|
|
|
Thomas Kunz |
|
Systems and Computer Engineering |
|
Carleton University |
|
http://kunz-pc.sce.carleton.ca/ |
|
tkunz@sce.carleton.ca |
|
|
|
|
Overview of WAP (Wireless Application Protocol)
and Trace Collection |
|
The Big Picture |
|
Sessions |
|
Activity Factors |
|
Conclusions and Future Work |
|
|
|
|
|
“de-facto world standard for wireless
information and telephony services on digital mobile phones and other
wireless terminals” |
|
published a global wireless protocol
specification based on existing Internet standards such as XML and IP for
all wireless network |
|
Marketing Cloud: |
|
Handset manufacturers representing over 75% of
the world market across all technologies have committed to shipping
WAP-enabled devices. |
|
Carriers representing nearly 100 million
subscribers worldwide have joined WAP Forum |
|
|
|
|
|
|
|
|
|
|
|
|
|
Users typically make a sequence of requests,
defined as a “browser session” |
|
A wireless channel is allocated at beginning and
released at the end: |
|
90 seconds timeout |
|
User terminates browser application |
|
Phone is powered off |
|
Problems: |
|
IP addresses assigned dynamically, so we cannot
track users |
|
When session times out, new IP address may be
assigned, even though same “user session” |
|
Number of sessions closely follows average daily
traffic |
|
Sessions are, on average, rather short (90% are
less than 3.77 minutes) |
|
|
|
|
|
|
|
|
Wireless link is scarce resource, but is it well
used by user/browser? |
|
Activity Factor: percentage of time that channel
is used to transmit data |
|
Determination: |
|
How long is channel allocated to user |
|
Trivial for pure circuit-switched connection |
|
More complicated for CDMA, where channel is not
always at full rate |
|
How much time is spent transferring user data |
|
assuming link bandwidth of 19.2 kbps |
|
data volume per session known from traces |
|
Activity factors differ for uplink and downlink,
but are constant for period studied |
|
Uplink: 11% |
|
Downlink: 30% |
|
|
|
|
|
|
Some similarities: |
|
Periodicity |
|
Daily patterns (and to some extent half-daily
patterns) |
|
Weekly patterns (and to some extent half-weekly
patterns) |
|
Not enough trace data to confirm/test for
seasonal patterns |
|
Self-similarity (Hurst Parameter between 0.79
and 0.82) |
|
But also some differences: |
|
Smaller packets (95% of all packets less than
220 bytes) |
|
Shorter sessions |
|
Traffic more balanced (dowlink traffic “only”
about 3 times as much data as uplink traffic) |
|
No growth trend in data presented in paper, but
long-term growth trend clearly visible since |
|
|
|
|
|
|
Installed trace collection infrastructure,
continuous trace collection effort |
|
Analyzed traffic, derived a number of
properties, compared to WWW traffic |
|
Future Work: |
|
Reconfirm findings/invariants for traces
collected since January 1, 2000 |
|
Refine analysis with additional data sources
(can we match users to IP addresses over time?) |
|
Use traces to build performance prediction
models based on LQN (layered queuing models): |
|
Impact of more users |
|
New applications |
|
Notes