Web Caching Overviews

Web Caching is used to improve the Internet faster and save bandwidth to the Internet with a Proxy Server using a Web store that was then called the Proxy Server.

Although many people do not know the word web caching fact, but they have to use web caching Browser and then managed by the browser, most documents will be collected and how that information came from. web on memory or disk each time you click the browser’s back button to return to previously viewed web pages viewing. Page will be from memory, which does not need to draw from a Web Server is a caching-style first class (first level of caching).

Basic idea of caching is to store documents that a user can run up to the position that can be used together. To the second user to use the same set of documents to retrieve from the origin server directly to another user that is second to retrieve documents from co-location (common place), instead of this nature is. necessary when your organization is very large corporations in Europe. And inbound traffic from the other side most of the Atlantic is linked to a lower speed.

The main advantage of the method is ??? can use documents from the Web quickly and reduce the traffic to the Internet This medium-sized organizations with 50-100 users, the size of caching can provide up to 60% of URL requests from the local cache.

Differences between the browser cache with proxy-cache server is a browser cache to cache for a single user and is in a position in the final user workstation as proxy-cache is a program that represents a number of web browser. clients without the client allowed to read a document that had been requested by others before this proxy-cache server is typically a common server situated between the local network with the Internet browser will request documents from a proxy server. 1 shows the structure of this type of proxy-cache server.

proxy-cache is not only a solution to solve the problem of bandwidth. But is also required when organizations want to file insurance for Wall’s network security. In this case, the proxy-cache on the Server that all clients can access the browser. But it is a separate client machine from all networks with the Internet Server is a must have 2 network interfaces to act as the internal network to external network access from Internet network access only to specific external network Figure 2. Show config this work by using a proxy-cache server should be allowed only within a specific network to use caching.

Multi-Level Web Caching

Another step in the advancement of Web Caching is the concert, I made a set hierarchy (cache hierarchy), in which connection may be used to connect a hierarchy and a proxy-cache server from the 2nd to the proxy-cache with a role that might be different in each two-tier It depends on the form of network (network topology), resources, systems and policies of the ISP that a cache nearby. (neighor (sibling) cache) will provide a unique document in the cache itself and then only in the order above (parent cache) can make it from pulling from the other cache. or higher in order of source (Source) depend on whether the cache is connected to the other parent or neighbor cache or not in class or not the parent cache should be used when not in use. The document can be retrieved from the cache in the same class.

Choose the cache topology that is very important because they can reduce traffic on the network, an organization can choose to have several sibling cache in each sub-unit and the parent cache in only one parent cache. part near the Internet network, the parent cache can be ?????? a document request from the parent cache other than the above example, the ISP has an agreement to establish a parent cache or sibling cache between users. ISP can reduce the traffic that is more than (overload) in the link or make traffic a web path through the non-IP traffic to a common web caching should be considered in the application-level, how do routing, using ICP (Internet cache protocol) is a pro cow collagen below the main image is an example of a multi-level web caching in the organization.

What’s ICP

ICP (Internet Cache Protocol) protocol used for communication between web caches is a lightweight protocol on top of the UDP will be used by ICP to determine the location of the web object in the cache side (neighboring caches) transfer. between the cache object is made by most TCP-based HTTP protocol, but to determine where to draw the object must use the quick and easy.

Cache in one of the best location to download the cache object will send ICP request packet to the parent cache sibling cache and cache, all these will send a reply with ICP HIT code HIT or MISS code which purport This is a cache object and agreed to provide the means MISS no object to the cache so that a request would know who is the object itself and it will want to take this information together with other factor such as round-trip time or a response each time. To choose whether to pass HTTP  requests to cache if the cache? MISS packet is a reply to a request that the cache will make a request from the parent cache document processing requests and replies ICP (ICP request / reply exchange). should occur in a short period, otherwise the cover will increase the time (latency time) to the browser, but this time the user often can not notice.

If the requested object through ICP object is small enough, it can be included in the HTTP redirect ICP HIT reply like this but do not are a party much use of the ICP will be used when the network is a multi-level cache that contains the parent cache and only sibling cache, using ICP needless work for the network in Figure 1 and 2 in the event of a cache only if the cache or the requested document from the higher-. lever cache the same thing that ICP is not only overhead.

The Multi-level Web Caching on Squid.

Squid can do on our Multi-level Web Caching by a config file / etc / squid / squid.conf. The value to set is called “cache_peer” This is a 2-cache_peer is

1. Parent Cache is a cache that can get documents from another cache in order to get higher than a document from the source (source), which has to be drawn from the cache or from source? Fact ???? Will depend on that the network has many Parent cache or a cache in the same class (neighbor) or not use this type of cache should be used when no opportunities to get documents from the cache in the same class.
2. Sibling Cache (may be called a neighbor) is a cache of documents that are currently only available in a self that is only if the documents requested are not in use can not glue himself to service.

The model uses a command cache_peer is as follows: 
cache_peer hostname type http_port icp_port.

Examples of active such as 

cache_peer parent.foo.net parent 3128 3130 [proxy-only]. 
cache_peer sib1.foo.net sibling 3128 3130 [proxy-only]. 
cache_peer sib2.foo.net sibling 3128 3130 [proxy-only].