DDG: An Efficient Prefetching Algorithm for Current Web Generation

Abstract

Abstract-Web prefetching is one of the techniques proposed to reduce user's perceived latencies in the World Wide Web. The spatial locality shown by user's accesses makes it possible to predict future accesses based on the previous ones. A prefetching engine uses these predictions to prefetch the web objects before the user demands them. The existing prediction algorithms achieved an acceptable performance when they were proposed but the high increase in the amount of embedded objects per page has reduced their effectiveness in the current web. In this paper we show that most of the predictions made by the existing algorithms are useless to reduce the user's perceived latency because these algorithms do not take into account how current web pages are structured, i.e., an HTML object with several embedded objects. Thus, they predict the accesses to the embedded objects in an HTML after reading the HTML itself. For this reason, the prediction advance is not enough to prefetch the objects and therefore there is no latency reduction. As a result of a wide analysis of the behaviour of the most commonly used algorithms, in this paper we present the DDG algorithm that distinguishes between container objects (HTML) and embedded objects to create a new prediction model according to the structure of the current web. Results show that, for the same amount of extra requests to the server, DDG always outperforms the existing algorithms by reducing the perceived latency between 15% and 150% more without increasing the computing complexity.

Key factors in web latency savings in an experimental prefetching system

“…However, when the predictor is at the proxy server, it takes advantage of the multiuser and multi-server information gathered at this element to perform the predictions [8,13,16]. Finally, if the engine is located at the server, it makes predictions based on multi-user accesses to the same website [6,17,3,18].…”

“…The main objective of this comparison is to show that the benefits achieved by real proposals are still far from the performance capabilities of web prefetching. These algorithms are the Prediction by Partial Match (PPM) [5], the Dependency Graph (DG) [3], the Double Dependency Graph (DDG) [6], and the Referrer Graph (RG) [24].…”

“…Intensive research works focusing on prediction and prefetching algorithms to improve web performance have been published, for instance [1][2][3][4][5][6]. However, few studies [7][8][9][10] focus on the maximum performance achievable through web prefetching and the main constraints to reach it.…”

The Impact of the Web Prefetching Architecture on the Limits of Reducing User's Perceived Latency

“…This engine can be located at any part of the web architecture: clients [7], proxies [2], [1], and servers [8], [5], [4], or even in a collaborative way between several elements [3]. To make the predictions, a high amount of algorithms that learn from the past access patterns to predict future ones have appeared in the literature, e.g., [4], [3]. These patterns of user's accesses differ depending on the element of the architecture in which the prediction engine is implemented.…”

“…A proxy where user's requests are predicted and requested to the server in advance is described in [2], whereas [3] proposes different levels of proxies and servers collaborating to predict users' requests. Other works [4], [5] implement a system where servers predict future accesses and then, clients download those objects in advance.…”