When we think of catastrophe’s like car accidents or fatal illnesses we tend to think “that’s what always happens to other people”. Because of this, we are shocked when these things happen to us but the important thing to remember is that to someone else, you and I are the other people that these happen to. It’s no different with our IT environments. If you don’t have a disaster recovery plan in place now, part of your justification may be that you think it will never happen to you but the reality is that disaster can strike anywhere and doesn’t always look like we would think it does.
First, we need to clarify what a disaster means in terms of your datacenter. When you hear disaster your mind likely jumps immediately to hurricanes in the Gulf of Mexico, Earthquakes in California, tornadoes in the Mid-West or structure fires, but hey, that’s what the sprinklers are for right? However IT disasters can be due to much more than just natural disasters. They can include cyber security breaches where a hacker steals, takes hostage, or even wipes your data, or they can come from a disgruntled employee who just found out today was his last day doing the same. Now these are all extreme situations but an IT disaster can even come a system failure that sets off a chain reaction of failures around your environment. At its base, an IT disaster is any event or series of events that cripples or destroys your data.
So that’s what an IT disaster looks like but why do you need a plan in place to protect yourself? Studies have shown that 80% of businesses that are hit with a disaster close their doors within 18 months, and 90% are closed within 2 years. I wish there was a way I could give you to be 100% safe from a disaster occurring but, while I can’t do that, you can be prepared to deal with a disaster before it puts your data in a critical state.
Now that you know you want to set up a DR environment you’ll have to decide what type of site you want. There are a number options utilizing both cloud options and co-location options. For co-location DR you have the option to utilize a hot site, cold site, or a warm site, each of which has its own advantages and disadvantages.
The Cold Site
A cold site essentially means that you either have a data center that has all the racks, cooling, and power supplies you need but no storage or servers, or you have this all installed but everything is turned off. As far as co-location options go, this is naturally the cheapest as you’re only paying for the space or at most the physical infrastructure but saving on your opex by not using any power and not paying for resources to manage the site on a day to day basis. However, because the site is turned off this option isn’t going to work well for an organization with a small RTO due to the need to start up and configure your infrastructure and load up your backups. However if you have data workloads that you need to protect but don’t necessarily need to be up within a particular timeframe, a cold site may be an effective, low-cost option for your disaster recovery environment.
The Hot Site
On the other end of the spectrum we have the DR hot site. As the name would imply, if a cold site is infrastructure that is sitting in a data center but not turned on, a hot site is a second operational site. While certainly more expensive to operate, as you need to purchase not only additional infrastructure but resources to manage and operate the infrastructure as well as deal with increased power and cooling costs, this option allows you to mirror your production site, minimizing downtime in the event of a disaster. By replicating your critical workloads to your hot site in near real time, you don’t have to worry about your production site getting hit by a disaster because your data is already running in the other location. This setup is perfect if you have a short RTO and have the extra funds to keep a second site fully running.
A Warm Site
In between these two extremes sits the warm site, clever, I know. The warm site takes the best of both setups and minimizes the cons. Whereas in the cold site nothing is on and the hot site everything is running just like your main data center, in the warm site you have a backup datacenter installed and it’s on, but nothing is happening. With this setup you’re able to spin up the new datacenter more quickly than a cold site but not have as large of expense as you would if you were running a hot site.
DR in the Cloud
Now, you probably remember that I mentioned cloud options earlier right? So now that you’ve got your strategy for your co-location DR, let’s talk about how you can run your DR in the cloud with the concept of the Pilot Light. The Pilot Light is essentially the concept of running a hot or warm site by utilizing the cloud. So instead of paying all of the capex and opex costs associated with keeping a second datacenter up and running while mirroring your environment, you pay for a cloud environment and replicate your critical workloads to one of their locations. Now, if you’re production site is compromised by a disaster, you can immediately turn around and spin up a production environment in the cloud. Prefer to have an on-premise production site? No problem, if you need the short RTO but like the cost savings of a cold site, then utilizing a pilot light environment can allow you to quickly spin up your production environment in the cloud and then once you get your cold site set up and ready to go, you can move your workloads back onto your new on-premise data center.
Now that you know how many options there are for a disaster recovery environment and how important it is to have a disaster recovery plan, let’s have a conversation about which setup is right for you and how you can start your path towards being confident in the face of disaster.