Electrical computers and digital processing systems: multicomput – Distributed data processing – Processing agent
Reexamination Certificate
1999-01-20
2004-06-08
Dinh, Dung C. (Department: 2153)
Electrical computers and digital processing systems: multicomput
Distributed data processing
Processing agent
C709S217000, C709S239000
Reexamination Certificate
active
06748416
ABSTRACT:
FIELD OF THE INVENTION
This invention relates generally to improving the availability and performance of network-mediated applications. More particularly, it relates to a client-based dynamic server switching method for use in a distributed system including multiple intermediary servers.
BACKGROUND OF THE INVENTION
The demand for remote networked services is steadily increasing. Many networked applications must access intermediaries in order to reach remote services. Network intermediaries include security firewalls, file caches, and directory services. For example, many corporations establish security firewalls to protect their internal networks from unauthorized access, e.g. from the Internet. At the same time, internal users need to cross these firewalls to access remote services, for instance, to access public search engines (e.g. for price comparisons in procurement). One common way to enable crossing the firewall boundaries, is for client applications (e.g. Web browsers) to access SOCKS servers. These servers are trusted network intermediaries that are enabled to communicate, across the firewalls boundaries, with remote Internet services. Typically, client applications that access remote services across the firewalls are statically configured to use only one such intermediary, e.g., a given SOCKS server. Hence, such applications become dependent on the performance and availability of the given intermediary in order to perform many of their functions. Therefore, if the SOCKS server fails, all of the Web browsers using that socks server will be unable to communicate with any remote server outside the firewalls. From the end-user perspective, connectivity to the Internet has been lost, while in fact, that is not necessarily the case, since an alternative SOCKS server may be able to provide the equivalent service at the same time.
A second problem faced by these client applications is that many of their communications are often unnecessarily being routed via intermediaries. For example, a Web browser may be statically configured to access a given SOCKS server for all remote services, even if some of these services are “inside the firewall”, that is, are local to the network domain, and hence do not require access via a SOCKS server intermediary. Such requests create unnecessary load on the intermediaries, which results in additional delays for all interactions.
A third problem with prior art approaches is the lack of configurable policies to allocate constrained network resources. For example, on a bandwidth-constrained network, the download of a large news file may significantly slow all other interactions via the network. In particular, real-time applications, such as live broadcasts of multimedia streams may be unnecessarily delayed, since the current Internet infrastructure is lacking in effective priority mechanisms. It is desirable to provide a means and method for controlling the priority and pace of requests and acknowledgments of IP packets at the clients.
In order to improve the performance and availability of networked applications, multiple functionally-equivalent intermediary servers are needed to provide service to a large number of clients. It is also desirable in such a system to direct clients transparently to one of a multiplicity of intermediary servers to obtain the requested service; and, if an intermediary is not functioning as desired, the client request would ideally automatically be routed to another intermediary.
One method known in the art that seeks to increase the performance and availability of networked applications, is to create a cluster of computing nodes (also called a “multi-node cluster”) to share the load. This multi-node cluster approach is particularly useful for increasing the processing capacity at “hot” sites on the Internet. For the sake of consistency herein, the term “the Internet” refers to the collection of networks and gateways that use the Transmission Control Protocol/Internet Protocol (TCP/IP) suite of protocols, TCP/IP being a well known protocol for communications between computers (see e.g., D. E. Comer,
Inter-networking with TCP/IP, Principals, Protocols and Architecture
, Prentice Hall). The multi-node cluster is encapsulated, such that it is made to appear as one entity to clients, in order to make the added capacity provided by the multi-node cluster be transparent to clients. Client requests are distributed among nodes in the multi-node cluster. Many load balancing techniques are known in the art; see for example, Dias et al., “A Scalable and Highly Available Web Server”, Proc. 41st IEEE Computer Society Intl. Conf. (COMPCON) 1996
, Technologies for the Information Superhighway
, pp. 85-92, Feb. 1996; see also U.S. Pat. No. 5,371,852, issued Dec. 6, 1994 to Attanasio et al., entitled “Method and Apparatus for Making a Cluster of Computers Appear as a Single Host.”
A key problem with the static allocation of intermediaries is that it often results in a poorly balanced workload, thereby wasting network resources. For example, one SOCKS server can have a large number of clients simultaneously connecting to it while others are serving only a few clients. The workload across the intermediaries can become skewed and highly imbalanced, potentially causing significant performance degradation due to server overloads. In the event that an intermediary server fails or is overloaded, there is no automatic mechanism for redirecting a client to another server. A need exists, therefore, for a client-based means to automatically switch to an alternate intermediate server in order to receive remote services without interruption, in the event of service degradation or failure of a network intermediary.
Similar problems have been addressed by (1) server-side solutions, like Network Dispatcher, and (2) application-specific solutions. A server-side solution does not address the same problem, since it does not have any control over the clients. An application-specific solution can address a similar problem for a given intermediated application, at the cost of increased complexity in the application, and the intermediate servers.
What is desirable, and is an objective of the present invention, is to provide a solution to the foregoing problems, which solution is client-based and can support an arbitrary set of networked applications.
An objective of the present invention is, therefore, to provide a local client-resident intermediary broker, that can intelligently route the traffic among intermediaries so as to: (1) provide better availability of the intermediary services, (2) share the load of requests across several intermediaries (e.g. load balancing), and (3) bypass intermediaries whenever possible.
In addition, it is an objective of the present invention to provide remote administrative control of the local intermediary brokers, enabling them to utilize their shared, limited networking resources, according to specific policies. For example, important interactive data, such as e-commerce traffic can be given higher priority than non-critical information. Remote administrative control will, thereby, prevent the “tragedy of the commons” syndrome, where each client tries to locally maximize its private utilization.
Accordingly, it is yet another objective of this invention to provide an improved client-based system for accessing network intermediaries.
SUMMARY OF THE INVENTION
These and other objectives are realized by the present invention wherein a method and an apparatus are implemented as an intelligent local broker, the client agent, that can passively keep track of the state of network intermediaries, and can allocate network traffic according to different policies. In addition, an optional management platform running at a policy server provides an active network monitor of intermediaries and implements administrative policies. Under the present invention, an intermediary-aware application (e.g., a browser) will be configured to use the local client agent instead of a specific intermediary. The policy server can remotely load conf
Carpenter Geoffrey C.
Goldszmidt Ziman German S.
Dinh Dung C.
Dougherty Anne V.
Perez-Pineiro Rafael
LandOfFree
Client-side method and apparatus for improving the... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Client-side method and apparatus for improving the..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Client-side method and apparatus for improving the... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3314785