Upload
others
View
8
Download
0
Embed Size (px)
Citation preview
in request rate, number of requests, and diversity of
population. The most recent tracing studies have been
larger and more diverse. In addition to static analysis,
some studies have also used trace driven cache
simulation to characterize the locality and sharing
properties of very large traces and to study the effects
of cookies, aborted connections, and persistent
connections on the performance of proxy caching [2].
Gonzaley, et.al had studied six replacement
algorithms LRU, LFU and LFUDA, other three are
specially developed for web documents GD Size,
GDSF and GD, It finally concluded that no
replacement policy outperforms the other for all
content type [3].
Prischepa analyzes the effectiveness of LRU-K
replacement policy for the purpose of caching on
proxy server [4]. Cao et.al introduced greedy Dual
Size, which incorporates locality with cast and size
concern in a simple and non parameterized fashion
for high performance [5]. G Golan[6] proposes an
optimal offline algorithm for replacement in
multilevel cache, based on an algorithm for the
relaxes list updated problem and the DEMOTE
operation.
Shiva Shankar Reddy P, Swetha L. [7] proposes a
new method of caching HTTP Proxy servers which
takes lower bandwidth by maintaining a cache of
internet objects. V. Sathiyamoorthi and Dr.Murali
Bhaskaran, discusses various data preprocessing
techniques that are carried out at proxy server access
log which generate web access pattern. These
patterns are used for further applications [8].
Martin Arlitt, Ludmila Cherkasova, John Dilley,
Richard Friedrich, Tai J in [9] introduces virtual
caches, an approach for improving the performance
of the cache for multiple metrics simulteneously.
Yong Zhen Guo, Kotagiri Ramamohanarao and
Laurence A. F. Park [10] proposes web page
prefetching technique, they must be able to predict the next set of pages that will be accessed by users,
and ―Page Rank-Like Algorithm‖ is proposed for
conducting web page prediction.
R. Gupta and Tokekar [11] have presented a
preeminent pair of replacement algorithms for L1 and
L2 cache for proxy server. According with them the
access pattern of L1 and L2 cache are different. Thus
the replacement algorithm which is giving efficient
results for L1 may not be suitable for L2 cache. They
concluded that the pair of algorithm effort more
efficient than the savme used algorithm.
Most of studies for replacement policies have
been done for L1. Replacement algorithms can be
recency based or frequency based or may follow both
aspects, such as Least Recently Used (LRU) [12,13].
Least Recently Used-K (LRU-K) [14], Most Recently
Used (MRU) [12,15]. Michael et al proposes a policy
karma which uses application hints to partition the
cache and to manage each range of blocks with the
policy best suited for its access pattern[16].
John Dilley et. al reports on the implementation
and characterization of two newly proposed cache
policies, LFU with Dynamic Aging (LFUDA) and
Greedy Dual Size – Frequency (GDS-F) in the squid
cache, The combination of replacement algorithm
and offered workload determines the efficiency of
cache in optimizing the utilization of system
resources [17].
3. Issues of Web Proxy Caching Due to the explosive and ever growing size of the
web, distributed caching has received considerable
attention. The major aim of cache is to move the
frequently accessed information closer to the users.
Caching system should improve performance for end users, network operators, and content providers.
Caching can be recognized as an effective way to:
speed up web access, reducing latency perceived by
the users, reduce network traffic, reduce server load,
and improve response time to the users.
A. Load Balancing
The situation occur at any time for large number of clients who wishes to simultaneously
access data or get some services from a local
cache with single server. If the site is not
provisioned to deal with all of these clients
simultaneously, service may be degraded or lost.
Several approaches to overcoming this issue have
been proposed. The most frequently used method
is caching. This caching strategy stores copies of
popular pages or services throughout the Internet;
this spreads the work of serving a page or service
across several servers.
B. Transparency
Transparency of cache systems enables users
to get the benefits of caches without knowing that
they exist, or without knowing their physical
location. The advantages of this technique are
easy to use, no configuration required by the end
user and no users can bypass the cache.
Yogesh Niranjan et al, Int.J.Computer Technology & Applications,Vol 4 (2),221-225
IJCTA | Mar-Apr 2013 Available [email protected]
222
ISSN:2229-6093
C. Scalability
It is vital that the cache system be scalable as
the number of users and servers increases. It can
be clustered or cooperative and stand-alone
caches. Stand-alone caches are better suited for
individual systems and are easier to maintain.
However, cooperation between caches could
provide more information about cached data, which could be communicated between caches
without referring to the originating servers.
D. Cache miss
Cache systems should be capable of efficiently
handling cache misses. When a request cache
misses, a decision should be taken on where to
forward the request. And also a cache system
should decide on which data to be cached or
should all cached data be treated equally.
4. Design and Implementation of Proposed
Proxy Caching Algorithm
Web proxy caching is a well-known strategy for
improving performance of Web-based systems by
keeping Web objects that are likely to be used in the
near future closer to the client. Most of the current
Web browsers still employ traditional caching
policies that are not efficient in Web caching.
Web caching is an emerging technology in Web
and in Web caching if the client is requesting a page
from a server, it will fetch from the server and will
give response to the server. According to the
locations where objects are cached, Web caching
technology can be classified into three categories i.e.,
client‘s browser caching, client-side proxy caching,
and server-side proxy caching.
A Client Side Proxy Caching (CSPC) is a caching
server that acts as an intermediary for requests from
clients seeking resources from other servers. A client connects to the proxy server, requesting some
service, such as a file, connection, web page, or other
resource available from a different server. The proxy
server evaluates the request according to its filtering
rules. For example, it may filter traffic by IP address
or protocol.
If the request is validated by the filter, the proxy
provides the resource by connecting to the relevant
server and requesting the service on behalf of the
client. A proxy server may optionally alter the client's
request or the server's response, and sometimes it
may serve the request without contacting the
specified server. In this case, it 'caches' responses
from the local cache.
Figure 1 show a proxy server with cache memory
which runs with many features such as Reduces
network traffic, Reduces Latency time, reduce load
on web server. This architecture also inherently helps speeder browsing of web pages. In this system when
proxy cache saturated and new page request arrives at
proxy a page replacement algorithm decides which
page has to be evict from the cache. Efficiency of
system depends on the page replacement algorithm.
Client Side Proxy Caching Algorithm (CSPC)
There has been extensive theoretical and
empirical work done on exploring web caching
policies that perform best under different
performance metrics. Many algorithms have been
proposed and found effective for web proxy caching.
These algorithms range from simple traditional
schemes such as Least-Recently Used (LRU), Least-
Frequently Used (LFU), First-In First-Out (FIFO), and various size-based algorithms, to complex hybrid
algorithms such as LRU-Threshold, which resembles
LRU with a size limit on single cache elements,
Lowest-Relative Value (LRV), which uses cost, size
and last reference time to calculate its utility, and
GreedyDual, which combines locality, size and cost
considerations into a single online algorithm.
Figure 1 Proxy Server Caching
C1
C2
Cn
Cache
Processor
Clients Proxy Server Server
Internet
Yogesh Niranjan et al, Int.J.Computer Technology & Applications,Vol 4 (2),221-225
IJCTA | Mar-Apr 2013 Available [email protected]
223
ISSN:2229-6093
The table I show the proposed Client Side Web
Proxy caching algorithm. This algorithm focused on
the aspect i.e. algorithm of replacing documents.
With the study of web cache characteristics going
further; algorithms of replacing documents based on
the statistics of collected web data are proposed.
Following consider factors into its scheme:
Document reference frequency
Document size
Consistence of documents
Freshness of document
Efficient schemes combine more than one of factors
in their implementation of web cache. Some
algorithms consider different cache architecture to
improve caching performance.
Table I: Client Side Proxy Cache Algorithm (CSPC)
5. Experimental Result
A proposed algorithm (CSPC) is devloped in
windows XP using C# .Net. and the unique
identification number is allot to unique URL‘s of proxy server log. These numbers are taken as
reference string that become input to the algorithms.
The result of algorithm in the form of Hit Rate are
shown in Table II.
6. Conclusion This paper is basically concentrated to explore the
client side proxy caching algorithm which is best
suited for proxy server. Real trace of web references
is achieved with the help of log details of proxy
server. For the simulation numeric reference string
was obtained by giving numeric identity to each of
the URLs. After simulation it concluded that the
proposed client Side Proxy Caching Algorithm
perform well than other algorithms like LRU, LFU
and FIFO. The CSPC algorithm improves Hit Ratio
approx 10.67%.
After exhaustive simulation experiments it is
concluded that for proxy caching the SCPC hit ratio
performance better than others algorithms.
1. WHILE there is a page p in Cache in the current window.
2. Serve first such p and mark the page
3. IF all pages in the cache are marked
4. Unmark all the pages
5. Evict randomly an unmarked page from the
cache
No of
Request
Replacement Algorithms
LRU LFU FIFO CSPC
100 0.400 0.410 0.390 0.460
300 0.486 0.496 0.483 0.563
500 0.490 0.496 0.480 0.560
800 0.548 0.551 0.545 0.625
1000 0.591 0.596 0.585 0.664
1200 0.635 0.640 0.631 0.691
1500 0.634 0.638 0.630 0.701
1800 0.658 0.658 0.653 0.711
2000 0.677 0.679 0.671 0.737
Figure 2: Hit Ratio Analysis
Yogesh Niranjan et al, Int.J.Computer Technology & Applications,Vol 4 (2),221-225
IJCTA | Mar-Apr 2013 Available [email protected]
224
ISSN:2229-6093
7. References
[1] David A. Malts and Pravin Bhagwat (march 1998). ―Improving HTTP caching proxy performance with
TCP tap‖. Technical report, IBM.
[2] A. Feldman et. al.; Performance of Web Proxy Caching in Heterogeneous Bandwidth Environments,
Proceeding of INFOCOM 99, 1999.
[3] F.J. Gonzalez- Canete, E Casilari, Alicia Trivino – Cabrera, ―Characterizing Document Types to Evaluate
Web Cache Replacement Policies‖, International
conference on Information Technology ITNG 2007.
[4] Valdimir V. Prischepa, ―An efficient Web Caching
Algorithm based on LFU-K replacement policy‖
Spring Young Researcher‘s Colloquium on Database
and Information System, 2004.
[5] P.cao and Irani, ―Cast aware WWW proxy caching
Algorithms‖, in proc. USENIX Symp. Internate
Technologies and System, Monterey, CA, 1997.
[6] Gala Golan ―Multilevel cache management based on
application Hints‖ computer science department,
Technion Haifa 32000, ISRAEL. November 24, 2003.
[7] Shiva Shankar Reddy P,Swetha L ―Analysis and
Design of Enhanced HTTP Proxy Cashing Server ―
paper published in International Journal of computer Technology, Volume 2 (3), 537-541.
[8] V. Sathiyamoorthi and Dr.Murali Bhaskaran ―Data
Preprocessing Techniques for Pre-Fetching and Caching of Web Data through Proxy Server‖
International Journal of Computer Science and
Network Security, VOL.11 No.2011.
[9] Martin Arlitt, Ludmila Cherkasova, John Dilley,
Richard Friedrich, Tai J in "Evaluating Content
Management Techniques for Web Proxy Caches",
published in ACM SIGMETRICS Performance Evaluation Review Volume 27 Issue 4, March 2000.
[10] Yong Zhen Guo, Kotagiri Ramamohanarao and
Laurence A. F. Park ―Personalized PageRank for Web Page Prediction Based on Access Time-Length and
Frequency‖ This paper published in 2007
IEEE/WIC/ACM International Conference on Web
Intelligence.
[11] R. Gupta, Tokekar .‖Preeminent pair of replacement
algorithms for L1 and L2 cache for proxy server‖.
First Asian Himalayas International Conference AH-ICI 2009.
[12] Abraham Silberschatz and Peter Baer Galvin,
Operating System concepts. Addison Wesley. 1997.
[13] A. Dan and D. Towsley, ―An Approximate Analysis
of the LRU and FIFO Buffer Replacement
Schemes‖, in Proceedings of ACM SIGMETRICS.
Boulder, Colorado, United States, 1990, pp. 143—
152.
[14] E.J O‘Neil, P.E. O‘Neil and G. Weikum, ―The
LRU-K page replacement algorithm for database
Disk Buffering‖ proc.ACMSIGMOD Int ‗l Conf.
Management of Data, pp. 297-306, May 1993.
[15] M.J. Bach, ―The Design of the UNIX Operation
system‖. Engle wood Cliffs, NJ: Prentice-Hall,
1986.
[16] Michael Factor, Assaf Schuster, Gala Yadgar,
―Multilevel Cache Management Based on
Application Hints‖, Technion- Computer Science Department Technical Report CS-2006.
[17] John Dilley, Martine Arlitt and Stephane Perret
―Enhancement and Validation of Squid‘s Cache Replacement Policy‖ Internet Systems and
Applications Laboratory HP Laboratories Palo Alto
HPL- 1999-69, May 2009.
Yogesh Niranjan et al, Int.J.Computer Technology & Applications,Vol 4 (2),221-225
IJCTA | Mar-Apr 2013 Available [email protected]
225
ISSN:2229-6093