After the main concepts behind high availability are covered, the third section
discusses existing highly-available solutions for Websites. Once the pros and cons of those
existing solutions are mentioned, section four will explain the proposed architecture that
provides automatic replication for Web applications at low cost. Limitations and advantages
of the model are detailed, and a set of common scenarios highlight the range of situations
where the architecture would do its best. Following the architecture overview, section five
explains each component, giving enough details to allow an expert reader to experiment
with them. Finally, section six concludes the chapter with some remarks about automatic
replication.
High-Availability Concepts
High availability is usually pursued by server replication. The same piece of server software
is placed into different computers, so that in case failures occur in one server, the other servers
can take over its tasks. Clients access the replicated service by sending invocations to
the servers. When invocations reach the server set, they coordinate themselves to perform
the requested actions and send the corresponding replies to the clients. As there can be a
number of concurrent clients requesting services, server coordination has to be designed
carefully to show an acceptable performance.
Pages:
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529