1 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY ISP Backbone Traffic Inference Methods to Support Traffic Engineering Olivier Goldschmidt Senior Network Consultant
2 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Outline 1. Problem Description 2. Inputs to the Models 3. Constraints of the Models 4. Inference Methods: Pseudo-Inverse Method Linear Programming 5. Test Results 6. Conclusion and Open Issues
3 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY RATIONALE A major headache for Internet Service Providers is to estimate the end-to-end traffic volumes on their backbone network. Reliable traffic estimates between ingress and egress points are essential to traffic engineering purposes such as ATM PVC or LSP layout and sizing.
4 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Problem Description An "easy" solution is to turn on NetFlow or IP-Accounting on all ingress and egress interfaces. But such solution is - Costly - Impractical
5 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Problem Description Objective of traffic inference is to "guess" end to end aggregate traffic using limited information.
6 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Inputs to the model Deterministic Information Measured Information Usage Information
7 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY DETERMINISTIC INFORMATION Network Topology Types of routers and links Routing paths between end points
8 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY MEASURED INFORMATION Baselining Information on network interfaces using SNMP Partial RMON/RMON2 information using selective probes (NetFlow or IP account.)
9 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY USAGE INFORMATION Data that can be correlated with the traffic on the network Allows to derive additional constraints on the network traffic.
10 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Ingress-Egress points Internal routers WAN Link
11 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Assume that reading are symmetric. Interface flow reading
12 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY 3 3
13 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY
14 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY CONSTRAINTS
15 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY
16 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY PSEUDO-INVERSE METHOD
17 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY LINEAR PROGRAMMING METHOD
18 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY OBJECTIVE FUNCTION COEFFICIENTS Choice of coefficients for the objective function will determine the precision of the end to end traffic estimates. Obvious choice is to set all coefficients to 1 and to maximize or to minimize the objective function But this choice is not neutral
19 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY EXAMPLE Assume these are the true traffic demands 10 Notice that all interface flows are equal to 20
20 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY If all objective coefficients are equal to 1 If objective function is maximized 0 20 If objective function is minimized
21 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY But if coefficient are equal to the number of hops of demand route Is a solution whether objective function is maximized or minimized
22 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Another advantage of the LP method Allows to add constraints that represent usage information. For instance constraint the very unlikely end-to-end traffic to be close to zero. Also known traffic from NetFlow or IP accounting readings can be included as constraints in the linear program.
23 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Test Results 60 Routers 114 WAN Links 529 Traffic demands Bandwidth from 0 to 256 Kbps NETWORK
24 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Test Results 1. Route the demands 2. Compute the resulting interface flows 3. Apply the Linear Programming method to estimate the end-to-end traffic demands 4. Compare those estimates with the original traffic demands in % of absolute difference |estimate-true value|/true value The following charts show % of demands with given relative error
25 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Objective coefficients = number of hops
26 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Objective coefficients = number of hops
27 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY All objective coefficient = 1
28 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Netflow turned on on five random routers
29 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Netflow turned on on five most used routers
30 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Netflow turned on on ten random routers
31 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Comparison of different results
32 ISMA Backbone Traffic Inference MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Conclusions Objective coefficients in LP need to be scaled Turning NetFlow on a few selected interfaces can greatly improve the traffic estimates.