89
as
N
p(i), where N is the number of the most frequently accessed data items (e.g.,
i=1
files) in the cache. Thus,
N
N
C
R
hit
(N ) =
i
- 
)
-1
.
(5.3)
i
,
where C = (
i=1
i=1
The temporal locality, (R
hit
(N )), is asymptotically equal to lnN (when   = 1)
or C
1- 
(when 0 <   < 1) [18]. The above equation implies that the cache hit ratio
is determined by two parameters; 1) the number of web pages in the cache, and 2) the
slope of log log plot of the trace file, (i.e.,  ).
Figure 5.6 presents the cache hit ratio obtained using the traces and the equation
5.3 as a function of cache size. From this figure, we observe that the temporal locality of
the trace files and the temporal locality obtained from Equation (5.3) are pretty close.
In other words, the temporal locality of traces follow the Zipf distribution (using the
  listed in Table 5.4). The PSU trace shows the highest locality among three traces.
Figure 5.6(c) depicts that less than 512MBytes is enough to provide more than 95%
cache hit ratio. This is because the frequently accessed data items are concentrated on a
small part of the whole data set, which shows very little sign of the heavy tail property,
even though the size of the PSU data set is more than 2 GBytes. Figure 5.6(a) shows
that the CSE trace has about 87% and 98% cache hit ratios when the cache size is set at
1 GBytes and 2 GBytes, respectively. However, the UCB trace shows very low temporal
locality compared with the other trace files. With 1 GBytes data cache, only 70% cache
hit ratio is achieved.












  

Home

About Services Network Support FAQ Order Contact
 

 

Clan Web Hosting

Our partners:Jsp Web Hosting Unlimited Web Hosting Cheapest Web Hosting  Java Web Hosting Web Templates Best Web Templates PHP Mysql Web Hosting Interland Web Hosting Cheap Web Hosting PHP Web Hosting Tomcat Web Hosting Quality Web Hosting Best Web Hosting  Mac Web Hosting 

Lunarwebhost.net  Business web hosting division of Vision Web Hosting Inc. All rights reserved