* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download MethodNRPTM
Survey
Document related concepts
Asynchronous Transfer Mode wikipedia , lookup
Computer network wikipedia , lookup
Cracking of wireless networks wikipedia , lookup
Distributed firewall wikipedia , lookup
Deep packet inspection wikipedia , lookup
List of wireless community networks by region wikipedia , lookup
Transcript
Backbone Traffic Inference ISP Backbone Traffic Inference Methods to Support Traffic Engineering Olivier Goldschmidt Senior Network Consultant ISMA 2000 1 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference 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 ISMA 2000 2 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference 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. ISMA 2000 3 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference 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 ISMA 2000 4 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference Problem Description Objective of traffic inference is to "guess" end to end aggregate traffic using limited information. ISMA 2000 5 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference Inputs to the model Deterministic Information Measured Information Usage Information ISMA 2000 6 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference DETERMINISTIC INFORMATION Network Topology Types of routers and links Routing paths between end points ISMA 2000 7 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference MEASURED INFORMATION Baselining Information on network interfaces using SNMP Partial RMON/RMON2 information using selective probes (NetFlow or IP account.) ISMA 2000 8 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference USAGE INFORMATION Data that can be correlated with the traffic on the network Allows to derive additional constraints on the network traffic. ISMA 2000 9 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference WAN Link Ingress-Egress points Internal routers ISMA 2000 10 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference 3 3 Assume that reading are symmetric. 3 3 Interface flow reading ISMA 2000 11 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference 3 3 ISMA 2000 12 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference 1 2 1 2 ISMA 2000 13 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference CONSTRAINTS ISMA 2000 14 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference 3 3 2 1 1 3 ISMA 2000 2 3 15 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference PSEUDO-INVERSE METHOD ISMA 2000 16 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference LINEAR PROGRAMMING METHOD ISMA 2000 17 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference 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 ISMA 2000 18 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference EXAMPLE 10 10 10 Assume these are the true traffic demands Notice that all interface flows are equal to 20 ISMA 2000 19 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference If all objective coefficients are equal to 1 0 20 20 If objective function is maximized 20 0 0 If objective function is minimized ISMA 2000 20 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference But if coefficient are equal to the number of hops of demand route 2 1 10 10 1 10 Is a solution whether objective function is maximized or minimized ISMA 2000 21 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference 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. ISMA 2000 22 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference Test Results NETWORK • 60 Routers • 114 WAN Links • 529 Traffic demands • Bandwidth from 0 to 256 Kbps ISMA 2000 23 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference 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 ISMA 2000 24 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference Objective coefficients = number of hops % demands 12 10 8 6 4 2 0 5 20 35 ISMA 2000 50 65 80 95 0 11 5 12 0 14 5 15 0 17 5 18 0 20 25 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference Objective coefficients = number of hops % demands (cumul) 120 100 80 60 40 20 0 5 20 ISMA 2000 35 50 65 80 95 0 1 1 5 2 1 0 4 1 5 5 1 0 7 1 5 8 1 0 0 2 26 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference All objective coefficient = 1 % demands 70 60 50 40 30 20 10 0 5 20 35 ISMA 2000 50 65 80 95 0 11 5 12 0 14 5 15 0 17 5 18 0 20 27 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference Netflow turned on on five random routers % demands 35 30 25 20 15 10 5 0 5 20 ISMA 2000 35 50 65 80 95 0 11 5 12 0 14 5 15 0 17 5 18 0 20 28 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference Netflow turned on on five most used routers % demands 70 60 50 40 30 20 10 ISMA 2000 20 0 18 5 17 0 15 5 14 0 12 5 11 0 95 80 65 50 35 20 5 0 29 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference Netflow turned on on ten random routers % demands 70 60 50 40 30 20 10 ISMA 2000 20 0 18 5 17 0 15 5 14 0 12 5 11 0 95 80 65 50 35 20 5 0 30 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference Comparison of different results 70 60 50 40 30 20 10 0 5 10 15 20 25 30 35 40 45 coeff 1 ISMA 2000 50 # hops 55 60 rand net 65 70 75 80 85 90 95 100 opt net 31 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY Backbone Traffic Inference Conclusions Objective coefficients in LP need to be scaled Turning NetFlow on a few selected interfaces can greatly improve the traffic estimates. ISMA 2000 32 MAKE SYSTEMS THE NETWORK RESOURCE PLANNING COMPANY