Participants University of Wrzburg Thomas Zinner, Dominik Klein, - - PDF document

Institute of Computer Science Chair of Communication Systems

• Prof. Dr.-Ing. P. Tran-Gia

MultiNext: Measuring Concurrent Multipath Transmissions in an Experimental Facility

• T. Zinner (Uni Würzburg), K. Tutschku (Uni Vienna), T. Zseby (Fraunhofer)

7th EURO-NF CONFERENCE ON NEXT GENERATION INTERNET NGI 2011

2

Thomas Zinner

Participants

University of Würzburg Thomas Zinner, Dominik Klein, Christian Schwartz, Daniel Schlosser, Phuoc Tran-Gia University of Vienna Kurt Tutschku, Albert Rafetseder Fraunhofer Fokus Tanja Zseby, Carsten Schmoll, Christian Henke (TU Berlin) Tel Aviv University Yuval Shavitt http://www3.informatik.uni-wuerzburg.de/research/projects/multinext/

3

Thomas Zinner

Agenda

Aims of the project MultiNext Architecture and transport mechanisms Federation concept Concurrent multipath transmission Measurements and results Comparision of different testbeds Active and passive measurements Lessons learned and conclusion

4

Thomas Zinner

Aims of the MULTINEXT Project

Gain experience in the usability of federated testbeds

• Possibility to setup a (larger scale) routing overlay

to emulate multipath transport

• Large distributed set of nodes to get a high QoS

diversity Federated measurements with the most appropriate measurement equipment

• Advanced measurement methods for high

precision and hop-by-hop one-way delay measurements Validate a performance model for multipath transport in a Federated Future Internet

5

Thomas Zinner

ARCHITECTURE AND TRANSPORT MECHANISM

6

Thomas Zinner Backbone #2 Backbone #2 Backbone #1 Backbone #1

Compute Cluster #2 Compute Cluster #1

Application-specific Slice

Application which coordinates the slice Quality islands, e.g. obtained by DiffServ or Overprovisioning

ISP 1 ISP 2

Application

ISP 3 ISP 4 ISP 5 ISP 6 ISP 7

Compute Cluster #3

Enhanced Federation Concept (Tutschku, 2009, Geni)

Federation: Selection of co-operative resources (networks, links, nodes) for the life time of application-specific virtual network (aka !slice") Medium variability of slice (not flow-by-flow not static, not VPN-like) Account for limited life-time and variable performance of resources Challenge: optimal resource selection needs to measure temporal behavior of e.g. churn of resource networks or variability of E2E delay

7

Thomas Zinner

Concurrent Multipath Transmission

Concurrent Multipath Transmission: Resource pooling: $ Parallel usage of resources $ Forming a virtualized single !transport resource" Aiming for: Improved resilience High capacity (e.g. for video streaming) New business models e.g. $ Combine multiple operators $ Dynamic resource selection

8

Thomas Zinner

Concurrent Multi-Path Transfer

Superset of available physical resources Routing topologies

• f provider I

Routing topology

• f provider II

Aim: Very high and reliable throughput between two end hosts pooled transport pipe

POP Logical topology Aim: Very high and reliable transmission between two end hosts Solution: Transport Virtualization: Combine multiple paths (even from different providers)

9

Thomas Zinner

Performance Parameters

Input: Number of paths Scheduling Output: Steady-state re-sequencing buffer occupancy distribution Path delay distributions Path capacity

Source Destination

delay distribution delay distribution buffer

• ccupancy

10

Thomas Zinner

Each time unit, a packet is transmitted over each path Constraint: packets do not overtake each other on a single path, i.e. current delay previous delay $ interdeparture time But: packets overtake each other on other paths:

Assumptions

path 1 destination path 2 packets in re-sequencing buffer

2 3 1 5 4 6 2 2 4 4 2 1 4 3 6 5

% %

time

11

Thomas Zinner

MEASUREMENTS AND RESULTS

12

Thomas Zinner

Experimental Facilities

Feature G-Lab PLE PLC VINI Scope Germany Europe World Mainly US Exclusive Reservation Yes No No No Routing with own tools with own tools with own tools Yes Bandwidth and QoS with own tools Planned Planned Yes (service) Openess/ Federation No (tests planned) Yes (SFA) Yes (SFA) Yes (SFA) Tools/Packet Tracking individually Yes (service) individually individually Clock Sync NTP NTP, some GPS NTP NTP

Investigation of different testbeds Find most suitable testbed w.r.t. experimental requirements

13

Thomas Zinner

Measurement Setup

Active measurements ETOMIC boxes at end points Application layer overlay enables routing Passive measurements Measurements with multi-hop packet tracking VINI overlay routing

14

Thomas Zinner

Multi-hop Packet Tracking

15

Thomas Zinner

Active Measurements – Results

Varying one way delays on the paths Higher delays via Tromso Good fit between model and measurements

16

Thomas Zinner

Passive Measurements – Results

Variable path delay via VINI (due to an additional queue) Constant delay via GLAB path Good fit between model and measurements

17

Thomas Zinner

SUMMARY

18

Thomas Zinner

Lessons Learned – Experiments

Booking of resources o.k. Easy booking of PLE, PLC and VINI nodes via PlanetLab SFA G-Lab manually connected via overlay Routing needs more support PlanetLab restrictions: Overlay of end systems, only one interface configurable VINI provides preconfigured topology Observation tools as a joint effort Tools needed for precise measurements (active / passive) Agreements on tools essential for comparability and share- ability of results, standardization Clock sync support across testbeds

19

Thomas Zinner

Lessons Learned – Testbeds

G-Lab is exclusive resource PRO: exclusive, better controllable, arbitrary software CON: closed, small number of nodes, scope Germany PLE offers many additional functions PRO: many research tools, high precision measurement infrastructure (passive & active) CON: routing support VINI is best for routing PRO: Routing support, easy booking via SFA CON: few nodes, not all nodes on public Internet VINI is too good (low delays, low jitter) high precision tools required delay variation achieved by additional queue

20

Thomas Zinner

Conclusion

Aims of MultiNext Project Gain experience in the usability of federated testbeds High precision and hop-by-hop one-way delay measurements Validation of a performance model for multipath transport in a Federated Future Internet Discussion of testbeds and experimental setup Presentation of results for active and passive measurements Lessons learned Future Work: Combination of active and passive measurements Investigation of more complex scenarios

21

Thomas Zinner

Questions?

Thank you'