Business continuity can take many forms, from simply making sure
key personnel are contactable at all times to having an separate
facility built to relocate to in case of a disaster. Yet one aspect
that is nearly always consistent is data backup and recovery:
making sure vital information vital is saved in a safe place, and
can be recovered when needed. The latest technology to provide
benefits for backup and recovery is
virtualisation, allowing multiple machines to be created in a
single physical box.
The benefits virtualisation provides to a
business continuity plan are quite straightforward. Rather than
having to spend large amounts of money and time duplicating
physical systems, organisations can put in place a virtual disaster
recovery environment. The major benefit comes from the ease and low
cost of setting up virtual machines; it is much easier to create
"spare" machines, and to have them distributed more widely, greatly
reducing the chances that a disaster will cripple the business.
However, although virtualisation has the potential to be a
wonderful tool for business continuity, it is not without its
problems. The primary concern is one of management: virtual
machines are not like physical ones, and it can be very easy to
lose track of when and where they are created. Other than a small
notification of the original action, there need not be any record
of a virtual machine's creation. As a result, it can be easy to
either create multiple unnecessary copies of machines, or to
overlook an area of data that has not been duplicated. Left
unchecked, virtualisation can easily grow out of control, with a
sprawl of virtual machines which may or may not contain vital data.
For the purposes of business continuity, this is quite evidently
unacceptable.
On top of this, the tools and processes used to manage backup
and recovery of physical servers simply are not appropriate for
virtual machines. For example, traditional backup systems are
predicated on the old one-to-one ratio: one physical server to one
operating system, with one agent on each to back up the content. In
the virtual world, the ratio is one-to-many. It could easily become
one physical server to ten operating systems, with ten agents - a
recipe for real performance problems. Why make one physical server
bear the burden of ten backup processes, when only one is required?
Newer generation backup products designed with the virtual
"one-to-many" ratio in mind are much more efficient, both in
backing up and recovering. In addition, organisations will find
that when they need to recover data, and the clock is ticking, data
might be lost and recovery takes far too long. Without the correct
tools, when backing up and recovering virtual machines only large
amounts of data from a single point in time can be moved, as a
single file - a somewhat inelegant "all or nothing" approach.
Luckily, solving these problems need not be difficult or
expensive. As with most management, common sense will solve the
majority of problems and tools to fix others are readily available.
First, organisations need to rigorously map their virtual machines.
IT must work to identify and record all virtual architecture and
then have a system for monitoring to ensure they are alerted to any
changes. Protocols must be put in place regarding the creation of
virtual machines and these must be followed and enforced to the
letter. Creation should be standardised as much as possible, both
to increase efficiency and, more importantly, to make it easier to
control. With correct monitoring and reporting, it will be easy to
ensure that all relevant data is backed up and that machines are
only present where they are wanted, rather than spread at
random.
Secondly, IT departments should make sure that they have the
correct tools for backing up and restoring virtual machines. While
this may involve more initial expense than attempting to make do
with older, physical-system-based tools, this pales into
insignificance compared to finding out that your virtual machines
have not been backed up when disaster strikes. It is not just about
backup, though; restoration is also key. Organisations should be
testing to ensure that they have the correct tools and processes in
place to bring applications and data on backed up systems on line
within minutes if and when needed.
Implementing virtualisation brings along a sea change in
managing IT. As long as they understand this, organisations can
ensure that their business continuity takes full advantage of this
new technology, rather than being hamstrung by it. By using virtual
machines for disaster recovery, and ensuring that management adapts
its methods and tools to cope with the technology, businesses can
rest assured that their operations are more secure than ever
before.
Ratmir Timashev is CEO of Veeam Software