distributed QoS management protocols should be
  scalable
• Distributed query optimization using QoS
  information
• Depending on the users QoS requirements,
  different optimization criteria may be considered
• low cost, short delay
• Optimization algorithm should take available
  network QoS into account

6
Progress (1)Definition of several distributed
algorithms for load sharing and management
between brokers, users and servers

• We have defined an architecture that introduces a
  brokerage function between clients and servers.
  Brokers continuously monitor the performance of
  affiliated servers and assign them to clients
  according to a previously defined server
  selection policy. We have defined several server
  selection policies to optimize the capacity of
  the system based on server performance
  characteristics and users requirements. In order
  to make realistic predictions about the
  performance of the different servers, we have
  developed an approach to estimate the server
  performance as a function of the number of
  concurrent requests that may share the server at
  a given time (report in preparation). This
  estimation is based on a server performance
  model, which is obtained by monitoring the server
  and its behavior under various load conditions.
  We have performed an analysis study of the
  performance of our architecture and its different
  load-sharing algorithm by simulation.
• A prototype demonstration of this load sharing
  architecture is provided in the EC Major Project
  demonstration prototype. The selection of a
  server at the broker is based on performance
  information obtained by monitoring several Apache
  servers running of different computers.

7
Progress (2)Query optimization strategies
integrating QoS-based cost models and translation
of user requirements into query optimization
criteria

• We have investigated distributed query
  processing, particularly cost-based query
  optimization, and we have proposed an approach in
  which considers QoS (Quality of Service) both
  from the user's requirements perspective and from
  the network service availability. We have also
  proposed an adaptive cost model for distributed
  query processing. Our cost model is adaptive in
  the sense that first, it combines multiple
  optimization criteria, response time and money
  cost, into a simple cost model and second, it can
  give a more precise communication cost estimation
  according to the information captured by the QoS
  manager. This cost model is flexible because it
  can capture the user's willingness to pay for the
  query and the performance dynamics of the
  computer system. Accordingly, we can also
  consider two different optimization criteria the
  users criteria considering the delivered
  response time versus the cost of the query (based
  on existing tariff structures), and the systems
  criteria considering overall optimal resource
  utilization, the satisfaction of the users
  response time requirements and the net income
  from the usage charges. We also identified two
  network QoS parameters end-to-end delay and
  available bandwidth, and introduced methods for
  measuring them.
• Given the general approach described above, we
  have build a prototype implementation. Three data
  models are of interest the user profile, the
  global catalog of distributed database schemas
  and the measured QoS information concerning the
  network and server load. The user profile is
  helpful for translating QoS requirements into
  query optimization criteria it is also useful
  for guiding the optimizer to choose the correct
  cost model. The global catalog and the QoS
  information are mainly used by the cost models.

8
Progress (3)Algorithms for adjusting CPU
priorities in order to maintain several classes
of service on a single server

• We specifically examined applying algorithms for
  service differentiation to Web servers (since it
  is the cornerstone of many Web applications) and
  implemented a version of such an algorithm in the
  context of an Apache server. When a user
  request comes into the Apache server, it gets
  assigned its own process. This process registers
  with the QoS module. Its reference handle is put
  into a queue and the process is put to sleep. The
  QoS module has a scheduling algorithm that wakes
  up processes based on policies. For example,
  assume that we have two classes of service A and
  B. Assume that the "A" class is the premium
  class. One policy is that there can be at most
  "M" B class processes executing and at most "N" A
  class processes executing (M lt N). The
  disadvantage of this approach is that if there
  are few A class processes then the CPU is not
  being effectively used. Another algorithm treats
  A and B classes equally until there are a certain
  number of complaints from A processes indicating
  that they are taking too long to process. The
  number of non-sleeping B class processes is
  reduced. We have developed and experimented
  with several scheduling algorithms. The QoS
  module was designed so that it is relatively easy
  to change the scheduling algorithm.

9
Milestones for 2000-2001

• Analysis of the performance of several
  distributed algorithms for load sharing, using
  experiments with a prototype implementation and
  performance simulations.
• Performance analysis of query optimization
  strategies based on experimentation with the
  prototype and simulation studies improved
  optimization strategies.
• Design and prototype implementation of algorithms
  for adjusting multiple resource allocations in
  order to maintain several differentiated classes
  of service on a single server.

10
Some recent publications

• M. Katchabaw, S. Howard, H. Lutfiyya, A.
  Marshall, and M. Bauer, ìMaking distributed
  applications manageable through instrumentationî,
  In Press, The Journal of Systems and Software,
  1999.
• H. Lutfiyya, A. Marshall, H. Bauer, P. Martin,
  and W. Powley, ìConfiguration Maintenance for
  Distributed Application Managementî, Journal of
  Network and Systems Management, In Press, 1999.
• Bochmann, G. v., Kerhervé, B. and Mohamed-Salem,
  M., Service management issues in electronic
  commerce applications, in Electronic Commerce
  Technology Trends Challenges and Opportunities,
  W.Kou and Y. Yesha (eds), IBM Press, 2000, pp.
  227-238.
• H. Lutfiyya, A. Marshall, M. Bauer, and D.
  Stokes, ìA Policy-Driven Approach to Availability
  and Performance Management in Distributed
  Systems, Journal of Network and Systems
  Management, cond. accepted, 1998.
• M. Katchabaw, H. Lutfiyya, and M. Bauer,
  ìDriving Resource Management with
  Application-Level Quality of Service
  Specificationsî. First International Conference
  on Information and Computation Economies (ICE98),
  October, 1998. Also to be published in Journal of
  Decision Support Systems.
• M. Katchabaw, H. Lutfiyya, and M. Bauer, ìA
  Model of Resource Management to Support
  End-to-End Application-Driven Managementî. First
  International Conference on Information and
  Computation Economies (ICE98), October, 1998.
  Also to be published in Journal of Decision
  Support Systems
• H. Ye, B. Kerhervé, G. v. Bochmann, QoS-aware
  distributed query processing, DEXA Workshop on
  Query Processing in Multimedia Information
  Systems (QPMIDS), 10th International Workshop on
  Database Expert Systems Applications, Florence,
  Italy, 1-3 September, 1999, Proceedings published
  by IEEE Computer Society, 1999.
• H. Ye, G.v. Bochmann, B. Kerhervé, An adaptive
  cost model for distributed query processing, UQAM
  Technical Report, November 1999.