security and error correction

1. Security and Error Correction/Detection in 802.1x and GSM Routing and Traffic Engineering in Generalized Multi-Protocol Label Switching (GMPLS)

2. 2

3. 3 Problem Description Due to increased adoption of high speed access technology - Convergence of IP and optical networks requires: Handling the traffic due to increased Internet-based services. Guaranteed quality of service (QoS) IP is a connectionless, best-effort technology that was not designed for voice or any other real-time service IP traffic is unpredictable and unstable and hence demands: Bandwidth management and service requirements for next generation Internet Protocol (IP) based backbone networks Flexibility and ability to promptly react to traffic changes Overprovisioning used in telecom networks is not a cost effective solution to the problem of unpredictable bottlenecks Migration of all services over IP requires Traffic Engineering in Generalized Multi-Protocol Label Switching (GMPLS) for new generation of multilayer networks.

4. 4 Problem with Current Routing Approach Traffic entering a network can vary with time , both in predictable and unpredictable ways e.g. traffic carrying traditional telephone services is easily predictable using well known models while Internet traffic is not Telephone services traffic uses off-line routing approach for achieving global optimization of route calculation. Optimization of network resources requires long computational time, which increases with the network and traffic size Pure off-line approach can result unsatisfactory in case of Internet traffic. Overprovisioning is not a cost effective solution A pure-online routing approach Evaluation the routes �on-demand� But does not lead to efficient use of the network resources � no global optimization

5. 5 Proposed Solution Motivation: Traffic characteristics and practical implementation aspects related to routing Integrated Traffic Engineering (TE) system for new generation multilayer networks based on GMPLS: which dynamically reacts to traffic changes fulfills quality of service (QoS) requirements for different classes of service Hybrid Routing Approach makes use of both off-line and on-line methods to accommodate traffic requests Bandwidth Management allows elastic use of the bandwidth among different data flows based on their priorities allows pre-emption and rerouting

6. 6 Introduction to GMPLS Future data and transmission networks will consist of Routers and Switches DWDM systems Add-drop multiplexer (ADMs) Photonic cross-connect (PXCs) Optical cross-connect (OXCs)

7. 7 Label Extension:Label Switched Path (LSP) Hierarchy GMPLS allows a single control plane to handle heterogeneous Labeled Switched Paths (LSPs) A single instance of the control plane can span multiple technologies LSP of low order can be tunneled into an already existing LSp of higher order that acts as a link.

8. 8 Evolution of MPLS to GMPLS MPLS works for packet switching GMPLS includes packet, time slots, wavelengths, physical ports for switching Interface Packet Switch Capable (PSC) interface Layer-2 Switch Capable (L2SC) Time-Division Multiplex Capable (TDM) Lambda Switch Capable (LSC) Fiber-Switch Capable (FSC) GMPLS extends the MPLS-TE protocols except LMP (Link Management Protocol) Signaling : RSVP-TE and CR-LDP Routing : OSPF-TE and IS-IS-TE LMP Extremely many parallel links between neighboring nodes GMPLS uses IP for control channel

9. 9 LSP Hierarchy LSP Hierarchy Packet Switches (PS) LSP : Label ? virtual circuit Time Division Switches (TDM) LSP: Label ? time slot Wavelength Switch (WS) LSP: Label ? wavelength Fiber Switch (FS) LSP : Label ? fiber

10. 10 GMPLS Key Extensions to MPLS-TE Various link types and payload FA (Forwarding Adjacency) Label Suggestion Label Restriction Bidirectional LSP setup Rapid Failure Notification

11. 11 Scalability Enhancements Forwarding Adjacencies (FA) Unnumbered links Links (or interfaces) that do not have IP addresses Link bundling Link Management Protocol

12. 12 Forwarding Adjacencies Nested LSP (LSP inside LSP) for scalability enhancements LSP Hierarchy No need of forwarding state for inside LSP advertise an LSP as a link ? FA Aggregation LSR creating a TE LSP LSR forming a FA out of that LSP Allowing other LSRs to use FAs for path computation Nesting of LSPs originated by other LSRs into that LSP

13. 13 Link bundling Link bundling When a pair of LSRs is connected by multiple links, it is possible to advertise several (or all) of these links as a single link into OSPF and/or IS-IS Information aggregation/abstraction Restriction Same pair of LSRs Share common characteristics Modification of Routing and Signaling Unnumbered bundled link

14. 14 Link Management Protocol LMP: runs between neighboring nodes and is used to manage TE links Collection of useful procedures between adjacent nodes Control channel management Establish and maintain control channel Negotiation by Hello protocol and keep alive Link property correlation Aggregate links into bundled links or Exchange, correlate, or change TE link parameters Link connectivity verification Verify the physical connectivity of data links Fault management Fault detection Fault localization Fault notification

15. 15 Traffic Engineering The main goals of TE in new generation networks are: Optimization of the use of network resources Realization of the �Bandwidth-on-Demand� concept Support of different classes of service by guaranteeing the required QoS GMPLS is adopted by IETF to provide flexibility and efficiency in the use of network resource GMPLS can take advantage Suitable extension of signaling protocol to allow the reservation of network resources Priority mechanisms to assign resources to higher priority LSPs at the expense of lower priority LSPs to provide efficient QoS Previous works done by D. Papadimitriou et.al and Jong T. Park deal with specific TE functions such as routing, wavelength assignment, and preemption algorithms in an optical layer. Their work was based on solving the following problems separately : Design of logical topology of the optical network Routing of the data flows at the IP/MPLS layer

16. 16 Hybrid Routing Approach The proposed hybrid routing solution combines the offline and the on-line method In the presence of more than one class of service, the flexibility can be enhanced by means of the module called bandwidth engineering (BE) Two main groups of LSPs are identified: Higher priority (HP) LSPs - The HP traffic carries presumably mission-critical services, such as voice and video communications. Lower priority (LP) LSPs - The traffic carried by LP LSPs can be of various types of possible Internet services, with different QoS requirements.

17. 17 Off-Line Routing The TE system is designed to serve on-demand both HP and LP traffic, with a difference: HP traffic routes, once calculated, remain fixed during their life unless another off-line procedure is activated LP traffic routes can be dynamically changed from their originally assigned routes, according to the actual network status Off-Line Routing: The Global Path Provisioning (PR Module) Based on statistical estimation of traffic pattern

18. 18 On-Line Routing On-Line Routing: The Dynamic Path Selection Evaluation of �on-line� route for the entering LP LSP request based on: source and destination nodes bandwidth requirements, taking into account the updated link state status of MPLS and WDM layers The DR module aims at better utilizing network resources, by using less-congested paths instead of shortest, but heavily loaded paths. Two criteria for DR Module finding the shortest route that minimizes congestion, evenly distributing the traffic at MPLS layer selecting the lightpath in the logic link The two criteria can be fulfilled by using a shortest path algorithm with a weight function.

19. 19 Hybrid Routing To facilitate the integration of the PR and DR an opportune flexibility factor, a �[0,1] is introduced during the provisioning phase. scaling the physical topology during the off-line procedure by reducing the bandwidth of each wavelength Hence DR operates on an enforced topology, at the expense of an increment of physical network resources utilization The factor a depends on the network load the relationship between the expected and the actual traffic e.g. If the wavelength capacity is bw , the wavelength bandwidth used during the PR procedure, is limited to abw While during the DR operation those lightpaths, constituting the logical topology, are considered with their actual bandwidth, i.e., bw

20. 20 Bandwidth Engineering BE functions operate so that the temporarily unused reserved bandwidth of a higher priority LSP can be released and put at disposal of lower priority requesting LSPs. BE makes use of two key elements Bandwidth handling algorithm (BHA), which selects those LP LSPs that need to be moved to make available the bandwidth required by the HP LSPs Previously mentioned DR algorithm, which aims at rerouting those selected LP LSPs on alternatives paths.

21. 21 Expected Results The described TE system is based on two key novel components a hybrid routing scheme a system able to handle priority, preemption, and rerouting, called bandwidth engineering. The former allows an optimization of the use of the network resources an improvement of the dynamic performance of the network and the robustness against traffic unpredictability. The latter improves the performance of the network by achieving an elastic use of the bandwidth, so that the temporarily unused bandwidth by HP traffic is not wasted, but put at disposal of LP traffic. As a result, the proposed TE Guarantees QoS requirements to be fulfilled Optimization of network resources use, Increases the flexibility of the network Allows a large amount of traffic to be accommodated

22. 22 Simulations to be done In order to assess the effectiveness of the considered TE system, the performances have to be evaluated for three case studies: Case 1: the current traffic distribution and volume do not appreciably differ from the ones predicted by the traffic matrix Case 2: the traffic volumes are the same, but the current traffic distribution appreciably differs from the predicted one Case 3: the current traffic volume and distribution appreciably differ from the predicted ones.

23. 23 Thanks a lot �

24. 24