The Need for Disaster Recovery and DR Basics
Modern organizations have various applications that are used internally by partners
and by customers. These applications range from ones that are “nice to have” but not
essential to doing business to those that would shut down the company if not
available. Even the briefest outage of these business-critical applications can cause
organizations harm in multiple ways, including the following:
Financial loss through not being able to perform normal business functions
Damage to reputation through publicly visible outages that erode confidence in the
organization for external parties
Potential compliance gaps to regulatory requirements
It is therefore important to ensure that business-critical applications are always
available, both within the primary datacenter through high-availability technologies
and in alternate locations through disaster-recovery technologies. (Often a single
technology can be leveraged for both high availability and disaster recovery.) To
provide disaster recovery, the data related to an application must be available in the
alternate location, which means data replication. There must be a means to run the
application and connect to it, so compute and network resources are required.
It is important to understand which applications are critical to the organization, and
that can be ascertained only with the involvement of the business groups. Once the
business-critical applications are identified, you must understand the dependent
applications and services of those applications, because protecting the business-
critical applications without their dependencies would result in a nonfunctional
solution in the event of a system outage or disaster scenario.
As an example, consider a typical line-of-business application that may run on one or
more application servers. That application may leverage a SQL database that runs on a
separate infrastructure, it may publish services through a corporate reverse proxy that
is Internet facing, and it may require Active Directory for authentication. For the line-
of-business application to be functional, all of those dependent services must be
available. In fact, when planning for high availability and disaster recovery, it’s
necessary to protect the applications and services that the target application depends
on at the same or higher protection level.
There are many ways to provide resiliency and availability to services locally, within a
location and between locations, and there is no single “best” technology; rather, it is
important to utilize the best availability technology for specific applications and
services. Many availability solutions leverage the Failover Clustering feature that was
covered in the previous chapter. A cluster-enabled application is protected from the
failure of a node and will either seamlessly transition to another node or restart on
another node without any administrator intervention.