Resource Pooling A system exhibits complete resource pooling if it behaves as if there was a single pooled resource. I propose ‘extent of resource pooling’

Slides:



Advertisements
Similar presentations
A DISTRIBUTED CSMA ALGORITHM FOR THROUGHPUT AND UTILITY MAXIMIZATION IN WIRELESS NETWORKS.
Advertisements

Data Center Networking with Multipath TCP
Improving Datacenter Performance and Robustness with Multipath TCP
PATH SELECTION AND MULTIPATH CONGESTION CONTROL BY P. KEY, L. MASSOULIE, AND D. TOWSLEY R02 – Network Architectures Michaelmas term, 2013 Ulku Buket Nazlican.
Restless bandits and congestion control Mark Handley, Costin Raiciu, Damon Wischik UCL.
1 Traffic Engineering (TE). 2 Network Congestion Causes of congestion –Lack of network resources –Uneven distribution of traffic caused by current dynamic.
Technical: what might I mean by multipath? 1.Via an overlay network 2.End-points are multi- homed, and expose multiple IP addresses; routing works as it.
Hot Sticky Random Multipath or Energy Pooling Jon Crowcroft,
1 EL736 Communications Networks II: Design and Algorithms Class8: Networks with Shortest-Path Routing Yong Liu 10/31/2007.
Resource Pooling A system exhibits complete resource pooling if it behaves as if there was a single pooled resource. The Internet has many mechanisms for.
PROMISE: Peer-to-Peer Media Streaming Using CollectCast Mohamed Hafeeda, Ahsan Habib et al. Presented By: Abhishek Gupta.
The War Between Mice and Elephants Presented By Eric Wang Liang Guo and Ibrahim Matta Boston University ICNP
Traffic Engineering With Traditional IP Routing Protocols
Comparing flow-oblivious and flow-aware adaptive routing Sara Oueslati and Jim Roberts France Telecom R&D CISS 2006 Princeton March 2006.
Charge-Sensitive TCP and Rate Control Richard J. La Department of EECS UC Berkeley November 22, 1999.
Network Bandwidth Allocation (and Stability) In Three Acts.
Benefits of coordination in multipath flow control Laurent Massoulié & Peter Key Microsoft Research Cambridge.
Rethinking Internet Traffic Management: From Multiple Decompositions to a Practical Protocol Jiayue He Princeton University Joint work with Martin Suchara,
Dennis Ippoliti 12/6/ MULTI-PATH ROUTING A packet by packet multi-path routing approach.
Combining Multipath Routing and Congestion Control for Robustness Peter Key.
Multipath Protocol for Delay-Sensitive Traffic Jennifer Rexford Princeton University Joint work with Umar Javed, Martin Suchara, and Jiayue He
Multipath Routing Algorithms for Congestion Minimization Ron Banner and Ariel Orda Department of Electrical Engineering Technion- Israel Institute of Technology.
Jennifer Rexford Princeton University MW 11:00am-12:20pm Wide-Area Traffic Management COS 597E: Software Defined Networking.
1 A Distributed Algorithm for Joint Sensing and Routing in Wireless Networks with Non-Steerable Directional Antennas Chun Zhang *, Jim Kurose +, Yong Liu.
MATE: MPLS Adaptive Traffic Engineering Anwar Elwalid, et. al. IEEE INFOCOM 2001.
Second year review Resource Pooling Damon Wischik, UCL.
Grid simulation (AliEn) Network data transfer model Eugen Mudnić Technical university Split -FESB.
Flow Models and Optimal Routing. How can we evaluate the performance of a routing algorithm –quantify how well they do –use arrival rates at nodes and.
DaVinci: Dynamically Adaptive Virtual Networks for a Customized Internet Jennifer Rexford Princeton University With Jiayue He, Rui Zhang-Shen, Ying Li,
The teleology of Internet congestion control Damon Wischik, Computer Science, UCL.
Lecture 15. IGP and MPLS D. Moltchanov, TUT, Spring 2008 D. Moltchanov, TUT, Spring 2015.
Multipath TCP design, and application to data centers Damon Wischik, Mark Handley, Costin Raiciu, Christopher Pluntke.
A Fair and Dynamic Load Balancing Mechanism F. Larroca and J.L. Rougier International Workshop on Traffic Management and Traffic Engineering for the Future.
FAIR CHARGES FOR INTERNET CONGESTION Damon Wischik Statistical Laboratory, Cambridge Electrical Engineering, Stanford
1 A Distributed Algorithm for Joint Sensing and Routing in Wireless Networks with Non-Steerable Directional Antennas Chun Zhang *, Jim Kurose +, Yong Liu.
Models of multipath resource allocation Damon Wischik, UCL.
Congestion control for Multipath TCP (MPTCP) Damon Wischik Costin Raiciu Adam Greenhalgh Mark Handley THE ROYAL SOCIETY.
DaVinci: Dynamically Adaptive Virtual Networks for a Customized Internet Jiayue He, Rui Zhang-Shen, Ying Li, Cheng-Yen Lee, Jennifer Rexford, and Mung.
Some questions about multipath Damon Wischik, UCL Trilogy UCL.
Intradomain Traffic Engineering By Behzad Akbari These slides are based in part upon slides of J. Rexford (Princeton university)
June 4, 2003EE384Y1 Demand Based Rate Allocation Arpita Ghosh and James Mammen {arpitag, EE 384Y Project 4 th June, 2003.
Jennifer Rexford Fall 2014 (TTh 3:00-4:20 in CS 105) COS 561: Advanced Computer Networks TCP.
Network teleology Damon Wischik
We used ns-2 network simulator [5] to evaluate RED-DT and compare its performance to RED [1], FRED [2], LQD [3], and CHOKe [4]. All simulation scenarios.
1 Slides by Yong Liu 1, Deep Medhi 2, and Michał Pióro 3 1 Polytechnic University, New York, USA 2 University of Missouri-Kansas City, USA 3 Warsaw University.
Chapter 11.4 END-TO-END ISSUES. Optical Internet Optical technology Protocol translates availability of gigabit bandwidth in user-perceived QoS.
HELSINKI UNIVERSITY OF TECHNOLOGY Visa Holopainen 1/18.
BUPT COMTEST R&D CENTER Network Coding-based Protection Strategy with Data Traffic Prediction Scheduling Algorithm Shasha Liang
1 Traffic Engineering By Kavitha Ganapa. 2 Introduction Traffic engineering is concerned with the issue of performance evaluation and optimization of.
D. AriflerCMPE 548 Fall CMPE 548 Routing and Congestion Control.
1 Three ways to (ab)use Multipath Congestion Control Costin Raiciu University Politehnica of Bucharest.
Delay-based Congestion Control for Multipath TCP Yu Cao, Mingwei Xu, Xiaoming Fu Tsinghua University University of Goettingen.
Confluent vs. Splittable Flows
Multipath TCP and the Resource Pooling Principle
Constraint-Based Routing
Presented by Tae-Seok Kim
ECE 544: Traffic engineering (supplement)
Improving Datacenter Performance and Robustness with Multipath TCP
Traffic Engineering with AIMD in MPLS Networks
TCP Congestion Control
ISP and Egress Path Selection for Multihomed Networks
LMP calculations Congestion evaluation
AMP: A Better Multipath TCP for Data Center Networks
Uncooperative Flow Control
The Impact of Multihop Wireless Channel on TCP Performance
Flows and Networks Plan for today (lecture 6):
Javad Ghaderi, Tianxiong Ji and R. Srikant
Internet congestion control
Understanding Congestion Control Mohammad Alizadeh Fall 2018
Congestion control and the P2P value chain
Presentation transcript:

Resource Pooling A system exhibits complete resource pooling if it behaves as if there was a single pooled resource. I propose ‘extent of resource pooling’ as a metric for evaluating multipath routing algorithms. Damon Wischik, UCL

Resource pooling is not just a fancy name for load balancing. Load balancing means ‘the load on all resources is roughly equal’ Complete resource pooling means ‘the system behaves as if there was one resource’

I am interested in the behaviour: ‘allocation of flow rates to long-lived flows’ Kelly’s Multipath TCP allocates transmission rates so as to solve an optimization problem I want to know: is this equivalent to solving a single-resource optimization problem? If so, there is complete resource pooling.

I am interested in the behaviour: ‘allocation of flow rates to long-lived flows’ Kelly’s Multipath TCP allocates transmission rates so as to solve an optimization problem I want to know: is this equivalent to solving a single-resource optimization problem? If so, there is complete resource pooling.

What is the behaviour when there is complete resource pooling? If there is complete resource pooling, The network’s total capacity is shared fairly between flows Adjusting the capacity of a single resource will rebalance traffic over all the other resources A localized surge in traffic will distribute itself over the pool

Resource pooling does not imply balanced load nor equal flow rates Resource pooling does not imply balanced load nor equal flow rates. This is a good thing. Here, three resources behave like a single pool, but one is more congested Flow y1 takes twice as much from the pool as y2

Sometimes the pool is made up of only some of the resources — the bottleneck cut [See Laws, Resource pooling in queueing networks with dynamic routing, 1992, Example 3.2. He finds generalized cut constraints of this form for a version of this problem with fixed capacity constraints and demands.]

Typically the network will be split into several pools Typically the network will be split into several pools. This is called partial resource pooling. Partial resource pooling is a sign that there aren’t enough flows with the right path diversity Note that the resource pools depend on routing & capacity & traffic. With different traffic loads, there will be different pools.

To improve resource pooling, we need better path diversity. To achieve better resource pooling, we need to give better path choices This should shift traffic onto the less congested pool If enough traffic can shift, the pools will merge

How to find the resource pools Create a multipath routing Measure the total rate ys for each flow s and the drop probabilities pj at each link j I propose to find these by running the fluid model of mTCP Find subsets of resources where y and p satisfy a certain relationship; these are the resource pools

What help do I need to get started? What do I want to achieve? I need a sensible topology to experiment on I need a sensible multipath routing algorithm I need a sensible traffic demand matrix I would like a plausible picture of p2p traffic The resource pool algorithm works just as well with mTCP as with a suitable congestion-sensitive p2p protocol I want a simulator which will identify the pools We can then see where the multipath routing algorithm has not done very well, and see which extra paths it should give Maybe we’ll find that p2p traffic is sufficient to achieve pretty good resource pooling! (with appropriate congestion control and peer choice) Probably, we only need the large flows to have path diversity, since they contribute most of the bytes on the Internet.