1、<p>　　4300單詞，2.1萬英文字符，5500漢字</p><p>　　出處：Broder A Z, Najork M, Wiener J L. Efficient URL caching for world wide web crawling[C]// 2003:679-689.</p><p>　　Efficient URL caching for world wide

2、 web crawling</p><p>　　Andrei Z. Broder, Marc Najork, Janet L. Wiener</p><p>　　ABSTRACT Crawling the web is deceptively simple: the basic algorithm is (a)Fetch a page (b) Parse it to extract

3、all linked URLs (c) For all the URLs not seen before, repeat (a)–(c). However, the size of the web (estimated at over 4 billion pages) and its rate of change (estimated at 7% per week) move this plan from a trivial progr

4、amming exercise to a serious algorithmic and system design challenge. Indeed, these two factors alone imply that for a reasonably fresh and complete crawl of the web</p><p>　　Our main conclusion is that cach

5、ing is very effective – in our setup, a cache of roughly 50,000 entries can achieve a hit rate of almost 80%. Interestingly, this cache size falls at a critical point: a substantially smaller cache is much less effective

6、 while a substantially larger cache brings little additional benefit. We conjecture that such critical points are inherent to our problem and venture an explanation for this phenomenon.</p><p>　　Keywords: Ca

7、ching, Crawling, Distributed crawlers, URL caching, Web graph models, Web crawlers 1. INTRODUCTION A recent Pew Foundation study [31] states that “Search engines have become an indispensable

8、utility for Internet users” and estimates that as of mid-2002, slightly over 50% of all Americans have used web search to find information. Hence, the technology that powers web search is of enormous practical interest.

9、In this paper, we concentrate on one aspect of t</p><p>　　Both the stream of local URLs and the stream of URLs received from peer crawlers flow into the Duplicate URL Eliminator (DUE). The DUE discards URLs

10、that have been discovered previously. The new URLs are forwarded to the URL Frontier for future download. In order to eliminate duplicate URLs, the DUE must maintain the set of all URLs discovered so far. Given that toda

11、y’s web contains several billion valid URLs, the memory requirements to maintain such a set are significant. Mercator can be config</p><p><b>　　中文譯文</b></p><p>　　萬維網(wǎng)爬行的有效URL緩存</p&

12、gt;<p>　　摘要 要在網(wǎng)絡(luò)上爬行非常簡單：基本的算法是：（a）取得一個網(wǎng)頁（b）解析它提取所有的鏈接URLs（c）對于所有沒有見過的URLs重復(fù)執(zhí)行 （a）-（c）。但是，網(wǎng)絡(luò)的大?。ü烙嬘谐^40億的網(wǎng)頁）和他們變化的頻率（估計每周有7%的變化）使這個計劃由一個微不足道的設(shè)計習(xí)題變成一個非常 嚴(yán)峻的算法和系統(tǒng)設(shè)計挑戰(zhàn)。實(shí)際上，光是這兩個要素就意味著如果要進(jìn)行及時地，完全地爬行網(wǎng)絡(luò)，步驟（a）必須每秒鐘執(zhí)行大

13、約1000次，因此，成員檢測 （c）必須每秒鐘執(zhí)行超過10000次，并有非常大的數(shù)據(jù)儲存到主內(nèi)存中。這個要求有一個分布式構(gòu)造，使得成員檢測更加復(fù)雜。 一個非常重要的方法加速這個檢測就是用cache（高速緩存），這個是把見過的URLs存入主內(nèi)存中的一個（動態(tài)）子集中。這個論文最主要的成果就是仔細(xì) 的研究了幾種關(guān)于網(wǎng)絡(luò)爬蟲的URL緩存技術(shù)。我們考慮所有實(shí)際的算法：隨機(jī)置換，靜態(tài)cache，LRU，和CLOCK，和理論極限：透視cac

14、he和極 大的cache。我們執(zhí)行了大約1800次模擬，用不同的cache大小執(zhí)行這些算法，用真實(shí)的log日志數(shù)據(jù)，獲取自一個非常大的33天的網(wǎng)絡(luò)爬行，</p><p>　　關(guān)鍵詞：高速緩存、爬行、分布式爬蟲URL緩存、網(wǎng)絡(luò)圖模型,網(wǎng)絡(luò)爬蟲1.介紹 皮尤基金會最新的研究指出：“搜索引擎已經(jīng)成為互聯(lián)網(wǎng)用戶不可或缺的工具”，估計在2002年中期，初略有超過1半的美國人用網(wǎng)絡(luò)搜索獲取信息。因此，一 個強(qiáng)大的搜

15、索引擎技術(shù)有巨大的實(shí)際利益，在這個論文中，我們集中于一方面的搜索技術(shù)，也就是搜集網(wǎng)頁的過程，最終組成一個搜索引擎的文集。 搜索引擎搜集網(wǎng)頁通過很多途徑，他們中，直接提交URL，回饋內(nèi)含物，然后從非web源文件中提取URL，但是大量的文集包含一個進(jìn)程叫 crawling 或者 SPIDERing，他們遞歸的探索互聯(lián)網(wǎng)?；镜乃惴ㄊ牵篎etch a pageParse it to extract all linked URLsF