. General-purpose.hardware: The proposed solution does not imply to use any particular
cluster or expensive hardware. It can be deployed over general-purpose hardware,
much cheaper, and powerful enough for most applications.
2.. No. software. recoding: Many highly-available solutions require a particular Web
server, or there is a need to implement the software using some specific programming
language, or there is a need to make substantial modifications to the running software.
In all those cases, cost due to the highly-available support increases. The solution
proposed in this chapter allows some applications to be plugged to the highly-available
solution, without any of those requirements.
3.. Unconstrained.replica.placement: Most current solutions for Web applications use
clusters placed in local area networks. Even though the availability level achieved
that way may be sufficient, local area networks should be considered themselves as
points of failures. The Internet provider??™s network connection may fail, or it may suffer
disasters such as flooding, fire, or long power outages. The architecture explained in
the next pages will allow system administrators to place replicas at distinct administrative
domains, so that unavailability risks can be reduced.
Pages:
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549