Transcript MB - NG Status of Task 2 Traffic Generation and Measurement

MB - NG
Status of Task 2
Traffic Generation and Measurement
MB-NG Meeting UCL 1 Nov 02
R. Hughes-Jones Manchester
1
MB – NG SuperJANET4
Development Network
MB - NG
MAN
MCC
OSM-4GEWAN-GBIC
OSM-4GEWAN-GBIC
OSM1OC48POS-SS
WorldCom
SJ4 Dev
C-PoP
Warrington
12416
SuperJANET4
Production
Network
Gigabit Ethernet
2.5 Gbit POS Access
2.5 Gbit POS core
MPLS Admin. Domains
SJ4 Dev
C-PoP
Reading
12416
RAL/
UKERNA
OSM-4GEWAN-GBIC
Leeds
WorldCom
SJ4 Dev
C-PoP
London
12416
WorldCom
Dark Fiber (SSE)
MB-NG Meeting UCL 1 Nov 02
R. Hughes-Jones Manchester
UCL
UCL
OSM-4GEWAN-GBIC
OSM-4GEWAN-GBIC
OSM1OC48POS-SS
2
MB - NG
The Scope and Progress
 Task: Define Traffic and measurement systems needed to
characterise the network



Network Characteristics
Measurements and systems
Background Load traffic
 Two Face2Face and several VC to progress the work
 4 Documents in preparation:




Network Metrics and Methodologies Framework
Measurements and Methodology for MB-NG
Proposed MB-NG QoS tests
Specification of background generation
 Consideration of tools
MB-NG Meeting UCL 1 Nov 02
R. Hughes-Jones Manchester
3
MB - NG
Network Metrics and Methodologies Framework
 Has set down understanding of Metrics (Network Characteristics)









Latency – 1-way, roundtrip
Loss – rate, distribution
Throughput
Connectivity and availability
Very much in line with GGF – although terminology a bit different
Like GGF has taken note of IETF IPPM work
Topics raised include
Accuracy and systematics/bias of the measurements – time
Status: Only needs editorial tidyup
MB-NG Meeting UCL 1 Nov 02
R. Hughes-Jones Manchester
4
MB - NG
Measurements and Methodology for MB-NG
 Many good discussions – now clear what MB-NG requires to observe
Network Characteristics - Both UDP and TCP traffic considered:









Latency – 1-way, roundtrip
Loss – rate, distribution
Throughput
Connectivity and availability
Web100 essential
A set of these measurements will be used for the Policies defined.
Useful experience from the play-pen / DataGrid & DataTAG / ATLAS
Need to include Protocol behaviour and MIB information.
Status: Document stalled due to EU demo
MB-NG Meeting UCL 1 Nov 02
R. Hughes-Jones Manchester
5
MB - NG
Proposed MB-NG QoS tests
 Sets down how the measurement and background traffic could be
combined and applied to various QoS scenarios:




BE (as reference)
EF & BE
BE & LBE
EF, BE & LBE
 Some proposals for the parameters (BW, scheduling) for QoS Policies
 Useful experience gained from the play-pen
 Extending Task2 towards Task 8
 Status: In Draft
MB-NG Meeting UCL 1 Nov 02
R. Hughes-Jones Manchester
6
MB - NG







Defining Background Traffic
Gained experience with using Smartbits – seems flexible – some scripts in use
Regular traffic - constant size packet, regular spaced in time
Poisson traffic - constant size, exponential spacing to form transient queues
Variable packet sizes too
Tools include UDPmon, iperf, …
Investigating code to generate traffic from long-tailed distributions
Still discussing:



IETF traffic mix - different sizes and different probability of each size sent
Play back of real traffic patterns generated from packet headers pre-recorded from suitable
points of the production network. This might include:
•
•
Video Conference traffic -> play back - rude/crude tools
General traffic captured at edge of a site, e.g. Manchester
Web-bursty traffic – web mirror – wget
MB-NG Meeting UCL 1 Nov 02
R. Hughes-Jones Manchester
7
MB - NG
MB-NG Meeting UCL 1 Nov 02
R. Hughes-Jones Manchester
8
MB - NG Measurements and Methodology for MB-NG
MB-NG Meeting UCL 1 Nov 02
R. Hughes-Jones Manchester
9