More and more I see companies who are backing up their data in quality ways, using a combination of good equipment, strong processes, and industry recommended best-practices. Make no mistake – I’m happy about this. But I’m also seeing many of these same companies who are considering their backup to be their disaster recovery plan. Unfortunately, no data backup, no matter how good, is a disaster recovery plan. Now, I’m not trying to preach about disaster recovery…well, okay, that’s a lie. I’m planning on doing a little preaching. But truly, there is a difference between a disaster recovery plan and standard data backup plan. Ultimately, disaster recovery is the means to an end with the “end” being business continuity. And everyone likes business continuity, right? So whether you decide to put together a full-blown disaster recovery plan or just stay with your current data backup, knowing the difference can save your business lots of money. Okay, I’m done preaching. Here’s some valuable information about what makes up a good disaster recovery plan, and how to put one together…
Note: A true disaster recovery plan includes phones, facilities, data, and more. I’m just focusing on the data portion of things in this article.
Generally, “disasters” come in three flavors:
- Hardware failure (bad hard drive, motherboard, network card, etc.)
- Software failure (“bluescreen” is a familiar term for a reason)
- Physical failure (flood, fire, etc.)
So let’s keep things simple. Good disaster recovery generally includes a mix of the following components:
- Quality data backup (preferably onsite and offsite)
- Imaging
- Redundancy
Let’s explore each of these in a bit more detail.
Data backup:
- Onsite – use a hard drive backup system (preferably a quality internal/external chassis that has removable drives). In terms of reliability, ability to restore quickly, and length of service, hard drives are simply better than comparable tape systems.
- Offsite – Offsite backup solves a number of security & operational concerns, and is very cost effective these days (about $1 per GB per month is pretty common). Because it’s automated, you don’t need to worry about people remembering to take data offsite and you can control the amount of data that is backed up offsite.
Imaging: Imaging is the process of taking a “snapshot” of an operating system. When applied to server systems, the use of imaging can literally save tens of thousands of dollars in recovery costs. Imaging fails when it comes to restoring data granularly (a single e-mail, a single document, etc.), but for true disaster recovery, imaging can’t be beat.
Redundancy: Creating a network devoid of “single points of failure” is almost always too expensive to employ, but the concept can be used effectively to significantly reduce the chance and effect “failure” can have on your business. Some good uses of redundancy would be:
- Quality server design (multiple hard drives, fans, power supplies, etc.)
- Use of a quality SAN (Storage Area Network) in a virtualized server environment (VMware software is great in these environments)
- Server redundancy (two servers deployed in a fully redundant architecture). Stratus Avance is amazing software that accomplishes this at a price small and medium sized businesses can absorb easily.
- Network switches deployed in a balanced architecture with enough capacity to assume full connectivity should one fail.
The last piece of a good disaster recovery plan is the plan itself. Now that you have all these quality pieces in place, you still need to have a plan in place should a disaster strike. Here are a few good ways to start putting your plan together:
- Know your operational costs (by the day, by the hour, etc.). Granted, this is only half the equation (operational costs don’t include the impact downtime will have on your clients, nor does it include other opportunity costs you may have), but it is an absolutely necessary ingredient for creating your plan.
- Define your tolerance for downtime. Some businesses can handle a day or two of downtime. Others can’t afford an hour of downtime. Determining how your tolerance for downtime will go a long way in putting your disaster recovery plan together
- Define a budget. If you know your operational costs are $25,000 per day (for instance), you then have good information from which you can create an appropriate budget for avoiding downtime.
- Design your network appropriately (data backup, imaging, redundancy, etc.).
- Put it on paper. Just like everything else in business, putting it on paper helps make it happen.
Tags: data backup, disaster recovery, disaster recovery plan, disaster recovery planning, hardware, Information Technology Professionals, phone systems, Virtualiztion
Networking, Technology Mangement, Technology Mangement | Joe Ulm |
March 3, 2011 5:06 pm |
Comments (0)
It makes no difference whatsoever how amazing a technology is if it doesn’t provide a defined value for your business. Below are the typical values businesses can expect to receive from deploying virtualization technology:
- Agility – Virtualized environments provide an incredible platform for adapting your company’s technology quickly in support of changing business goals and direction.
- Scalability – Adding new functionality to your company’s technology almost always presents numerous, critical, and complex concerns ranging from loss of productivity to significant capital expenditure. Virtualization simplifies this issue by eliminating hardware concerns from the equation. In fact, a properly designed virtualization infrastructure will support relatively large changes in your company’s technology with very little complexity or out-of-pocket expense in comparison to a non-virtualized environment.
- Disaster Recovery – The ability to take an “image” of your server has been around for some time. This ability is valuable from a disaster recovery standpoint because instead of rebuilding your server from scratch, you can simply reinstall the existing image of your server on a different piece of hardware which saves incredible amounts of time. However, taking an existing image and installing it on a different hardware platform has always been at best difficult, and at the worst, nearly impossible. With a virtualized environment, however, this concern is eliminated because virtualization breaks the one-to-one ratio between your server hardware and software.
Note: Management – Managing your company’s technology is a significant expense regardless of the size of your company or the market you’re in. Yet, because technology is critical to productivity for almost every business these days, it is an expense that is absolutely necessary. Virtualization doesn’t necessarily add to the management time required, and in certain circumstances, it can help reduce management time by removing the complexity of managing different hardware platforms. However, it can also add some processes to your existing management routine, and if you don’t have professional, in-house IT support, you may need to leverage the services of your existing IT consultant a little more.
From a very high level virtualization architecture is pretty straight forward. You take one very powerful server and you install virtualization software on it. Once you have the virtualization “layer” installed, you then install any desired server systems on top of it. This effectively creates a platform for a one-to-many hardware-to-software relationship. Of course, all the hardware and software specifications need to be taken into account cumulatively to properly design and deploy a virtualized environment.
The simplest way to explain virtualization may be: the ability to run multiple operating systems on a single piece of hardware. Of course, this is my personal definition and is specific to the context of how SMB’s typically use virtualization. For a more formal definition, Wikipedia has a pretty good one. Anyhow, in practice, this means that if your business employs 3 physical servers running Windows Server operating systems you could purchase one server and run all three instances of your Windows Server operating systems on that single piece of hardware.

At first glance it may not seem that revolutionary, but virtualization is changing the way technology is used across all market segments. Really.
Virtualization as a technology has been around for quite some time, but it has only become prevalent within the SMB community over the past few years. There are many reasons for this including cost, market viability, network infrastructure requirements, and more, but those barriers have been mostly removed over the past few years, making virtualization one of the most interesting and potentially valuable technologies for business today. Yet, as we enter 2011, many businesses still don’t understand the values virtualization can provide them, nor do many businesses understand how to analyze whether virtualization is a good fit for them. Well, read on and find out the practical aspects of utilizing virtualization technology for your business.
There’s no doubt that virtualization is becoming more and more prevalent for small businesses of all sizes. There are lots of reasons for this including:
- Reduced downtime
- Increased performance
- Strong disaster recovery
- Scalability
- Excellent business flexibility
But I find it curious that many companies I meet with are initially confused by the concept. So…here’s my attempt to simplify virtualization.
At its most basic level, virtualization for small businesses generally involves running multiple server operating systems on one physical server. That’s it. Oh sure, more than one physical server can be deployed to create better redundancy, less downtime, etc. But just knowing that simple concept – multiple server systems running on a single piece of hardware (server) – puts you way ahead of many small businesses.
You may have heard or read the hype over the last 10 years about the “cloud” and “virtualization” to consolidate your network hardware. It all sounds “neat” and a little confusing but how does it really help? Let me share a real story of a time when virtualization has made a difference in our business. (Note names have been changed but the facts in this case are real)
I had been working hard getting a 2 NFR licsenses from Lexis Nexis Time Matters enterprise working on 2 servers, one in our main office in Milwaukee and one in our Madison office. The purpose was to demonstrate the use of SQL replication technology. That is SQL’s ability to keep to SQL databases perfectly in sync. This mixed with Time Matters 9.0 Enterprise can create 2 or more databases all over the U.S. which stay in sync in real time but individuals can still connect locally not using VPN, Terminal Server or Citirix. Neat stuff and very useful for some of our mid to large clients.
I digress. I was using a vmware virtual machine of Server 2003 to demonstrate mostly because I could work on it on my mac in vmware fusion when I had time and then could ultimately transfer it to a more powerful machine in the Madison office whenever I needed it. Using a virtual machine or a “VM” allowed me to configure the server and have all those settings maintained no matter which machine I hit “play” on. Really it can be that simple. Vmware stores the entire OS, settings, files in a matter of a couple files which can be transfered and with vmware player (free, http://www.vmware.com/download/player/) you push a green play button and your in action.
True story, the morning of the demo for the client the hard drive on the machine I was going to demo it on failed. I did not panic I did not even call our main office. Instead I drove to our datacenter with the vm on a usb hard drive. Plugged it into one of our existing servers and hit “play”. The demo was never reschedule on our client and I think our sales team at ITP may have even been a little impressed.
Vmware, Hyper-V or Xen – virtualization is coming to the SMB market and to the home user. I answer virtualization questions daily from friends, family and colleagues. I called a maintence guy in Florida to ask about the internet setup for one of our traveling clients and he said “Hey, you got a minute for a question about vmware fusion…?”
I will work up a post on virtualiztion basics so that we can demystify virtualiztion. It is not marketing this technology makes business sense.