SEQUIN Premium IP

1. SEQUIN Premium IP

2. Project info • SEQUIN - SErvice QUality across Independently managed Networks(IST-1999-20841) • Duration 18 months (Nov 2000...Apr 2002) • Partners: • Project Web site: http://www.dante.net/sequin • Technical Web site: http://www.switch.ch/lan/sequin • DANTE, DFN, GARR, GRNET, PSNC, RENATER, SWITCH, UKERNA

3. Overview • SEQUIN has defined and implemented an end-to-end approach to Quality of Service (QoS), operating across multiple management domains and exploiting a combination of IP and ATM technology • The project has specified the implementation architecture for the Premium IP service, which aims at offering the equivalent of an end-to-end virtual leased line service at the IP layer across multiple domains. • The architecture is targeted at the GÉANT (The pan-European Gigabit Research Network) and is applicable to each connected National Research and Education Network (NREN) across Europe and local DiffServ domains

4. Definition of QoS • Work of WP2 ->D2.1 • Top-down approach • Interview with international user groups on their requirements • Bottom-up • Network performance parameters that may be influenced by configuration of equipment

5. QoS parameters • From users’ requirements and technical considerations : • One-way delay (OWD) • IP packet delay variation (IPDV) • Available bandwidth • One-way packet loss (OWPL) • The set is common to IETF and ITU-T • Naming and definitions are chosen to be comply to RFC 2330 (Framework for IP Performance metrics) and follow the ongoing IPPM IETF working group work.

6. QoS user requirements (from user’s questionnaire)

7. Premium IP Specification • Differentiated Services Architecture and use the expedited forwarding per hop behavior (EF PHB) • Interface definition between domains that behaves as an EF PHB • Do not starve best effort traffic (limited percentage of link capacity devoted to Premium IP) • Initial provisioning structure: static, no dynamic signaling • IETF IPPM QoS parameters measurement framework • QoS parameters monitoring system is a key element

8. Premium IP Service Implementation • Basic principles • minimize number of actions per node • do not use a signaling protocol • modular approach that allows different implementation schemes at every hop or domain and allows domains to join the service when ready • Do not try to solve the most general problem, but rather develop a model that can be implemented in short time using available tools • Keep it simple

9. Simplifying the actions for each node In principle, each node might perform an awful lot of tasks: - admission control and classification - marking - policing - scheduling - congestion control - shaping - QoS rules propagation - monitoring and accounting

10. Admission control Use the information in the IP header: - IP source and destination (prefixes) as near to the source as possible - the DSCP (or IP precedence equivalent value) along the path - perform an optional, suggested, admission control based on AS source and destination at inter-domain links (safety measure) - rules might be based on additional parameters, such as time-of-day

11. Admission control (continued) The consequences are: - allowing the computation of total requested Premium IP capacity at each network node in the default case (and for main backup cases too) - short access list near users’ premise (few users) - simple control at backbones (IP addresses are not propagated) - choosing destination aware service (next slide)

12. Admission control (continued) C A user 5 B user 4 Destination Aware Vs destination Unaware E user 3 F D user 2 user 1 G

13. Examining the tasks for each node In principle, each node might perform an awful lot of tasks: - admission control and classification always - marking - policing - scheduling - congestion control - shaping - QoS rules propagation - monitoring and accounting

14. Marking - Mark each “EF” legal packet at first classification point - Use the same DSCP value on all domains (decimal 46 [RFC 2474] to have interoperability with ToS only capable hardware) - strongly suggested - - valid DSCP coupled to invalid IP addresses may imply discard to allow easy debugging - packets with other DSCP values are left untouched Marking is mandatory at the first classification point, remarking is optional.

15. Examining the tasks for each node - admission control and classification always - marking - policing - scheduling Selected locations - congestion control - shaping - QoS rules propagation - monitoring and accounting

16. Policing • Microflow policing should be done as close as possible to the source according to agreed (through SLA) Premium IP capacity. This step is mandatory • Policing will be done using a token bucket. The depth of the token bucket will be two MTU close to the source and increase to 5 or more along the path if additional policing is required • It is suggested to perform only one additional policing stage at the ingress to GÉANT from an NREN, with a larger aggregated capacity value than the sum of the agreements. • “Avoid unwanted packet loss” is the motto.

17. Policing (continued) • The additional policing stage at the ingress to GÉANT from an NREN serves the purpose of protecting Premium IP traffic from misconfiguration/DoS coming from a single source. • It creates virtual “pipes” for the aggregated Premium flows from each NREN to each other (when needed). The failure of one “pipe” does not influence the others.

18. Classify by DSCP Police by (AS source,dest) aggregate capacity on all border nodes N3 CORE Policing not needed Policing can be avoided at ingress when receiving from a trusted backbone N1 Classification on IP addresses Strict policing N2 L1 L2 L1, L2 : end user domain (for example LANs) N1, N2, N3 : intermediate transport domains (for example NRENs backbones) CORE : interconnection domain (for example GÉANT) : router/switch Sample multidomain network

19. Examining the tasks for each node - admission control and classification always - marking - policing Selected locations always - scheduling Selected locations - congestion control - shaping - QoS rules propagation - monitoring and accounting

20. Application TCP IP Scheduling Network Interface Shaping The sending source is required to shape the traffic it produces. Shaping can be done by the network close to the user Shaping inside the sending host itself is the preferred way, shaping by the network will in most case lead to packet losses No Packet/Data losses host

21. Shaping C A user 5 B Multiple aggregation-separation points and link speed changes. E user 3 F D user 2 user 1 G

22. Selected locations Selected locations Examining the tasks for each node - admission control and classification always - marking - policing Selected locations - scheduling always Selected locations NO Done by source not needed - shaping - congestion control - QoS rules propagation - monitoring and accounting

23. Police by (AS source,dest) aggregate capacity on all border nodes Policing can be avoided at ingress when receiving from a trusted backbone Shape ONLY here Classify (IP pair prefixes) Police - Strict, Capacity Mark Summary Classify (DSCP) High priority queueing on all nodes Do not police on egress Do not shape

24. Grey Areas Exact configuration of buffering and token bucket depth in routers. As a rule of thumb the token bucket depth can be assumed to be 1.2 * (Diffserv active interfaces on router) Scalability - the maximum amount of aggregated Premium IP capacity the network can offer - hardware capabilities Fast provisioning of the service Widespread availability and tuning of “last mile” (LANs)

25. Example (one direction) Domain 2 ATM Dedicated PVC Domain 1 802.1p VLAN Or dedicated wire Classification (DSCP) Scheduling Classification (IP) Policing (strict 2 MTU) Marking - scheduling Domain 3 Backbone Domain 4 Classification (DSCP) Policing (AS aggregate) Domain 5 802.1p VLAN Or dedicated wire

26. “Proof of Concept” • Initial implementation of the testing methodology by implementing a “Proof of Concept” test-bed involving user groups • Goals: • access to a controlled environment composed of a variety of hardware and underlying technology • functionality verification of each component required toimplement Premium IP • The set of tests performed included: • laboratory tests for basic router functionality • wide area tests for network calibration (understand the performance users can expect & the interaction between different network technologies) • tests involving users to verify the QoS provisioning processes

27. H.323 users tests • H.323 users from TF-STREAM Task Force • TF-STREAM, http://www.terena.info/task-forces/tf-stream/ • Tests • Core network (GÉANT): 10Gbit/s & 2.5 Gbit/s POS and Juniper routers. • 4 high (2.5 Gbit/s POS) and lower (2x155Mbit/s ATM access) speed national networks connecting six testing locations • Traffic tests with measurement tools with/without Premium IP enabled • Objective and subjective quality assessments of H.323 videoconferencing

28. H.323 tests topology

29. Test scenarios • End-to-end setup, between each pair of the participants • Videoconference initiated – users assessment of audio and video quality • ICMP Ping tool was used to measure end-to-end RTT • The videoconference session was terminated • Use of RUDE/CRUDE tool with traffic pattern imitating videoconference stream in both directions for recording jitter and packet loss • NETPERF throughput test was used to assess the bandwidth available for Premium IP service

30. Test results I

31. Test results II

32. Test results III

33. Testing with IST projects • AQUILA (IST 1999-10077) • Enhanced architecture for QoS in Internet • PL (Warsaw) - AT (Vienna), 2.5 Mb/s • activated on 15 April 2002 • MOICANE (IST 2000-26137) • QoS support in access technologies • IT, GR, PT, RO • target time April/May 2002

34. Monitoring

35. Conclusions • SEQUIN has shown HOW to deploy Premium IP • NRENs are invited to implement it as a replacement of ATM-based MBS • The service provisioning model and debugging procedures need to be further elaborated • Support sought for development of monitoring tools, which is fundamental for the provisioning of the service