Browse Prior Art Database

Mechanism for creating edge servers from cooperating P2P nodes

IP.com Disclosure Number: IPCOM000010768D
Original Publication Date: 2003-Jan-20
Included in the Prior Art Database: 2003-Jan-20
Document File: 4 page(s) / 69K

Publishing Venue

IBM

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 34% of the total text.

Page 1 of 4

Mechanism for creating edge servers from cooperating P2P nodes

The invention provides a mechanism by which independent nodes in the internet web can be used as edge servers on demand, thus reducing the cost of maintaining and managing dedicated edge servers as well as having better distribution of load in through the network.

PROBLEM SOLVED BY THIS INVENTION (Summary) Web servers often face unpredictable surge in demands that not only brings down the quality of service provided by the web servers but also leads to network congestion. Existing solutions that attempt to ameliorate this problem use one of the following two techniques. The first is to quickly scaling up the central web server facility and then using load balancing to distribute the load among the cluster of central servers.

The second technique is to redirect requests to edge servers, which then perform a subset of roles of the central web servers. However, the above two solutions have the drawbacks that they are expensive to manage, they do not provide an instantaneous solution to a sudden spurt in demand on the central servers and they lead to excess capacities in form of server power and bandwidth as both the solutions need to maintain excess capacities to handle sudden rise in demand.

The invention provides a mechanism to use existing node capacities to handle unexpected spurt in demand on central web servers.

Prior Art

1. An existing solution to the problem is to quickly scale up the central web server facility and using load balancing to distribute the load among the cluster of central servers. This technique solves the problem of excess load on existing servers.
2. Another existing technique is to redirect requests to dedicated edge servers, which then perform a subset of roles of the central servers. This solves both the issues of excess load on the web server as well as network congestion since the traffic gets redistributed to the edge of the network.

3. Uprizer solutions : KARMATM moves popular information closer to users who want it. KARMATM achieves this by breaking up large files into "data chunks." These "data chunks" are widely distributed throughout the network and are retrieved hyper-locally on edge computers demanding the content. KARMATM content distribution products achieve this by
- shortening the path between the content and the end-user.

- storing the content traffic locally on KARMATM nodes. - reducing content providers' bandwidth costs and total cost of network ownership. - optimizing network performance

Deficiencies in Prior Solutions All these systems have the following deficiencies.
a) They are expensive to manage and maintain
b) They do not provide an instantaneous solution to a sudden spurt in demand on the central servers
c) They lead to excess capacities in form of server power and bandwidth as both the solutions need to maintain excess capacities to handle sudden rise in demand.

d) They do not allow the distributed nodes to price their r...