Electrical computers and digital processing systems: multicomput – Computer network managing – Network resource allocating
Reexamination Certificate
1998-05-29
2002-01-22
Meky, Moustafa M. (Department: 2153)
Electrical computers and digital processing systems: multicomput
Computer network managing
Network resource allocating
Reexamination Certificate
active
06341311
ABSTRACT:
BACKGROUND OF THE INVENTION
1. The Field of the Invention
The field of the present invention is that of the proxy servers used in connection with accessing data over the World Wide Web (“WWW”) through the Internet or other Wide Area Network (“WAN”). More particularly, the present invention involves an array of multiple proxy servers configured together to act as a single distributed cache of information identified through the use of Uniform Resource Locators (“URL”). Specifically, the invention treats the routing of a URL request received by one array member to the array member having the cached information associated with the URL.
2. Present State of the Art
Generally speaking, the concept of a “cache” or “caching” as used in computer terminology and applications typically means making a more accessible copy of some piece of data for a performance advantage. For example, information that is cached is in many instances more accessible than it otherwise would be so that processing speed is increased since accessing the cached information is quicker.
Since having excess copies of data can create its own sort of overhead and because cache size is limited, data is not typically retained in a cache indefinitely and will eventually be overwritten after a certain amount of time if the cache is being fully utilized. This may occur according to a Leased Recently Used (“LRU”) algorithm, an expiration time, or any other relevant criteria for a particular application. Caching commonly exists at the microprocessor level with instruction and data caches so as to avoid excessive accessing of system RAM and may also exist elsewhere in a computer system or in a network of computer systems.
Another form of caching is commonly used in relation to accessing data or information over the WWW. A user having Internet access can directly receive a URL identified data object, such as a web page, and then display it locally on a browser. Each time the user receives such a data object, there is a delay as the Hyper Text Transport Protocol (“HTTP”) request travels across the Internet to the location identified by the URL and the destination server processing the URL request responds with the requested data object. Because the data object can be quite large, the time to access the data objects with the HTTP request and response creates a significant amount of information traveling over the Internet connection causing delays and excess overhead.
In some organizations, many of the same data objects are requested by the various users within the organization. It is therefore common to introduce a proxy server that will receive user access requests from a client application, such as a web browser. Referring to
FIG. 1
, the use of a proxy server acting as a cache is shown. A client application
20
will direct all URL requests to a proxy server
21
that will serve as a cache for any information (i.e., URL data objects) associated with the URL. Should the proxy server
21
not have the URL data object within its cache, it will, in turn, make access over the Internet
22
to the destination server found within the URL in order to place the data object into its cache and then respond to the client
20
URL request. Thereafter, should another client make a request to the proxy server
21
, it can be serviced directly from the cache without necessary access over the Internet to the destination server found in the URL itself.
Note that the client
20
can be any software application capable of communicating or directing requests to the proxy server
21
using the HTTP protocol and would include other proxy servers, web browsers, Internet “enabled” applications, etc. Furthermore, HTTP requests contain a variety of information that can be used to “force” the proxy server
21
to access the URL data object over the Internet in order to assure that the “freshest” copy has been accessed. Such operation of proxy servers and their use as caches are generally know in the art for use in servicing IITTP requests.
It should be noted that the term URL may indicate either the address of where a data object is originally located and accessed, or the data object itself. To distinguish, a URL itself would be an address or location of the data object whereas a URL data object would be the actual web page file that is transmitted across the Internet to the client application. Though the appropriate usage is readily identifiable by context, efforts will be made throughout this application to distinguish between the two.
There are benefits of having a proxy server acting as a cache for URL data objects. One benefit is that for cached items, the total access time for a user is generally reduced since the connection between the client
20
and the proxy server
21
is typically over a Local Area Network (“LAN”) rather than having to access the data object over the Internet or other Wide Area Network (“WAN”). Another benefit is for security purposes so that an organization may have a “firewall” to protect itself from unwanted outside penetration.
Larger corporations and other organizations may have many proxy servers servicing their needs. It becomes desirable in such situations to harness many proxy servers together as a single, logical distributed cache. Ideally, such a single distributed cache would have no duplication of URL data objects contained therein. Furthermore, a single distributed cache should have as little overhead as possible in servicing any given URL request that arrives at a member of the distributed cache. In other words, the actual URL data object may not be residing at the same server that originally receives the URL request and some form of forwarding, routing, or acquisition of the desired URL data object must occur in order to service that original URL request.
One attempt at creating a such a distributed cache is the Internet Cache Protocol (“ICP”) that coordinates the activity of an “array” of proxy servers. Though ICP allows an array of proxy servers to function as a distributed cache, it also has some drawbacks as will be explained hereafter.
Referring now to
FIG. 2
, the interaction of a client with a proxy server array is shown. In such an arrangement, a client
23
will contact one of the proxy servers in the array
24
in order to access URL data objects that are available over the Internet
25
. Typically, a client
23
is assigned to a particular proxy server within the proxy server array
24
and may itself be a proxy server. Since the URL data object requested by a client may exist in a different proxy server than the one contacted, a mechanism or protocol is necessary for routing the URL request from the receiving proxy server to the appropriate proxy server, or in some other way service that URL request.
Referring now to
FIG. 3
, proxy server array that is organized and configured according to the ICP protocol is shown. In the example shown in
FIG. 3
, a client
26
will direct HTTP requests to an assigned proxy server
27
that is part of an ICP proxy server array
28
. Assuming that a desired URL data object is contained in the distributed cache created by the ICP proxy server array
28
and located at the proxy server
29
, a scenario illustrating the operation of ICP is now shown. This scenario will also illustrate a number of problems that make ICP a less than optimal way of creating a distributed cache.
The URL request will originate at the client
26
and be received by the proxy server
27
as indicated by arrow
30
. After determining that the desired URL object does not reside at the proxy server
27
in its local cache storage, a query will be sent out to all proxy servers in the ICP proxy server array
28
as indicated by the query messages path
32
. In turn, every other proxy server that receives the query will give a response back to the proxy server
27
as indicated by the response message path
34
. Each individual response will indicate whether or not the specified URL data object resides at that particular responding proxy server. Note that more than one of the proxy servers i
Hurvig Hans
Smith Brian J.
Valloppillil Vinod V.
Meky Moustafa M.
Microsoft Corporation
LandOfFree
Directing data object access requests in a distributed cache does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Directing data object access requests in a distributed cache, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Directing data object access requests in a distributed cache will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2832074