How important is uptime?
For mission-critical businesses, they will strive for High Availability, typically targeting "five nines" of uptime (about 5 minutes per year of downtime), usually for websites/services. They invest quite a lot of money to ensure that their services stay online when others will go down. Most common is medical and financial applications that do this. But when five nines (99.999%) uptime isn't good enough, what do you do? When you simply cannot have any downtime, how do you ensure things keep working?
First of all, aiming for 5N (five nines) of uptime is usually quite expensive and can be complicated to setup. You have to ensure that everything in your network, systems, applications and more is able to automatically "fail over" to something else if there is a problem. Most companies that go this route will spend quite a lot of money on this. But in some situations, High Availability (HA) isn't good enough, what then?
Enter Continuous Availability. Systems that have no noticeable downtime, even with multiple failures. There are several ways to achieve this, both in programming, hardware and a combination of both. There are vendors that release hardware and software that runs in Lockstep, where application instructions going all the way down the hardware layer to the processor across one CPU is mirrored on another. There are products that use virtualization to simulate one physical machine when there are in fact more than 1 running identically. There are ways to achieve this with programming and combinations of DevOps that can help get you closer to this Continuously Available (CA) goal.
Off the shelf or custom?
If you have the money, especially if you are starting from scratch, it is going to be easier to purchase an "off the shelf" system instead of creating your own custom solution. No matter which way you go, you are going to need a strong DevOps team to implement your solution and it will depend on the size of the environment. To make things easier, you will want to design the system to be Continuously Available from the beginning, rather than trying to retrofit a solution onto something else if you can. There are many variables that can make it more attractive to use one method over the other, you will need to take many things into account before you decide.
To really know what to do, you'll likely need to consult with an expert.
Need more help?
Use the contact form on the right or give us a call to see how RDA can help your business solve your High Availability or Continuous Availability problem today!
High Availability on Wikipedia
Continuous Availability on Wikipedia
Lockstep Computing on Wikipedia
XenServer and Continuous Availability
Stratus everRun products
Tandem Computers - examples of early lockstep computing