Traffic Engineering in Segment Routing Networks Using MILP

• Published in: IEEE eTransactions on network and service management Vol. 17; no. 3; pp. 1941 - 1953
• Main Authors: Li, Xiaoqian, Yeung, Kwan L.
• Format: Journal Article
• Language: English
• Published: New York IEEE 01-09-2020; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Computer architecture; Integer programming; IP networks; Linear programming; load balancing; Load management; Mixed integer; Mixed integer linear programming; Multiprotocol label switching; Routing; segment routing; Shortest-path problems; Traffic control; Traffic engineering
• ISSN: 1932-4537; 1932-4537
• DOI: 10.1109/TNSM.2020.3001615

In segment routing, a packet is forwarded along a path identified by a segment list. A segment list consists of segment identifiers (SIDs). A node-SID identifies a shortest-path segment, and an adjacency-SID identifies a link segment. A <inline-formula> <tex-math notation="LaTeX">{K} </tex-math></inline-formula>-segment path is a path with no more than <inline-formula> <tex-math notation="LaTeX">{K} </tex-math></inline-formula> segments. In this paper, we study the problem of finding a set of K- segment paths to carry all the flows in a given traffic matrix such that the maximum link utilization in the network is minimized. We first show that the solutions found by <inline-formula> <tex-math notation="LaTeX">{K} </tex-math></inline-formula>-LP, an existing linear programming (LP) approach, are not optimal because <inline-formula> <tex-math notation="LaTeX">{K} </tex-math></inline-formula>-LP does not support adjacency-SIDs. Focusing on 2-segment paths, a new LP formulation, denoted by e2-LP, is designed to support adjacency-SIDs in part. To fully support adjacency-SIDs, a mixed integer linear programming (MILP), denoted by <inline-formula> <tex-math notation="LaTeX">{K} </tex-math></inline-formula>-MILP, is designed. Since solving <inline-formula> <tex-math notation="LaTeX">{K} </tex-math></inline-formula>-MILP is time-consuming, a simplified version (<inline-formula> <tex-math notation="LaTeX">{K} </tex-math></inline-formula>-sMILP) is also proposed. Finally, <inline-formula> <tex-math notation="LaTeX">{K} </tex-math></inline-formula>-sMILP is extended to prevent excessive flow splitting or using paths that are too long.