Capacity Planning for Internet Service
Networks
Geoff HustonNTW Track4
Issues
TCP/IP Protocol Behavior
IssuesUsage ProfileCapacity GuidelinesGrowth Levels
Planning Issues
Understand the domain of operation technical issues market issues competitive issues regulatory issues
TCP/IP Protocol Issues
TCP/IP is NOT a flow damped protocol end to end flow management sliding window protocol adaptive flow rate designed to probe and
use max available end to end bandwidth only limited by end system buffering size
bandwidth x delay
system buffers are getting larger as OS vendors come to understand the problem
TCP/IP Protocol Issues
TCP/IP Data Flow Rate AdaptationTCP/IP Data Flow Rate Adaptation
Time
Dat
a F
low
Rat
e
Steady State Available Bandwidth
Rate overflow loading into network queues
TCP/IP Protocol Issues
No network-based flow control mechanismNetwork-based packet loss signals end
systems to collapse transmission window sizeVarying window size allows adaptive flow
metrics to adapt to changing maximum available capacity
Sustained insufficient capacity leads to congestion induced collapse of data throughput
TCP/IP Protocol Issues
Many simultaneous TCP sessions interact with non-predictive non-uniform load (ftp://thumper.bellcore.com/pub/dvw/sigcom93.ps.Z)
Peaks start to synchronize with each other
Buffering evens out individual flows, but buffers themselves behave with fluctuating load
Buffering adds latency
TCP/IP Protocol Issues
TCP/IP efficiency under congestion loadTCP/IP efficiency under congestion load
Traffic Level
Dat
a T
hro
ugh
put
33% 66%
33%
66% Congestive Collapse -
The slide to miseryand packet loss
TCP/IP Protocol Issues
TCP vs UDPUDP-based applications
Internet Phone, Video, WorkgroupUDP Issues
no flow control mechanism sustained use forces precedence over TCP
flows increasing use of flow bandwidth negotiated
protocols for these applications (RSVP)
TCP/IP Protocol Issues
Damping network capacity is not a demand management tool
Network capacity must be available to meet peak demand levels without congestion loss
Usage Profile
Two major Internet use profiles: Business use profile
peak at 1500 - 1600plateau 1000 - 1730
Residential dial profilepeak at 2030 - 2330plateau 1900 - 2400
Usage Profile
Distance profiles12% Local18% Domestic Trunk70% International
Traffic mix due to: Distance invisible applications without user
control Distance independent user tariff
Capacity Guidelines
Link Utilisation Average weekly traffic level set to 50%
of available bandwidth.Core network capacity should be
dimensioned according to aggregate access bandwidth
Link Usage Profile - optimal
peak loading less than 10% timegreater than 50% loading for
50% timetraffic bursting visible
Link Usage Profile - overloaded
90% peak loading for 45% time60% peak loading for 60% timeno burst profile at peak loadsimbalanced traffic (import based)
visible plateau traffic load signaturesmall load increases cause
widening plateau
Link Usage Profile - saturated
Overall Growth Levels
Two growth pressures: serviced population the changing Internet service model
more network-capable applicationsusing more bandwidth
Overall Growth Levels
For a constant service model the growth curve will exhibit demand saturation
0
100
200
300
400
500
1 11 21 31 41 51 61
Overall Growth Levels
For a changing service model the saturation point will move
More intense network use by increasingly sophisticated applications
0
200
400
600
800
1000
1200
1 11 21 31 41 51 61
Technology shift
How to plan
Generate a market demand model forecast the number of services in operation
existing servicesgrowth ratemarket capture level (competitive position)
forecast the average demand per servicedial access, leased lineweb, ftp, usenetcaching trendsnew Internet services
How to plan
demand models are typically very uncertain indicators high level of uncertainty of externalities highly dynamic competitive position poorly understood (and changing) service
demand model
How to plan
Forward extrapolation assume existing traffic follows a general
growth model forward extrapolat the growth model
Good for short term planning (12 months)
Cannot factor latent demand market price sensivity
Trend forecasting
0
20000
40000
60000
80000
100000
120000
7/1/
96
7/11
/96
7/21
/96
7/31
/96
8/10
/96
8/20
/96
8/30
/96
9/9/
96
9/19
/96
9/29
/96
10/9
/96
10/1
9/96
10/2
9/96
11/8
/96
11/1
8/96
11/2
8/96
12/8
/96
12/1
8/96
12/2
8/96
1/7/
97
1/17
/97
1/27
/97
2/6/
97
2/16
/97
2/26
/97
3/8/
97
3/18
/97
3/28
/97
4/7/
97
4/17
/97
4/27
/97
Total Capacity
Daily Traffic IN
Daily Traffic OUT
historical usage vs capacity data
Growth Trends
0
100000
200000
300000
400000
500000
600000
1/7/
96
28/7
/96
24/8
/96
20/9
/96
17/1
0/96
13/1
1/96
10/1
2/96
6/1/
97
2/2/
97
1/3/
97
28/3
/97
24/4
/97
21/5
/97
17/6
/97
14/7
/97
10/8
/97
6/9/
97
3/10
/97
30/1
0/97
26/1
1/97
23/1
2/97
19/1
/98
15/2
/98
14/3
/98
10/4
/98
7/5/
98
3/6/
98
30/6
/98
27/7
/98
23/8
/98
in
out
Trend - high
Trend - mid
Trend - Low
97/98 bandwidth
Planning
undertake demand and trend forecast models
constantly review the model against generated usage data
recognise that the larger the capacity you need the longer the lead time to purchase it
recognise that the bigger the purchase the greater the requirement for capital