News

Backup and Recovery Solutions

 

Let us assist you in your planning, design, and implementation of your Disaster Recovery / Business Continuity project.  Varrow can offer solutions to failover (in minutes) complete datacenters to “hot” off-site locations.  To accomplish these solutions, Varrow uses appropriate technologies from the following vendors; VMWare, VirtualIron, Citrix, Neverfail, 2X, Platespin, Microsoft, Juniper, Cisco, Netriplex, UltraDNS, VizionCore, ESXpress and more.

We are now in a new paradigm of infrastructure computing and networking that allows us to accomplish recovery time objectives (RTO) and recovery point objectives (RPO) we never thought possible or affordable in the case of disaster or hardware/software failure.

Here is an article written by a Varrow engineer that describes how the typical approach to disaster recovery has changed.  This article is about how Virtualization can play a key role in your DR/BC strategy.

For the case of this writing, BC = Business Continuity = (DR)Disaster Recovery (there are cases where BC and DR should not be interchanged but for this writing, consider them interchangeable)

Why Virtualize?  Just BC!

Business Continuity (BC) plans and server virtualization go hand-in-hand.  This article provides a great introduction to the ‘Virtual Computing’ concept as well as providing useful information for those in need of creating an off-site BC plan as it applies to the IT infrastructure.

Yesterday’s BC Plans:

The typical company’s IT infrastructure couldn’t survive a total site-loss disaster.  When asked, most company’s said they couldn’t afford to mirror their physical IT server infrastructure, server-for-server, at an off-site location.

Some companies had an off-site disaster recovery infrastructure, but it typically took them hours or days to complete a full emergency fail-over and even more time typically to test the failover.

Today’s BC Plans:

Almost all leading companies can now survive a total site-loss disaster, and by leveraging virtual computing technology, almost all companies can justify the costs involved.  These companies have off-site data centers using server virtualization technologies that allow them to have one physical server at their off-site location for every ~10 to 20 physical servers at their production location.  The most agile of these companies can fail-over to their off-site datacenters in minutes instead of hours or days (because of geo-clustering technologies that can mirror data over low bandwidth links).  And they can fail-back to their production facility just as quickly if or when the disaster has passed.

What happened to Yesterday?

There are two proven technologies that make today’s ‘smarter BC plans’ a reality; virtual computing, and geo-clustering/replication solutions.

Virtual computing, also known as hardware/server virtualization, or infrastructure virtualization, allows a single piece of hardware to present itself as many pieces of hardware.  This concept, while built on software, would not have become a reality would it not have been for today’s hardware performance advancements.  In the typical production datacenter today, many physical servers run at less than 30% of their potential, most at less than 10% of potential.  Server virtualization technology was built to help use the other 70% to 90%.  Imagine putting 10 or 20 or even 30 unique servers, with different names and identities and functions, all onto one physical computer.  Imagine being able to reboot and manage each of those servers independently as if they were each a physical computer.  For those who are not quite ready to take virtualization technology to their production floor, then why not use virtualization technology for R&D?  Once comfortable with virtualization in the R&D environment, the next logical step is to use virtualization in the disaster recovery or BC plan.  Now instead of needing 1-for-1 physical servers at the disaster recovery location, an off-site replica of the production physical server infrastructure can be a created with purchasing only 1 physical server for every 10 servers that were needed in the production location.  And since most disaster recovery sites don’t run at the same heavy user load during disaster-times as during non-disaster-times, the reduction of server resources won’t be a problem.  An added bonus to this virtualization design, is that the off-site servers can still be used for R&D during non-disaster-times.  Varrow uses both the free VMWare Server product and the enterprise VMWare VI3 product as necessary to build the off-site or on-site virtual computing solutions for their clients.

Geo-clustering (or geographically separated clustering) is the other technology required to make today’s ‘smarter BC plans’ work.  Imagine a solution that could replicate email data or accounting/ERP data to an off-site location all-the-time.  Imagine that the data at the disaster recovery location was always in-sync-to-the-minute with the production data.  In many cases, no LAN-speed links are required for replication.  A simple broadband speed VPN tunnel from the production facility to the off-site facility can sustain the replication needs of many companies.  Now imagine that this replication solution could not only replicate the data but also help to replicate the physical server identities and software updates to the off-site virtual environment.  For example, a physical server named SERVER1 needs to be configured for off-site failover.  The off-site location would actually also have a virtual server named SERVER1.  The virtual SERVER1 would not be visible to the network unless the physical SERVER1 were to become unavailable.  The physical SERVER1 and virtual SERVER1 would always be in sync all-the-time.  Any changes in software or data on physical SERVER1 would be replicated to virtual SERVER1.  The virtual SERVER1 would take over in times of emergency and the data on the virtual SERVER1 would be correct up-to-the-time of site loss.  End users could now work from home or emergency offices and all that would be required for access to the emergency fail-over system is a computer with a web browser.  Users would simply go to http://emergency.fakecompany.com , and the appropriate applications would be available to them using technologies from Citrix or other MS Terminal Server technologies.   Varrow partners with Neverfail Group to implement geo-clustering solutions using the Neverfail product line.

Products used in many Varrow BC solutions come from the following partners: VMWare, Citrix, 2X, LeoStream, Platespin, Microsoft, Juniper, Cisco, Netriplex, UltraDNS, VizionCore, Neverfail and more.