To limit the caching
problem that DNS clients may impose, they provide DNS records with very small TTLs
(time-to-live). These solutions, as it happened with cluster-based solutions, are just concerned
about routing client requests to the proper servers, but again without actual support
to make the servers cooperate.
As a result, we have shown that there are many solutions for high availability in the market,
some of them for the large scale and some others for local area networks. Some of them are
intended to run on clusters and some of them are to be used as run-time support to build
applications. Interestingly, most of those solutions lack support for automatic strong replica
consistency enforcement.
Automatic.Replication.Architecture.Overview
The main goal of the architecture that we detail in this chapter is to enhance the availability
of existing Web-based applications, without the need of specific hardware or proprietary
software, without requiring the source code to be modified, allowing the combination of the
architecture with other highly-available solutions such as clusters and round-robin DNS,
and providing automatic strong replica consistency.
The full architecture can be thought of as a software module that distributes HTTP requests
to the set of pre-configured replicas.
Pages:
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547