Wcol: WWW Collector Home Page
| *** Contents *** |
|---|
| Overview (this page): Features | Platforms | Source | History | Links |
| Technical Topics : Prefetching concept | Proxy combinations |
| Developer's Information : ICP_NOTIFY |
The Wcol is a prefetching proxy server for WWW. The program prefetches referred pages from user retrieved page. It improves most of WWW latency.
The Wcol runs on most popular UNIX. You get a fast WWW service using Wcol.
The Wcol runs on most popular UNIX.
Its requirements are below:
WcolD and older version of Wcol runs more platforms.
WcolE will run such platforms.
WcolE is tested the following systems:
Features and Advantages from Other Proxies
Platforms
Please let me known that WcolE runs your system or not.
Source and Distributions
History
The Wcol have developing since 1994. The prefetching proxy server was big challenge for network service. Prefetching requires many technique and many resources. There were many trial of implementation. We got good prefetching proxy server and knowledge about prefetching through them. However, it is not best. We challenge to new features.
A: prototype of caching (1994/6,7)
- caching
- do fork() each request -- fork() generates large load. process table was run over.
B: prefetching (1994/8)
- HTML parser -- parse anchors(A-tag) and inline images(IMG-tag).
- conversion
C: asyncronus connections (1995/2)
- threading by asyncronus connection -- file descriptors were run over. busy loop is appeared.
D: pre-fork and fd-passing (1995/9)
- pre-fork -- reduce number of fork().
- file descriptor passing -- no file descriptor starvation.
- garbage collection -- delete unused or old resources.
- prefetching background images and figures -- parse background(BODY-tag), and figures(FIG-tag).
- catalyst mode (1996/10) -- separation of prefetching engine from proxy server
E: asyncronus and pre-fork (1997/1...) -- Current generation, wcolC+wcolD
- handle a large connections
