IEEE/ACM TRANSACTIONS ON NETWORKING, VOL. 12, NO. 6, DECEMBER 2004 1105 Segment Shared Protection in Mesh Communications Networks With Bandwidth Guaranteed Tunnels Pin-Han Ho, Member, IEEE, János Tapolcai, and Tibor Cinkler, Member, IEEE Abstract—This paper focuses on the problem of dynamic sur- vivable routing for segment shared protection (SSP) in mesh communication networks provisioning bandwidth guaranteed tunnels. With SSP, a connection is settled by concatenating a series of protection domains, each of which contains a working and protection segment pair behaving as a self-healing unit for per- forming local restoration whenever the working segment is subject to any unexpected interruption. We first discuss the advantages of using SSP—the ability to shorten the restoration time as well as achieve a higher throughput by saving spare capacity required for 100% restorability; then the survivable routing problem is formulated into an Integer Linear Programming (ILP), where the switching/merging node pair of each protection domain along with the corresponding least-cost working and protection segment pair can be jointly determined for a dynamically arrived connec- tion request. A novel approach of arc-reversal transformation is devised to deal with the situation that the working segments of two neighbor protection domains may overlap with each other by more than a single node. Due to a very high computation complexity induced in solving the ILP, a novel heuristic algorithm is proposed, named Cascaded Diverse Routing (CDR), to allocate protection domains for a connection request by performing diverse routing across a set of predefined candidate switching/merging node pairs. Experiments are conducted on five two-connected network topologies to verify the ILP and the CDR algorithm. We first determine the best diameter of protection domains for the CDR scheme in each network topology. Using the results of best diameters, CDR is compared with two reported schemes, namely PROMISE and OPDA. We demonstrate in the simulation results that the path-shared protection schemes are outperformed by the SSP schemes in terms of blocking probability under all possible arrangements in the experiment and that CDR yields better performance than PROMISE and OPDA due to the extra efforts in manipulating the location of working segments at the expense of longer computation time. Index Terms—Integer linear programming (ILP), segment shared protection (SSP), shared risk link group (SRLG), surviv- able routing, working and protection paths. I. INTRODUCTION S URVIVABILITY has emerged as one of the most important issues in the design of modern communications networks with bandwidth guaranteed tunnels. Survivable routing is rec- Manuscript received April 13, 2003; revised September 22, 2003; approved by IEEE/ACM TRANSACTIONS ON NETWORKING Editor A. Orda. P.-H. Ho is with the Electrical and Computer Engineering Department, Uni- versity of Waterloo, Waterloo, ON N2L 3G1 Canada (e-mail: pinhan@bbcr. uwaterloo.ca). J. Tapolcai is with High-Speed Networks Laboratory, Department of Com- puter Science and Information Theory, Budapest University of Technology and Economics, Budapest H-1117, Hungary (e-mail: janos@cs.bme.hu). T. Cinkler is with the Department of Telecommunications and Media Infor- matics, Budapest University of Technology and Economics, Budapest H-1117, Hungary (e-mail: cinkler@tmit.bme.hu). Digital Object Identifier 10.1109/TNET.2004.838592 ognized as one of the best strategies to equip the networks with service continuity by preplanning link-disjoint or node-disjoint protection paths for working capacity. With a diversely routed working-protection path pair, once the working path is subject to any unexpected interruption, the corresponding service can be restored by switching over the working traffic to the protection path such that the failure is circumvented. With the emergence of some commercially applications and delay-sensitive services addressing stringent requirements on data integrity and service continuity, the design of survivable routing algorithms should not only be both capacity- and computation-efficient, but also minimize the possible restoration time for a specific connection, such that the maximum benefits can be gained in the operation of carrier networks. Segment shared protection (SSP) is one of the best ap- proaches to meet the above design requirements, where a connection is provisioned by concatenating a series of protec- tion domains, each of which contains a working and protection segment pair behaving as a self-healing unit for performing local restoration when the working segment is subject to any unexpected interruption. As shown in Fig. 1, when the working path segment of protection domain 2 is impaired unexpectedly (e.g., either link E-F, F-G, G-H, or H-I is cut), the restoration is performed locally within protection domain 2 such that the affected flow switches over to the backup segment at node E (called switching node of the protection domain) and merges back to the original working path at node J (called merging node of the protection domain). Comparing with its counterpart—path shared protection [10]–[18], SSP has been reported to achieve a better throughput by maximizing the extent of spare capacity resource sharing [2], [7], [21]. It can also impose a stringent limitation on the restoration time for a specific application by constraining the length/hop count of the working and protection segment in each protection domain. The major difficulties of implementing SSP lie in the dependency between the working and spare capacities as well as the exponentially enlarged design space with the network size in identifying a set of switching/merging node pairs to form protection domains for a connection request. Thus, most previous studies focus on heuristic approaches to solve the problem. In [2], a framework known as Short Leap Shared Protection (SLSP) is proposed, which implements SSP by preassigning a series of switching/merging node pairs along a given working path. The capacity efficiency with SLSP is expected to be further improved if the working and backup segments can be jointly determined. In [3], an algorithm is developed to find the working path first followed by its backup path segments. The study is characterized by the fact that the 1063-6692/04$20.00 © 2004 IEEE