Design of Prefetching Framework for Mozilla Firefox

Fuzzy C-means clustering based prefetching to reduce web traffic and Exploring the Web Caching Method to Improve the Web Efficiency
Dr.AstonCole Profile Pic
Dr.AstonCole,United Kingdom,Researcher
Published Date:14-11-2017
Your Website URL(Optional)
Comment
I.J. Information Engineering and Electronic Business, 2015, 5, 7-12 Published Online September 2015 in MECS (http://www.mecs-press.org/) DOI: 10.5815/ijieeb.2015.05.02 Analysis and Design of Prefetching Framework for Mozilla Firefox Neha Sharma Northern India Engineering College, Delhi, India E-mail: neha.sh.2689gmail.com Sanjay Kumar Dubey Amity University, Noida, U.P. 201303, India E-mail: skdubey1amity.edu Abstract—The presence of number of web sites has web pages client is using. Now, the question is how to increased the user’s attraction towards web objects. This select the pages. For this, different prefetching techniques tremendous use came up with the future requests are used. These prefetching techniques are browser prediction depending upon the current and past access dependent. The browser with better prefetching scheme, behaviour. Use of Internet has boomed up a lot since the is more efficient. last decade. This use also came with the heavy load on Web requests or better say, web surfing patterns of the the internet. In today’s world, speed plays a significant user are almost same for the specific intervals. Even there role and hence the speed augmentation is one of the is some similarity between set of users 2. They might biggest issues. For this, web latency reduction by access the same pattern simultaneously. prefetching is one of the good ideas. For the same, web Prefetching is the technique, where user’s next prefetching is performed, where user’s next expected expected requests are loaded previously in the web cache, requests are prefetched in the web cache of the web by applying some predictive methods. It basically browser. A browser is basically a GUI based application improves the web performance and due to prefetched program, which provides a platform for running Internet. information, user can access the web objects faster and Mozilla Firefox is a web browser, which is in very much hence, user satisfaction towards the browser will improve. use these days. This paper provides an analysis of Prefetching can be performed on server-side as well as Mozilla Firefox prefetching technique (link and DNS browser-side. The browser-side prefetching is the work to prefetching) and then designs a new prefetching scheme be performed by different organisations to reduce the for the same. The experimental results are performed in load and to improve their work performance. A good Matlab 5.0. The results show that the designed prefetching technique at browser-side increases user’s prefetching framework is more efficient in terms of the satisfaction. cache hit ratio. Numbers of browsers are present today but only few among them are famous in users. Users choose a browser Index Terms—Mozilla Firefox, Prefetching, Cache, depending upon its few services, i.e. facility of add-ons, Browser, Link, DNS. extensions available, support and speed etc. If the browsing speed is good and the browser is user-friendly, then it meets the users expected requirements. Browsing I. INTRODUCTION speed depends upon the web latency, if web latency is less, it means waiting time for web objects fetching is less. Communication has grown a lot in the last decade via Less waiting time means, faster access. In today’s Internet. The immense growth of web for sharing scenario, everyone is short of time. So, run-time needs to information and entertainment made the Internet lively. be as short as possible. Hence, prefetching at the browser- This liveliness came with lots of load on it. As a result, side is a good option for enhancing the browser several issues arrived such as increase in web latency. capabilities and pleasing the users. Web cache is used to reduce web latency for short-term nd The paper works on the analysis of 2 most popular prefetching of web objects and to increase bandwidth. web browser, Mozilla Firefox, prefetching scheme. Recently and frequently used web pages are being stored Mozilla Firefox uses link prefetching technique and its in web cache. Reverse proxy is good technique to few versions also uses DNS prefetching scheme. The implement web cache. Instead of going to server, now paper provides a thorough study of the same and then, client will request to the proxy. The web objects will designs a new prefetching framework for the Mozilla directly be loaded from the proxy, resulting in the Firefox browser. A comparative report on the cache hit reduction in waiting time (web latency) and traffic in the ratio of the old and new prefetching techniques has been network. Upon receiving a new object, the proxy server, provided. The efficiency and effectiveness, depending provide a copy to the end-user and keep another copy to upon the cache hit ratio, is represented in the forms of its local storage 1. Web cache of browser, stores the Copyright © 2015 MECS I.J. Information Engineering and Electronic Business, 2015, 5, 7-12 8 Analysis and Design of Prefetching Framework for Mozilla Firefox graphs, to clear the picture. The results found make it future requests for web documents based on previous very much clear that new prefetching framework requests. produces comparatively better results than the old one. Wan et al. discovered latent factors of user browsing Hence, it results in web latency reduction too. behaviours 9. It was discovered on the basis of random The paper is divided into different sections for the ease. indexes and detecting clusters of web users according to Section 2 provides an overview of primarily research their activity patterns acquired from access logs. done in this field. A lot of work has been done previously Ramya and Sathiyamoorthi had re-evaluated principles in this field and this part of paper explains the same. and existing works of conventional and intelligent web Section 3 gives an overview of the research methodology caching 10. Categorisation of prefetching techniques used for the paper. Section 4 focuses on the Mozilla was also done and the main focus is history-based Firefox prefetching techniques, i.e. link and DNS prefetching approach. This three step approach’s first step prefetching. Section 5 concentrates on the designing of is data extraction from proxy web log. Then after, new framework for Mozilla Firefox given in the paper. In extracted data is preprocessed. Using clustering, this section 6, comparative report provided for old and new preprocessed data is mined and to know the patterns to be prefetching designs. Finally, the conclusion of the paper pre-fetched, sequence analysis is being performed. has been written, explaining the work done in brief with Singh et al. proposed a framework for prediction of the merits and scope of the work in the present scenario. web requests of users and accordingly, prefetching the content from the server 11. The proposed framework improved performance of web proxy server using web usage mining and prefetching scheme. They have II. LITERATURE REVIEW clustered the users according to their access pattern and Kosala et al. provided a review in the area of web usage behaviour with the help of K-Means algorithm and mining 3. The paper examines the connection between then Apriori algorithm is applied to generate rules for different web mining types and correspondingly related prefetching pages. This cluster based approach was agent paradigm. This paper focuses basically on applied on proxy server web log data to test the results representation issues, on the process, and on the learning using LRU and LFU prefetching schemes. algorithm, and the application of the recent works as the Sharma and Dubey provided the literature survey in the criteria in the field of web mining. area of web mining 12. The paper basically focuses on Pallis et al. explained the short-term prefetching the methodologies, techniques and tools of the web problem on a web cache environment using an algorithm mining. The basic emphasis is given on the three for clustering inter-site web pages 4. The proposed categories of the web mining and different techniques scheme efficiently integrates web caching and incorporated in web mining. prefetching. According to this scheme, each time a user requests an object, the proxy fetches all the objects which are in the same cluster with the requested object. III. RESEARCH METHODOLGY Sharma and Dubey proposed a framework for web traffic reduction 5. The paper presented a framework for Mozilla Firefox is the second best browser these the prefetching and prediction in web. According to the days13. Due to its speed problem, it is on second framework, previous web requests of the user will be number, instead of being on the top. Its prefetching extracted from the proxy web log. From this web log, scheme somewhere lacked and as a result, web latency strong rules will be generated using FP Growth algorithm. was quiet more as it was expected to be. While searching These rules will be used to prefetch the upcoming for Mozilla Firefox, so many relevant papers were found requests of the current user. on this topic. The relevant material is referred from Vijayan and Jayasudha addressed about the different different journals, conferences, websites and Mozilla prefetching and caching techniques, how they predict the official web page too. The papers and database collected web object to be pre-fetched. It also focused on what are was analysed, depending upon the objective and then the issues, challenges involved when these techniques are after, shortlisted accordingly. To provide relevant applied to a mobile environment 6. information for the Mozilla prefetching techniques, data Greeshma et al. analyzed web prefetching techniques collected form following sources: and other directions of web prefetching 7. The significance of these techniques is to reduce the network Mozilla Firefox official web page traffic and improve the user satisfaction. The authors Scopus Database gave an idea that web prefetching and caching can also be Google Scholar integrated to get better performance. Elsevier Kasthuri et al. explained that depending upon the past ACM requests, future references can be deducted and Springer implemented using predictive prefetching 8. The IEEE prediction engine can be present either in the client or server side. The papers prediction engine resides at client side. It used the set of past references to find correlation IV. MOZILLA FIREFOX PREFETCHING SCHEME and initiates prefetching that were finding out user’s Copyright © 2015 MECS I.J. Information Engineering and Electronic Business, 2015, 5, 7-12 Analysis and Design of Prefetching Framework for Mozilla Firefox 9 Mozilla Firefox different versions use different V. FRAMEWORK DESIGN-OPTIMIZED MOZILLA prefetching techniques. Link prefetching and DNS PREFETCHING TECHNIQUE prefetching are the two major techniques implemented in Mozilla Firefox is a second most popular browser, Mozilla Firefox browser and even in Netscape too. according to a survey 13. Its prefetching concern can be A. Link Prefetching one of the reason of it’s not be on the top. The paper provides a prefetching framework design for the Mozilla A link prefetching mechanism is a technique used for Firefox browser. Caching with prefetching is a good modifications to both the browser and the server 14. option to improve prefetching capabilities. Prefetching Fisher proposed a server-driven link prefetching helps in reducing the web latency and hence web objects technique in 2002 15. Fig. 1 presents the link can be prefetched at the client-side, i.e. browser-side prefetching method proposed by Fisher. It is designed to prefetching. Fig. 2 represents process of prefetching. The provide a solution to prefetching that is friendly to both prefetching framework for Mozilla Firefox has been client and server, specially addressing some of the provided in this section. The design of prefetching shortcomings of existing ad-hoc methods. The browser framework consists of number of steps, explained as follows special directives from the web server (or proxy below: server) that instruct it to prefetch specific documents. The server can inject these directives into an HTML A. User Requests document or into the HTTP response headers 16. The Initially user puts the requests on the Mozilla Firefox browser responds to the prefetch directives after it browser. At a particular time number of requests can be finishes rendering the document (or HTTP response) put on the Mozilla windows. So, all requests need to be containing the directives 17. This mechanism allows fulfilled by the browser for number of users. servers to control precisely what is prefetched by the browser, and it allows the browser to determine when B. Request Fulfillment best to prefetch documents based on local network As the requests come to Mozilla, Mozilla will check inactivity, browser idle time, etc. for the requests on L2 cache. If cache hit, then request will be fulfilled from their only. Otherwise, for the cache miss, connection with the server will be created and requests will be transferred to the server. After TCP connection creation, server will respond to the client’s requests. The server listen the request and reverts back with response. The corresponding response is being transferred to the user. C. Web Prefetch Web prefetch is the module where the user’s future expected requests will be stored depending upon their past behavior. User’s history is stored in web log. User web log is also being maintained on the web. Initial requests of the users will be saved in the web log. Now, depending upon this web log (history of web objects used by different users) users upcoming requests will be Fig. 1. Link Prefetching 15 prefetched. As the user logs in or requests web objects, depending upon its past behavior (stored in web log), its expected future requests are being prefetched and kept in B. DNS Preftching the web cache. In DNS prefetching, depending upon the user’s request, DNS of the expected next requests has been prefetched. WEB LOG PREFETCHING ENGINE DNS prefetching only helps in prefetching the Domain Name of the Server (DNS) for which request can be made, so time to convert domain name into IP address has been saved. It means that web pages are not prefetched in LISTENER PATTERNS GENERATION actually, only the DNS h\as been prefetched 18. Hence, the technique not actually accomplishes the task of prefetching. It just reduces the latency time to some extent. Mozilla Firefox also provides the facility to its CACHE PREFETCH PATTERNS users to control DNS prefetching 19. Firefox 3.5 (Web objects (Download Web Objects) performs DNS prefetching. It performs only domain repository) name resolution. Google chrome also has a prefetching Fig. 2. Process of Prefetching feature similar to the DNS prefetching scheme 20. Copyright © 2015 MECS I.J. Information Engineering and Electronic Business, 2015, 5, 7-12 10 Analysis and Design of Prefetching Framework for Mozilla Firefox From web log data will be extracted and grouped Fig. 3 represents the OMPT (Optimized Mozilla according to the users and their access behavior. For Prefetching Technique). prefetching, hints are being passed to the web prefetch module via L2 cache. These hints are basically the user information. Depending upon these hints, users past requested objects will be passed on to the DNS mapping module. D. DNS Mapping Module Users know about the domain names instead of IP addresses of the web objects. IP address is a 32-bit address used to define different networks in World Wide Web 21. So, whenever user’s put any request to the browser, they use domain names and browser also stores the domain names. So, in this step, domain names will be converted into their corresponding IP addresses, before applying any of the association rule mining algorithm. E. Prediction by Partial Matching (PPM) PPM is one of the efficient association rule mining Fig. 3. Optimized Mozilla Prefetching Technique algorithm. On the hints passed by previous module, association rule mining will be applied to find out the sequence of the web objects to be prefetched, which are likely to be requested by the users. PPM algorithm is used VI. RESULT ANALYSIS for pattern finding. PPM will find out the patterns to be Present section provides the result analysis of Mozilla prefetched depending upon the comparison of current prefetching scheme and OMPT. Depending upon the context with each Markov model 22. user’s scenario, prefetching capabilities of both the F. Patterns techniques has been analyzed graphically. There are two important parameters to compare the prefetching results PPM results in the frequent patterns generation. These of both the techniques: precision and recall. Precision in patterns are the expected future requests of the users, basic terms means number of web objects correctly found depending upon their access behavior. So, whenever a from the set of prefetched objects. Recall is the number of user re-enters in the browsing phase, their corresponding correctly fulfilled web objects requests from the number pattern is found and web objects are prefetched and kept of user requests from the cache. Precision is shown in in the browser cache. It’ll improve the chances of cache equation (1) and Recall is shown in equation (2). hit and reduce web latency too. Precision = Prefetch hits / Prefetches (1) G. Inverse DNS Mapping Module As in the DNS mapping module domain names were Recall = Prefetch hits / User requests (2) converted into the IP addresses, so now for the understanding of user and making matching of the Fig. 4 represents Precision vs. Threshold graph for request provided by the user, IP addresses of the web Mozilla and OMPT, whereas Fig. 5 represents Recall vs. objects to be prefetched will be converted into their Threshold graph. Fig. 6 depicts the Latency vs. traffic respective domain names. This is the task of reverse graph and Fig. 7 show the energy vs. Cache size. Fig. 8 domain in the networking 22. Now, these web pages focused on Cache Hit Rate vs. Prefetch Threshold. will be prefetched and stored in the browser cache. H. Optimal Page Replacemnet algorithm L2 cache is used in the design and prefetched web objects are stored here. It has limited size. So, pages need to be replaced in it after sometime. For page replacement, optimal page replacement algorithm is used. Optimal page replacement algorithm has the lowest page fault rate. It works well in the cases where the system has prior knowledge of future requests. Hence, it is best for the prefetching case. In this algorithm, the pages that will be used the farthest in the future are replaced 23. The algorithm works in two phases: firstly it runs to generate the reference traces and then to actually replace the web objects, that’ll be used farthest in the future. Following Fig. 4. Precision vs. Threshold Copyright © 2015 MECS I.J. Information Engineering and Electronic Business, 2015, 5, 7-12 Analysis and Design of Prefetching Framework for Mozilla Firefox 11 came into the picture in near about 2004 December. Its prefetching technique is the main focus of this paper. The paper presents an analysis of Mozilla Firefox prefetching technique. Two kinds of prefetching techniques are explained viz. link and DNS prefetching. The paper also designed a prefetching framework for the Mozilla Firefox browser. The design consists of L2 cache, web prefetch engine, DNS mapping, PPM, and inverse DNS mapping modules. PPM is used as association rule mining algorithm and it generates the patterns, which need to be prefetched. Accordingly, user requests will be saved in the browser cache. Optimal page replacement algorithm is used for page replacement in browser cache. Result Fig. 5. Recall vs. Threshold analysis is being provided for both Mozilla prefetching technique and OMPT (proposed design). The result analysis came with the fact that OMPT provides better results than the Link prefetching technique of Mozilla. Web latency is reduced more in the OMPT and cache hit ratio is also comparatively better. Proxy server design and implementing the work are under the future scope. REFERENCES 1 A. K. Chauhan, V. Chauhan and R. Gupta, ―Exploring the Web Caching Method to Improve the Web Efficiency‖, International Journal of Computer Applications in Engineering Sciences,vol. 1, issue 4,2011. 2 N. Sharma and S.K. Dubey, ―Fuzzy C-means clustering Fig. 6. Traffic vs. Latency based prefetching to reduce web traffic‖, International Journal of Advances in Engineering & Technology, vol. 6, issue 1, pp. 426-435, 2013, ISSN: 2231-1963. 3 R. Kosala and H. Blockeel, ―Web Mining Research: A Survey‖, ACM SIGKDD Explorations Newsletter, vol. 2 issue 1, June 2000. 4 G. Pallis, Vakali and J. Pokorny, ―A clustering-based prefetching scheme on a Web cache environment‖, Computers and Electrical Engineering 34, Elsevier,pp 309–323, 2008. 5 N. Sharma and S.K. Dubey, ―FP tree use in prefetching‖, Third International Conference on Advances in Computer Science (Elsevier Track), 2nd December, 2013. 6 G. G. Vijayan and S. J. Jayasudh , ―A survey on web pre- fetching and web caching techniques in a mobile Fig. 7. Energy Vs. Cache Size environment‖, Natarajan Meghanathan, et al.. (Eds): ITCS, SIP, JSE-2012, CS & IT 04, pp. 119–136,2012. 7 G. Vijayan Greeshma and J. S. Jayasudha, ―A survey on web pre-fetching and web caching techniques in a mobile environment‖ Natarajan Meghanathan, et al.. (Eds): ITCS, SIP, JSE-2012, CS & IT 04, pp. 119–136,2012. 8 I. Kasthuri, M.A. Ranjit Kumar , K. Sudheer Babu, and Dr. S. S. S. Reddy, ― An Advance Testimony for Weblog Prefetching Data Mining‖, International Journal of Advanced Research in Computer Science and Software Engineering, vol. 2, issue 4,2012. 9 M. Wan, A. Jonsson, C. Wang, L. Li, and Y. Yang, ―A Random Indexing Approach for Web User Clustering and Web Prefetching‖, pp. 40–52, Springer-Verlag Berlin Heidelberg 2012. Fig. 8. Cache Hit Rate Vs. Prefetch Threshold 10 T.V. Sree ramya and V. Sathiyamoorthi, ―Survey on Integrating Web Caching and Pre-Fetching‖ International Journal of Advanced Research in Computer Science and Software Engineering, Volume 3, Issue 4, April 2013, VII. CONCLUSION ISSN: 2277 128X. 11 N. Singh, A. Panwar and R. S. Raw, ―Enhancing the Mozilla Firefox is chosen, as it’s an open source, free, Performance of Web Proxy Server through Cluster Based old and second most popular browser. Mozilla Firefox Copyright © 2015 MECS I.J. Information Engineering and Electronic Business, 2015, 5, 7-12 12 Analysis and Design of Prefetching Framework for Mozilla Firefox Prefetching Techniques‖, International Conference on Authors’ Profiles Advances in Computing, Communications and Informatics (ICACCI), 2013, ISBN 978-1-4799-2432-5. Neha Sarma is Assistant Professor in 12 N. Sharma and S.K. Dubey, ―A Hand to Hand Department of Information Technology, Taxonomical Survey on Web Mining‖, International Northern India Engineering College Delhi, Journal of Computer Applications, vol. 60, issue 3, 2012, India. Her current research interest is in the ISSN 0975 – 8887. field of web mining. 13 http://www.w3schools.com/browsers/browsers_stats.asp accessed last on December 2013. 14 http://en.wikipedia.org/wiki/Link_prefetching accessed last on January 2014. 15 D. Fisher and G. Saksena, ―Link Prefetching in Mozilla: A Server-Driven Approach‖, Netscape/AOL, 2002. Dr. Sanjay Kumar Dubey is Assistant 16 http://en.wikipedia.org/wiki/Talk%3ALink_prefetching Professor in Department of Computer accessed last on December 2013. Science and Engineering in Amity School of 17 https://support.mozilla.org/en-US/kb/Firefox-cant-load- Engineering and Technology, Amity websites-other-browsers-can accessed last on December University Uttar Pradesh, India. He has 13 2013. year of teaching experience. He is member 18 https://developer.mozilla.org/en/docs/Controlling_DNS_pr of IET and ACM digital library. He has a efetching accessed last on February 2014. rich academics & research experience in 19 http://bitsup.blogspot.in/2008/11/dnsprefetching-for- various areas of Computer Science. His research areas include Firefox.html accessed last on March 2014. Object Oriented Software Engineering, Soft Computing, HCI, 20 http://dev.chromium.org/developers/design- Cloud Computing and Data Mining. He has published more documents/dns-prefetching accessed last on December than 73 research papers in indexed National & International 2013. Journals including ACM and in Proceedings of the reputed 21 A. Tanenbaum, ―Computer Networks‖, Prentice Hall International/ National Conferences (including IEEE Explore). Professional Technical Reference, 2002. These publications have good citation records. He has authored 22 J. Domenech , A. J. Gil, J. Sahuquillo, and A. Pont, 3 books also. In addition, he has also served as a Technical ―Evaluation, Analysis and Adaptation of Web Prefetching Program Committee Member of several reputed conferences. Techniques in Current Web‖, Universitat Politecnica de Valencia, 2007. 23 A. Silberschatz, P. B. Galvin and G. Gagne,‖ Operating System Concepts‖, Wiley Publishing, 2008. How to cite this paper: Neha Sharma, Sanjay Kumar Dubey,"Analysis and Design of Prefetching Framework for Mozilla Firefox", IJIEEB, vol.7, no.5, pp.7-12, 2015. DOI: 10.5815/ijieeb.2015.05.02 Copyright © 2015 MECS I.J. Information Engineering and Electronic Business, 2015, 5, 7-12

Advise: Why You Wasting Money in Costly SEO Tools, Use World's Best Free SEO Tool Ubersuggest.