Trace-Driven Queueing Results

Trace-Driven Queueing Results

Click <here > for a page of simple queueing examples.

The data in the table are the results from feeding each trace into a FIFO queue with infinite size and server rate C.
C was varied to be the greatest integer less than or equal to the value which gives the target utilization (which may introduce a small amount of granularity effect).
Queue length distributions were calculated after skipping the first 50,000 queue length values.
Utilization data: column 1 shows the queue size, column 2 shows the number served (departures) in that time bin.
Links to loglnylogx plots show log10(-ln(P(Q>x)) vs. log10(x). With a Weibull-like tail for P(Q>x) these plots should show a straight line with slope (2-2H). Where such a straight line can be seen, a regression line with slope and H estimate are included.
Links to logylogx plots show log10(P(Q>x)) vs log10(x). With a Pareto-like tail for P(Q>x) these plots should show a straight line. Where such a straight line can be seen, a regression line with estimated slope is included.


			Queue Summaries					Utilizations
Data	Traffic Feature	Length	Length Plots	Length Dist. Plots	25%	50%	75%	90%	99%
Synthetic: 18 clients, 10 servers, 80 Mbps		4,700,001	pdf	<logy> <loglnylogx>	data	data	data	data	data
Fractional Gaussian Noise (from Vladas) (after truncation)		7,200,000	pdf	<logy <loglnylogx>	data	data	data	data	data

Packet Counts
Tue. April 9 300 2002 1ms		7,322,437		<logy><loglnylogx><logylogx>	data	data	data	data	data
Tue. April 9 300 2002 10ms		732,243	pdf	<logy><loglnylogx><logylogx>	data	data	data	data	data
Tue. April 9 800 2002 1ms		7,342,488		<logy><loglnylogx><logylogx.	data	data	data	data	data
Tue. April 9 800 2002 10ms	uptrend with spikes	734,248	pdf	<logy><loglnylogx><logylogx>	data	data	data	data	data
Sat. April 13 1930 2002 1ms		7,340,807		<logy><loglnylogx><logylogx>	data	data	data	data	data
Sat. April 13 1930 2002 10ms	spikes	734,080	pdf	<logy><loglnylogx><logylogx>	data	data	data	data	data

Sun. April 6 1000 2003 10ms		732,178	pdf	<logy><loglnylogx><logylogx>	data	data	data	data	data
Mon. April 7 1000 2003 10ms	up trend	731,992	pdf	<logy><loglnylogx><logylogx>	data	data	data	data	data
Mon. April 7 1100 2003 10ms	center hump	720,128	pdf	<logy><loglnylogx><logylogx>	data	data	data	data	data

Byte Counts
Tue. April 9 300 2002 1ms		7,322,437		logy	data	data	data	data	data
Tue. April 9 300 2002 10ms		732,243	pdf	logy	data	data	data	data	data
Tue. April 9 800 2002 1ms		7,342,488		logy	data	data	data	data	data
Tue. April 9 800 2002 10ms	uptrend with spikes	734,248	pdf	logy	data	data	data	data	data
Sat. April 13 1930 2002 1ms		7,340,807		logy	data	data	data	data	data
Sat. April 13 1930 2002 10ms	spikes	734,080	pdf	logy	data	data	data	data	data

Sun. April 6 1000 2003 10ms		732,178	pdf	logy	data	data	data	data	data
Mon. April 7 1000 2003 10ms	up trend	731,992	pdf	logy	data	data	data	data	data
Mon. April 7 1100 2003 10ms	center hump	720,128	pdf	logy	data	data	data	data	data