Electrical computers and digital processing systems: multicomput – Computer-to-computer data routing
Reexamination Certificate
2001-06-29
2003-03-04
Geckil, Mehmet B. (Department: 2142)
Electrical computers and digital processing systems: multicomput
Computer-to-computer data routing
C709S246000, C707S793000, C370S394000
Reexamination Certificate
active
06529960
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to an improved data processing system and, in particular, to a data processing system for database maintenance.
2. Description of Related Art
Enterprise messaging requirements are evolving beyond traditional store-and-forward e-mail to include the integration of groupware/workflow applications and tight coupling with the “browsing” model of corporate intranets. Another key trend, made possible by the proliferation of Internet standards, is ubiquitous access to information across standards-based networks and data stores. At the same time, the messaging infrastructure must be extended beyond the enterprise to business partners, customers, and suppliers to provide a return on investment in electronic messaging technologies.
As a result of these new imperatives, enterprise and inter-enterprise message traffic is expanding quickly beyond the limitations of disparate legacy systems, loosely coupled mail and intranet systems, and the multitude of gateways connecting them. Indeed, companies are faced with the task of consolidating heterogeneous e-mail systems, delivering access to new information sources, and building a robust messaging infrastructure that meets current and expected enterprise requirements.
A known enterprise messaging solution is Lotus Notes®, which is a platform-independent, distributed client-server architecture. Domino® servers and Lotus Notes® clients together provide a reliable and flexible electronic mail system that “pushes” mail to recipients using industry standard message routing protocols, and that facilitates a “pull” paradigm in which users have the option to embed a link to an object in a message. The object can reside in a Domino® database, an HTTP-based “intranet” data store, a page on the World Wide Web, or even a Windows® OLE link. Lotus Notes® also tightly integrates groupware applications.
Groupware connects users across time and geography, leveraging the power of network computing. Ironically, networks present one of the biggest challenges to groupware implementation. Connections are sometimes unavailable or inadequate for demanding tasks. While this can be due to failure, there are many other reasons including, without limitation, mobile users, remote offices, and high transmission costs. Groupware has to keep users working together through all these scenarios. The technology that makes this possible is so-called replication. A replication mechanism puts information wherever it is needed and synchronized changes between replicas.
Thus, for example, using Lotus Domino® replication services, an organization wishing to deploy a Web application to multiple locations may set up servers in each location. As data is changed in each location, the architecture ensures that databases are synchronized through replication. As another example, a salesperson who pays frequent visits to customer sites also needs to stay connected to the databases and information at her home office. When she leaves the office with a laptop computer, she makes a copy or replica of the lead tracking and customer service databases that she needs. While out of the office, however, other account managers may make changes to the server database at the same time that she is making her own changes. The salesperson can re-synchronize the two by replicating again over a telephone connection. All updates, additions and deletions that were made to the server after she left the office are now replicated to the laptop database, and all updates, additions and deletions she made on the laptop database are replicated back to the server database. The replication process also detects update conflicts and flags them for the salesperson and other users to reconcile.
There are several known replication techniques that allow workgroup users to connect to a local server and at the same time keep information synchronized across geographically dispersed sites. Documents in the replicated database are composed of fields. When two servers desire to synchronize their respective version of a given document, the most recent field entry for each field of the document is often used for replication purposes. If timely replication is desired, updates to one replica are propagated to other replicas as soon as possible. For example, in Lotus Notes® Clustering Release 4.5, replication is effected by having every server convey an update to every other server whenever a local update occurs. This approach suffers from the drawback of not being readily scalable. Another approach is “scheduled replication”, wherein a pair of servers periodically wake up and compare data sets. This requires every data set on both servers to be compared and is strictly a two way operation. Scheduled replication is costly and cannot be done in a timely fashion, and it also creates a significant amount of undesirable network traffic.
Other known techniques (e.g., Microsoft Exchange®) provides a simple, first generation messaging-based replication scheme. This technique relies on store-and-forward mail to push changes from one server to other defined replicas on other servers. There is no comparison operation, however, to guarantee that replicas remain synchronized. Such a system significantly increases administrative and end-user burden. Moreover, if a user changes even a single property or field of a document, the entire document must be replicated rather than just the property or field. Netscape Suitespot® uses proxy servers for locally caching Web pages, which reduces network bandwidth requirements. This technique, however, is merely duplication—copying files from a distant place to a closer place—and there is no relationship between the copies. It is not a true replication mechanism.
There remains a need to provide enhanced replication schemes that address the deficiencies in the prior art.
SUMMARY OF THE INVENTION
It is a primary object of this invention to replicate data in a timely manner across a large number of nodes.
It is another primary object of this invention to provide replication enhancements in a distributed system that significantly reduce network traffic.
It is still another primary object of this invention to provide high performance, near-realtime replication in a geographically-dispersed network topology.
Still another primary object is to provide a simple replication mechanism that is highly scalable.
A particular object of this invention is to configure a replication mechanism within a hub and spoke network architecture.
Still another particular object is to enable sliding window acknowledgment through the hub on a broadcast to nodes in the network architecture.
Yet another object of the present invention is to enable spokes to issue periodic acknowledgments to the central hub, and for the central hub to issue periodic acknowledgments to the originating spokes, wherein such acknowledgments effectively indicate the vitality of the nodes within the system as well as any need for packet retransmission.
Still another object of this invention is to provide a hub recovery mechanism to enable a set of server nodes to transfer operation from a failed hub to a new central hub. Also, the invention provides a spoke failure mechanism for isolating a failed hub from a group of destination nodes that are targeted to receive update(s), and for selectively re-admitting the failed hub back into the destination group.
A further object of this invention is to implement a multilevel replication mechanism, for example, wherein first level hubs participate as second level spokes in a recursible architecture extendible to any depth desired.
Still another object of this invention is to enable a plurality of updates to be batched, collected and distributed together by the replication mechanism.
Another more general object is to synchronize multiple database replicas in a distributed computer environment. These databases, for example, may be local replicas of databases on a large number of servers, registry information servers, domain name servers, LDAP direc
Chao Ching-Yun
Hough Roger Eldred
Shaheen Amal Ahmed
Burwell Joseph R.
Geckil Mehmet B.
International Business Machines - Corporation
Judson David H.
LaBaw Jeffrey S.
LandOfFree
Method and system for replicating data in a distributed... 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 system for replicating data in a distributed..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for replicating data in a distributed... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3052743