You cant always get what you want...QoS in CWS.

1
You cant always get what you want...-QoS in CWS.

• Ruth Lennon
• University College Dublin, Ireland.
• Ruth.Lennon_at_lyit.ie

2

• What is and why have QoS?
• What has generative programming to do with it
  all?
• The proposed system.
• QoS and QoE.
• SLAs
• Monitoring.

3
What is QoS anyway?

• QoS is often a compromise between the customer
  requirements and the ability of the service
  providers and the underlying network.
• QoS versus QoE
• Quality of Service versus Quality of Experience.

4
Functional vs Non-Functional

• Functional requirements
• Response time
• Objective requirements.
• Non-Functional requirements
• Throughput
• Up-time
• Trust
• Subjective requirements.

5
Standards

• WSI, BPEL, SOAP, WSDL, WS-Management,
  WS-Reliabillity .
• De-facto standards.
• Where is the conformity?
• SOA Service Oriented Architecture.

6
Generative Programming

• Generative programming has much to offer
  distributed web service applications.
• It has been applied to give BPEL web service
  applications new behaviours.
• It has also be used to enhance communication
  behaviours.

7
Proposed System.

• We propose that the selection of service
  providers include the ability of the clients to
  append additional information regarding the
  required quality of service and the level of
  adherence (hardness) of these criteria.
• HERSOA !
• History, Environment and Rating Enabled SOA.

8
Basic System Structure
9
Generative programming to be applied to HER-SOA

• The development of an application using
  generative programming to adapt the proposed
  system to take into account updates in the
  standards governing web services.
• That is, the web service application may
  continuously carry out introspection to test its
  efficiency with generatively programmed modules
  generated to update the system to perform more
  efficiently under the latest knowledge gained.

10
Monitoring the System

• A monitoring tool will be required to measure the
  performance of the service provider in servicing
  a request.
• Oscillating network quality requires strict
  monitoring.
• Web is not just for disseminating information so
  new measure must be found.

11
Service Provider Rating

• Used by companies such as Amazon.
• Regulated to ensure no spoofing.
• Accuracy improves over time.
• Ratings initially provided by service providers
  but normally provided by clients.

12
QoS Applied.

• The definition of QoS requirements is in the form
  of a tuple (V, Vas, Vtp, Vval, Vw).
• V variable name
• Vas, variable association
• Vtp, variable types
• Vval, variable value
• Vw . variable weighting

13
Service Level Agreement

• SLA define
• Service name,
• Service description
• Server name
• Cost
• Service time
• Average delivery delay
• Security features
• Optional additional specifications.
• SLA between Client and Mediator.
• SLA between Mediator and SPs.

14
Simulations

• A number of simulations were carried out to test
  the basic theorem.
• Simulations of various scenario including
• SP downtime
• Network delay/faults
• Results are not presented here.

15
Possibilities

• Developing a future proof system is not an easy
  task.
• Generative programming can assist in the
  development of new modules to adapt system.
• System design and coding is on-going.

16
You cant always get what you want...-QoS in
CWS.But if you try sometimes, you might find,
you get what you need.

• Ruth Lennon,University College Dublin, Ireland.
• Ruth.Lennon_at_lyit.ie