Electrical computers and digital processing systems: multicomput – Computer-to-computer session/connection establishing – Network resources access controlling
Reexamination Certificate
1997-11-07
2001-02-06
Maung, Zarni (Department: 2758)
Electrical computers and digital processing systems: multicomput
Computer-to-computer session/connection establishing
Network resources access controlling
C709S223000, C709S225000
Reexamination Certificate
active
06185619
ABSTRACT:
COPYRIGHT NOTICE
A portion of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure as it appears in the Patent and Trademark Office patent file or records, but otherwise reserves all copyright rights whatsoever.
BACKGROUND OF THE INVENTION
The present invention relates generally to the field of distributed computer systems, and more specifically to computing systems and methods for assigning requests to one of a multiplicity of network servers based upon best criteria such as the speed of the underlying network infrastructure.
The explosive growth of the World Wide Web needs little introduction. Not only are members of the technical community finding an ever greater number of technical and informational resources available on the World Wide Web, but also the mainstream populace is finding favorite restaurants, car makes and churches sporting new websites. The popularity of the World Wide Web as a communications medium lies in the richness of its information content and ease of use. Information in this medium exists as objects in a widely distributed collection of internetworked servers, each object uniquely addressable by its own uniform resource locator (URL). Since its inception, the World Wide Web has achieved a global prominence in everyday life and commerce.
Yet this explosive growth has not been had without difficulty. The proliferation of commercial applications brings with it an ever increasing number of users making ever increasing numbers of inquiries. The problems of latency and bandwidth constraints manifest themselves in delay, lost information and the distraught customers.
Network architects respond using an array of solutions. Many responses fall within the category of solutions based upon supplying more computing power. This may encompass such alternatives as different web server software, web server hardware or platform, increases in RAM or CPU in the server, application rewrites, or increasing network bandwidth by upgrading hardware. Another class of solutions involves using multiple servers or locating servers strategically. One method in this class is to locate the server at the internet service provider. By selecting a service provider with a optimal pairing capability, colocating the server at the service provider's site can yield a much better connection to the rest of the internet. Another approach is the use of distributed servers. Place identical content servers at strategic locations around the world. For example, one in New York, one in San Francisco, and one in London. This distributes the load to multiple servers and keeps traffic closer to the requester. Another approach is to cluster servers. Clustering enables sharing of hard drive arrays across multiple servers. Another approach is the server farm. This entails the use of multiple webservers with identical content, or the segmentation based upon functionality. For example, two servers for web functions, two for FTP, two as a database and so forth. A variation on the server farm is the distributed server farm. This places server farms at strategic locations—essentially combining the server farm with the distributed server approach.
The multiple and distributed server approaches solve one problem at the expense of creating another. If there are multiple servers, how does the end user locate your site? Presently, names and universal resource locators (URLs) are resolved into unique single addresses by a domain name service (DNS). DNS servers maintain a list of domain names cross referenced to individual IP addresses. However, if multiple web servers or server farms are used, the DNS system must be modified. A common approach to this problem is to modify the DNS system to a one to many mapping of names to IP addresses. Thus the DNS will return a list of IP addresses for any particular web object. These may then be handed out to the various clients in a round-robin fashion. There are, however, several drawbacks to this approach. The round-robin paradigm returns IP addresses in a strict order with little regard to the location of the requester or the server. The scheme has no knowledge of server architecture or loading. The selection simply progresses down a simple list. One server may receive all heavy duty users. Additionally the weakest link determines the overall performance, so server platforms need to be kept in relatively parity. Another problem is that the DNS simply returns IP addresses with no regard as to whether the server to which the address corresponds is operational. Consequently, if one of the round-robin servers happens to be off-line for maintenance, DNS continues to give out the address, and potential users continue to receive time-out error responses. Thus, the round-robin modification of DNS makes a broad attempt to solve the distributed server problem. However, there is no regard to network traffic, low balancing among servers or reliability issues.
Several products on the market purport to address these problems, but these prior efforts all suffer the drawback that they require that the user software environment be modified in order to facilitate replicated server selection. A scheme that requires user software modifications is less desirable due to the practical problems of ensuring widespread software distribution. Such schemes are, at best, useful as optimization techniques.
One such class of approaches are those that rely on explicit user selection to assign a user request to a server. The user application may include additional steps that require the user to have sufficient knowledge, sophistication, and patience to make their own server selection. Such schemes are not always desirable for a number of reasons.
A technique based on selective host routing uses multiple replicated servers, all with the same network address, located at different points in the network topology. A router associated with each server captures incoming network traffic to the shared server address, and forwards the traffic to the specific server. This technique can only statically distribute the client request load to the nearby server, with no consideration of server load or other network characteristics.
The BIND implementation of the Domain Name System server may include techniques to bind server names to different network addresses, where one of a set of different multiple addresses are assigned either sequentially or randomly. Service providers assign a different address to each replicated server, and BIND directs user requests to the alternative servers. This technique can only statically distribute the client request load to an arbitrary server, with no consideration of server load or other network characteristics.
SONAR is an emerging IETF (Internet Engineering Task Force) protocol for distributing network characteristics, in particular, for topological closeness. SONAR includes a data format for representing query requests and responses, but does not specify a mechanism for determining the network characteristics.
The Cisco Local Director is a product that works as a network traffic multiplexor that sits in front of multiple local servers, and distributes new transport connections to each server based on the amount of traffic flowing to the servers. This product does not consider network characteristics in its decision, and further requires that the replicated servers be collocated. Cisco Systems is a company headquartered in San Jose, Calif.
The Cisco Distributed Director redirects user requests to topologically distant servers based on information obtained from network routing protocols. The Distributed Director intercepts either incoming DNS requests or HTTP requests, and provides the appropriate response for redirection. This product does not consider server load, and only considers the restricted set of information available from routing protocols; this information is also limited in accuracy by the aggr
Brittain Jason Alan
Dykes Barry A.
Hotz Steven Michael
Joffe Rodney Lance
Lippard James Joseph
Genuity Inc.
Maung Zarni
Suchyta Leonard Charles
LandOfFree
Method and apparatus for balancing the process load on... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for balancing the process load on..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for balancing the process load on... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2575232