Capacity Dimensioning Based on Traffic Measurement in the Internet

1
Capacity Dimensioning Based on Traffic
Measurement in the Internet
APT-HRD Collaborative research Program for
Exchange of Researcher/Engineers

• Dr. Mahmood Ashraf Khan
• Institute of Communication Technologies,
• Pakistan Telecommunication Company Limited
• Islamabad, Pakistan

2
Collaborative Institutes

• Multimedia Network Architecture Lab. Osaka
  University, Osaka, Japan
• Institute of Communication Technologies,
  Islamabad, Pakistan.
• Regional Telecom Centre, Islamabad, Pakistan
• Information Technology Infrastructure (ITI)
  Department, PTCL, Pakistan

3
Contents

• QoS in Telecommunications?
• Traffic Measurement Approaches
• Pakistan Internet Exchange (PIE) Network
• Conclusion
• Recommendations

4
QoS in Telecommunications?

• Target applications in telecommunications?
• Real-time Applications Voice and Video
• Require bandwidth guarantees, and thats all
• Past statistics
• Traffic characteristics is well known
• Single carrier, single network
• Erlang loss formula
• Robust (allowing Poisson arrivals and general
  service times)
• QoS measurement call blocking prob.
• Can be easily measured by carrier

5
What is QoS in Data Applications?

• The current Internet provides
• QoS guarantee mechanisms for real-time
  applications by int-serv
• QoS discriminations for aggregated flow by
  diff-serv
• No QoS guarantees for data (even in the future)
• For real-time applications, bandwidth guarantee
  with RSVP can be engineered by Erlang loss
  formula
• But RSVP has a scalability problem in the of
  flows/intermediate routers
• Distribution service for real-time multimedia
  (streaming)
• playout control can improve QoS
• How about data?
• Data is essentially greedy for bandwidth
• Some ISP offers the bandwidth-guaranteed service
  to end users
• No guarantee in the backbone

6
Fundamental Principles for Data Application QoS

• Data applications always try to use the bandwidth
  as much as possible. This is now supported by
• High-performance end systems
• High-speed access lines
• Neither bandwidth nor delay guarantees should be
  expected.
• Packet level metrics (e.g., packet loss prob. and
  packet delay) are never QoS metrics for data
  applications.
• It is inherently impossible to guarantee
  performance metrics in data networks.
• Network provider can never measure QoS
• Competed bandwidth should be fairly shared among
  active users.
• Transport layer protocols (TCP and UDP) is not
  fair.
• Network support is necessary by, e.g., active
  queue management at routers.

7
Fundamental Theory for the Internet?

• Queueing Theory is Useful?
• Only provides packet queueing delay and loss
  probabilities at the node (routers buffer at
  one output port)
• In Erlang loss formula,call blocking probability
  is just a user
• level QoS
• Behavior at the Router?
• Data application utilizes TCP of which congestion
  control performed.
• TCP is inherently a feedback system
• Input to the queue (i.e. to the router) must be
  affected by the feedback.
• Streaming media using UDP/RTP has its proprietary
  rate.
• User level QoS for Data?
• Application level QoS such as Web document
  transfer time.
• We need to develop a new fundamental theory in
  Networking research.
• Control theoretic approach seems to be promising
  to get the system stability and transient
  behaviors.

8
Challenges for Network Provisioning

• New problems we have no experiences in data
  networks
• What is QoS?
• How can we measure QoS?
• How can we charge for the service?
• Can we predict the traffic characteristics in the
  era of information technology?
• End-to-end performance can be known only by end
  users
• QoS prediction at least at the network
  provisioning level

9
Traffic Measurement Approaches

• Passive Measurements (OC3MON, OC12MON,,,)
• Only provides point observations
• Actual traffic demands cannot be known
• QoS at the user level cannot be known
• Active Measurements (Pchar, Netperf, bprobe,,,)
• Provides end-to-end performance
• Measurement itself may affect the results
• Not directly related to network dimensioning (The
  Internet is connectionless!)
• How can we pick up meaningful statistics?
• Routing instability due to routing control
• Segment retransmissions due to TCP error control
• Rate adaptation by streaming media
• Low utilization is because of
• Congestion control?
• Limited access speed of end users?
• Low-power end host?

10
Spiral Approach for Network Dimensioning

• Incremental network dimensioning by feedback loop
  is an only way to meet QoS requirements of users

No means to predict the future traffic demand
Traffic Measurement
Provides confidence on results
Statistical Analysis
Capacity Dimensioning
Flexible bandwidth allocation is mandatory (ATM,
WDM network)
11
Pakistan Internet Exchange Network

• The PIE network was designed due to the
  following
• There was no inter-exchange of Internet traffic
  domestically between the ISPs. All inter ISP
  traffic of Pakistan was first routed to the
  Internet backbone and routed back to Pakistan,
  resulting in wastage of International bandwidth,
  slow Internet response and little or no backup in
  case of loss of connectivity.
• Achieve economy of scales by aggregation of
  traffic.
• Foster PTCL business interest in the field of
  Data Communication.
• Develop a Network Infrastructure for convergence
  of Voice, Video and Data networks.
• Facilitate Law full intercept and monitoring
  facilities.

12
Existing PIE Structure
13
Proposed PIE

• Point of presence is being expanded from 13 POPs
  to 35 POPs.
• The present two155Mbps International would be
  upgraded to 2x622Mbps (2xSTM4) through SMW-4,
  which will make more than 1200mbps of Internet
  connectivity to the backbone available to PIE.
• PIE PoPs in the three cities, Rawalpindi, Lahore
  and Karachi are being expanded by adding two high
  capacity core routers at each site.
• 6 main routers and 26 distribution routers will
  be installed in the rest of the POPs.
• The links between the POPs will be at the level
  of E3/STM1.

14
Proposed PIE Structure
15
Conclusion

• Internet has become the backbone of all the
  activities in the World including education,
  business and research.
• Network Dimensioning is important in order to
  provide QoS to the customer on Internet and is
  crucial to keep the Internet link for best
  utilization
• Bottleneck Identification for traffic measurement
  is the prime factor.
• Experimental results have shown that measurement
  method can provide an accurate estimate of the
  link utilization.

16
Result /Output of the Project

• Good experience to work in collaboration with
  Japanese researchers.
• The present infrastructure for Internet was
  studies and analyzed.
• Based on the analysis new network dimensioning
  was proposed
• A design framework for Pakistan Internet Exchange
  (PIE) network has been proposed and changes have
  been made for the adequate link capacity on the
  basis of end-to-end measurement in the
  infrastructure of PIE network.
• The performance of the PIE network has been
  improved.

17
Recommendations

• The research Program was very valuable and this
  must be sustained.
• More involvement of Japanese side is required as
  those researchers have much knowledge and the
  developing countries can benefit from their rich
  research experience.
• A pool of Japanese Universities may be identified
  where such collaborative research program shall
  be conducted.
• An awareness of this ICT-HRD collaborative
  research program may be given to all the
  Universities at Japan so that they may
  collaborate in this program.

18
Thank you