<title>Transcoding characteristics of Web images</title>

Abstract

Transcoding is a technique employed by network proxies to dynamically customize multimedia objects for prevailing network conditions and individual client characteristics. Transcoding can be performed along a number of different axes and the specific transcoding technique used depends on the type of multimedia object. Our goal in this paper is to understand the nature of typical Internet images and their transcoding characteristics. We focus our attention on transcodings intended to customize an image for file size savings. Our results allow the developers of a transcoding proxy server to choose the appropriate transcoding techniques for the important classes of Internet images.We analyze the characteristics of images available on the Web through a representative trace. We show that most GIF images accessed on the Internet are small; about 80% of the GIF images are smaller than 6 KBs. JPEG images are larger than GIF images; about 40% of the JPEG images are larger than 6 KBs. We also establish the characteristics of popular image transcoding operations. We show that for JPEG images, the JPEG compression metric and a transcoding that reduces the spatial geometry are productive transcoding operations (saves at least 50% of the file size for 50% of the images). Our systematic study of image characteristics leads to some surprising results. For example, a naive spatial geometry reduction of GIF images by a factor of 2 along each axis actually causes to an increase in the file size compared to the original image for 40% of the images. Thus it is important to understand the characteristics of individual images before choosing the proper transcoding operation.

A distributed architecture of edge proxy servers for cooperative transcoding

“…Workload 1 is based on proxy traces belonging to the nodes of the IRCache infrastructure. We can assume that Workload 1 captures a realistic scenario, because we found that the statistical characteristics of the images contained in the workload (e.g., file size, JPEG quality factor, number of colors of GIF images) are very close to the results reported in [4]. Workload 2 is characterized by image files much larger than those contained in Workload 1, thus requiring in average a transcoding cost about three times higher.…”

“…The classes of devices range from high-end workstations/PCs which can consume every object in its original form, to midrange PCs or laptops, set-top boxes, hand-held PCs, personal digital assistants, and cellular phones. We consider that transcoding is applied only to image objects (both GIF and JPEG formats), as more than 70% of the files requested in the Web still belongs to this class (e.g., [4]). In the system model, we assume that traditional devices, such as PCs and laptops, are still more popular than other client devices; hence, 25% of client requests come from PCs not requiring any transcoding, 25% from laptops with low or null necessity of content adaptation, while the remaining 50% of requests is equally divided among the other four classes of devices with major transcoding requirements.…”

Content Adaptation Architectures Based on Squid Proxy Server

“…We downloaded the resources from their content servers and placed them on our content server. We performed some characterization on the images in the light workload, such as file size, JPEG quality factor, and number of colors of GIF images, and found that they are very close to the results reported in [6]. Hence, we can assume that our workload captures a realistic scenario of present Web requests.…”

“…In this paper, we consider that transcoding operations are applied only to image resources (both GIF and JPEG formats), as more than 70% of the files requested on the Web still belongs to this class [6]. However, in our experiments we also consider a likely future scenario where the transcoding operations have higher costs because of the larger percentage of non-textual resources.…”

“…The nature of typical images in Web workload and their transcoding characteristics have been analyzed in [6], where the authors found that 74.81% were GIF and 24.41% were JPEG images. GIF and JPEG images represent almost all the image To obtain a file size reduction, the transcoding algorithms seem to perform better if operating on spatial geometry, color depth, quality, animation, and MIME subtype.…”