In this sense, its role is somehow similar to that of
Gald??mez
Copyright ?© 2007, Idea Group Inc. Copying or distributing in print or electronic forms without written permission
of Idea Group Inc. is prohibited.
cluster front-ends. However, instead of having just one front-end for the whole cluster, every
replica will be equipped with the proposed software module.
This software module, that we will call HA-module, will distribute HTTP requests made
by the service clients (Web navigators) to the set of replicas, so that, in case one of them
fails, the rest of the replicas will take over the failed one. All this highly-available setting
will be achieved without any specific software requirements for clients and without clients
perceiving significant downtime intervals.
As shown in Figure 1, a Web server without replication is unavailable as soon as the server
fails. At the right part of the Figure, a replicated Web server, as it will be the case using the
HA-module, continues processing client requests using the set of working servers.
As we have mentioned, nowadays there are many highly-available solutions, some of them
actually quite successful; however, the architecture which we propose offers five advantages
that altogether could be important for many Web applications:
1.
Pages:
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548