A Flow Aggregation Method Based on End-to-End Delay in SDN

Transcription

1 A Flow Aggregation Method Based on End-to-End Delay in SDN Takuya Kosugiyama, Kazuki Tanabe, Hiroki Nakayama, Tsunemasa Hayashi and Katsunori Yamaoka Tokyo Institute of Technology, Japan Bosco Technologies Inc., Japan

2 Overview Background Goal Modeling Algorithm Result Summary

3 What is SDN? Software-Defined Networking (SDN) Controllable via API Application Plane Business Application Control Plane SDN Controller Northbound API Network Service Data Plane Network Device Southbound API Network Device Network Device Network Device

4 Problem Routing Mechanism in OpenFlow (OF) 1. Proactive. Reactive OF Controller OF Controller 1 Configure rule dst IP: A Port: Packet comes To: A 1 Forwarding OF Switch Request rule Configure rule 1 Packet comes dst IP: A Port: To: A 1 OF Switch Forwarding [1] S.H. Yeganeh, A. Tootoonchian, and Y. Ganjali, On scalability of software-defined networking, IEEE Communications Magazine, vol.51, no., pp.16 11, Feb. 01.

5 Problem 5 Routing Mechanism in OpenFlow (OF) 1. Proactive. Reactive OF Controller OF Controller 1 Configure rule dst IP: A Port: Packet comes To: A 1 Forwarding Request rule Configure rule 1 Packet comes dst IP: A Port: To: A 1 Forwarding OF Switch Rule update rate (1000 rules/s) Rule capacity (10 50 K) [1] OF Switch SDN scaling issues [1] S.H. Yeganeh, A. Tootoonchian, and Y. Ganjali, On scalability of software-defined networking, IEEE Communications Magazine, vol.51, no., pp.16 11, Feb. 01.

6 Goal 6 How to manage massive flows in SDN Aggregate flows and reduce number of flows Rule aggregation method based on bandwidth [1][] Rule division method to satisfy capacity [][] There are no works to aggregate flows based on End-to-End allowable delay Target: IoT / MM flow A flow aggregation method to minimize number of flows satisfying allowable delay [1] F. Giroire, J. Moulierac, and T. K. Phan. Optimizing rule placement in software-defined networks for energy-aware routing. In 01 IEEE Global Communications Conference, pp. 5 59, Dec 01. [] Xuan Nam Nguyen, Damien Saucez, Chadi Barakat, and Thierry Turletti. OFFICER: A general optimization framework for OpenFlow rule allocation and endpoint policy enforcement. Proceedings - IEEE INFOCOM, Vol. 6, pp , 015. [] Yossi Kanizo, David Hay, and Isaac Keslassy. Palette: Distributing tables in software-defined networks. Proceedings - IEEE INFOCOM, pp , 01. [] Nanxi Kang, Zhenming Liu, Jennifer Rexford, and David Walker. Optimizing the One Big Switch Abstraction in Software-Defined Networks. Conext 1, p. 17, 01.

7 Approach 7 Basis of flow aggregation Consider flows on same section as one flow f1 f1 f 1 f Tunnel and labeling f Link cost Route by aggregated flow ex) Full mesh flows (allowable delay = 80) flow: = 100 > 80 (a) Network (b) Maximum aggregation not satisfying allowable delay

8 Approach 8 Basis of flow aggregation Consider flows on same section as one flow f1 f1 f 1 f Tunnel and labeling f Route by aggregated flow ex) Full mesh flows (allowable delay = 80) Link cost (a) Network (b) Maximum aggregation not satisfying allowable delay (c) Maximum aggregation satisfying allowable delay

9 Model setting (1) 9 SDN as a directed graph Node: Link: Link cost: Flow: Allowable cost: Relation between flow and link

10 Model setting () 10 Constraints to route flow There is only one flow from a source node and to a destination node

11 Model setting () 11 Constraints to route flow There is no flow to a source node and from a destination node

12 Model setting () 1 Constraints to route flow Number of incoming flows equals to number of outgoing flows on a node

13 Model setting () 1 Constraints to route flow A node is visited only one time

14 Model setting () 1 Constraint of allowable cost Aggregate flows Flows on same link and same direction Objective: minimize number of flows

15 Model setting () 15 Deformation y and z are vectors (0, 1,,0) which correspond to the index of h and g Apply to 0,1-Knapsack Problem NP-hard

16 Heuristics 16 Overview 1. Ascending sort by (minimum cost - allowable cost) Fs. Select the first element of Fs f. [Flow compotision] Compose route of f using existing flows. [Flow aggregation] Aggregate flows in Fs into existing flows 5. Return to. unless Fs is empty = flexibility of changing route Existing flows Flow composition Flow aggregation

17 Flow composition 17 Path composition by Best-First Search (BFS) Evaluation value of node : : number of times existing rules are used : minimum cost to reach from to Prioritize two nodes Time complexity:

18 Flow composition: Ex.1 18 Flow composition by Best-First Search Evaluation value: {t, c} t : number of times existing rules are used c : minimum cost to reach from source to current 5 {0,0} 5 {0,5} {0,} Dashed arrows: existing flows

19 Flow composition: Ex.1 19 Flow composition by Best-First Search Evaluation value: {t, c} t : number of times existing rules are used c : minimum cost to reach from source to current 5 {0,0} 5 {1,5} {0,} {0,5} Dashed arrows: existing flows

20 Flow composition: Ex.1 0 Flow composition by Best-First Search Evaluation value: {t, c} t : number of times existing rules are used c : minimum cost to reach from source to current 5 {0,0} 5 {1,5} {0,} {1,8} {,9} Dashed arrows: existing flows

21 Flow composition: Ex.1 1 Flow composition by Best-First Search Evaluation value: {t, c} t : number of times existing rules are used c : minimum cost to reach from source to current 5 {0,0} 5 {1,5} {0,} {1,8} {,9} {,1} Dashed arrows: existing flows Number of adding flows:

22 Flow composition: Ex. Flow composition by Best-First Search 5 {0,0} 5 {1,5} {0,} {1,8} {,9} {,1} > Dashed arrows: existing flows

23 Flow composition: Ex. Flow composition by Best-First Search 5 {0,0} 5 {1,5} {0,} {1,8} {,9} Dashed arrows: existing flows

24 Flow composition: Ex. Flow composition by Best-First Search 5 {0,0} 5 {1,5} {0,} {1,8} {,9} {1,10} Dashed arrows: existing flows Number of adding flows:

25 Heuristics 5 Overview 1. Ascending sort by (minimum cost - allowable cost) Fs. Select the first element of Fs f. [Flow compotision] Compose route of f using existing flows. [Flow aggregation] Aggregate flows in Fs into existing flows 5. Return to. unless Fs is empty Existing flow Flow composition Flow aggregation f (f1, f) f1 f

26 Flow aggregation 6 Delete flows in Fs over existing flows Calculate flows can be reached within allowable delay by Dijkstra s algorithm Time complexity: f1 f Flow aggregation f (f1, f ) f1 f f can be routed using part of f1 and f

27 Simulation settings 7 Simulation Topology Model: ER / BA Size: V = 50, 500 Parameters Link cost: [5, 15], discrete uniform distribution alpha: ratio of allowable cost of flows to the maximum value of the minimum delay of flows Ex) min cost = 10, alpha = Cf = 0 Comparison Table: topologies Minimum cost / Minimum hop routing with aggreagtion

28 Performance evaluation (1) 8 Effect of allowable cost F = 1000 There is not much effect of alpha Fig. 1 Effect of allowable cost of flows (Scalefree-small)

29 Performance evaluation (1) 9 Effect of allowable cost F = 1000 Flow composition selects first route No more effect of allowable delay Fig. 1 Effect of allowable cost of flows (Scalefree-small)

30 Performance evaluation () 0 Effect of number of flows α = 5 HIGH link density more effective Fig. Random-large Fig. Scalefree-large

31 Summary & Future work 1 We propose a flow aggregation method to minimize number of flows satisfying End-to-End delay Simulation in several topologies Higher performance than simple comparison method Flexibility of changing route is important Future work Expand our model more realistic Bypassing on weak topology