Transcript Task 2 Traffic Generation and Measurement Definitions Pass-1

MB - NG
Task2 Traffic Generation and Measurement
Definitions Pass-1
MB-NG Technical Meeting 03 May 02
R. Hughes-Jones Manchester
1
MB - NG
Proposed Agenda
 Load Traffic



Task 2.1.1 Define Simulated and Background Traffic
Tools
Control - logging
 Measurement Systems









What do MB-NG NEED to measure
• Tech and Political
Task 2.2.1 Define low level Probe Measurements
Metrics Pragmatic
Tools
Control - logging
Relation to IETF and GGF Network Metrics
Draft reports
Next Technical “bits&bytes” Meeting
AOB
MB-NG Technical Meeting 03 May 02
R. Hughes-Jones Manchester
2
MB – NG SuperJANET4
Development Network (22 Mar 02)
MB - NG
WorldCom
OC48/POSSR-SC
MAN
MCC
OSM-4GEWAN-GBIC
OSM-4GEWAN-GBIC
Leeds
SJ4 Dev
C-PoP
Warrington
12416
OSM1OC48POS-SS
Gigabit Ethernet
2.5 Gbit POS Access
2.5 Gbit POS core
MPLS Admin. Domains
RAL
RAL
OSM-4GEWAN-GBIC
OSM-4GEWAN-GBIC
OSM1OC48POS-SS
SuperJANET4
Production
Network
SJ4 Dev
C-PoP
Reading
12416
OC48/POSLR-SC
SJ4 Dev
C-PoP
London
12416
WorldCom
WorldCom
Dark Fiber (SSE)
POS
MB-NG Technical Meeting 03 May 02
R. Hughes-Jones Manchester
SJ4 Dev
ULCC
12016
OC48/POSSR-SC
ULCC
UCL
UCL
OSM-4GEWAN-GBIC
OSM-4GEWAN-GBIC
OSM1OC48POS-SS
3
MB - NG




Defining Background Traffic
Regular traffic - constant size packet, regular spaced in time
Poisson traffic - constant size, exponential spacing to form transient queues
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
 UCL real conf playback tool
 Multicast ?
 General traffic captured at edge of a site, e.g. Manchester
 Do tests with a gen to see what gets dropped 0.5 Gbit typical poeak UCL
 Web-bursty traffic – web mirror – wget
Need to be able to reproduce traffic so no TCP as non-linear. Statistical reproducible OK
In general UDP best mto understand the net but some TCP bground is good too
 Self-Similar
What about generating MPLS frames??
 Consider UDP TCP flows
 Need ToS / QoS to be set
 How to control



Start Stop
measure load as function of time – links and Queues
Start and end numbers
How to get data from the core routers – eg need PC in the PoP?
MB-NG Technical Meeting 03 May 02
R. Hughes-Jones Manchester
4
MB - NG
Defining the Measurements (1)
 Suggested Approach





What do we need to measure for MB-NG
How do we analyse it ?
What do we expect ?
What does it tell us ?
NOT defining the measurement matrix – link to Task5 Policies
 General Conditions include



IP operation
MPLS
QoS/Diffserv




Apply techniques to understand the performance of:
PC-PC
Routers and (LAN) interconnects
Development Core – useful to know about the GSR performance details how
do we load it (PC only 1Gig I/f – may be Spirant – Dante OC48 kit.
 End-2-end
MB-NG Technical Meeting 03 May 02
R. Hughes-Jones Manchester
5
MB - NG
Defining the Measurements (10)
 UDP round-trip latency vs packet size



Sum of dt/dl transfer rates
Router performance
Stack performance
 UDP round-trip histograms


Indication of network load
Switching / transport effects – spikes
 UDP 1-way delay
 UDP throughput vs Packet size and transmit delay



Throughput behaviour
The effect of load on switch/router throughput
Offered vs achieved throughput
 UDP Packet Loss vs transmit rate and burst size





Loss rate ( note location of loss)
Packet loss distribution as function of time
Packet loss for simulated “TCP bursts”
Buffer sizes in the path
Detect packet re-ordering
 UDP Inter-frame Jitter as function of packet transmit spacing


Indication of network load
Behaviour of end system NICs
MB-NG Technical Meeting 03 May 02
R. Hughes-Jones Manchester
6
MB - NG
Defining the Measurements (11)
 TCP round-trip latency vs Message size




Sum of dt/dl transfer rates
(Router performance)
Stack / protocol performance
Detect Packet size dependencies
 TCP round-trip histograms



Indication of network load
(Switching / transport effects – spikes)
Stack / protocol performance
 TCP throughput vs Message size and transmit delay




Throughput behaviour cf UDP
Offered vs achieved throughput
Packet loss distribution as function of time + Re-transmit rate
Loss as func of pkt rate eg keep the data rate the same change pkt size – multi-streams
 TCP throughput vs Window size / TCP tuning
 TCP throughput vs number of streams



Stream throughput - benefits
Packet loss distribution as function of time + Re-transmit rate
Effect on Network
 TCP Protocol behaviour - tcptrace


What are the “burst” lengths
Effect of routers / end system NICs
MB-NG Technical Meeting 03 May 02
 All this for WRED Wt fair Qing data rate const and ch pkt size – chack how well the
R. Hughes-Jones Manchester
routers do the Qing
7
MB - NG
Next Steps
 Who does what investigation
 Draft reports
 Next Technical “bits&bytes” Meeting 24 May Manchester
Time
Discussion!
MB-NG Technical Meeting 03 May 02
R. Hughes-Jones Manchester
8