Cloud vs. Tin - Downtime and Disk Failure
When looking at your infrastructure and deciding on Cloud vs On Premise Servers (Tin), one thing that’s often overlooked is the cost of potential downtime caused by hardware failures, most commonly disk or RAID failures.
If you have a highly transactional database (like a CRM or stock management system) and if your physical server has a failure, then what does that downtime cost the business while you are waiting for a replacement part and installation?
Single disk failure could put your RAID array in danger
In our experience, if your server is relatively new, and assuming you opted to take a suitable warranty/support agreement from the manufacturer you would still be looking at downtime.
However, in cases where clients have older tin or it’s out of warranty, if one disk fails you have to replace it with a identical disk, finding and replacing this could take time and the extra load on the other disks can create a domino effect putting your entire RAID system in danger of failing!
In more than one instance we’ve seen this happen and have known companies to come to us having spent weeks trying to rebuild their RAID array.
Being out of action could cripple your business
24 hours in business is a long time but weeks of being out of action can be financially crippling to a business. Most companies’ backup their servers and have a DR plan in place, but on premise disk and subsequent RAID array failure can happen at any point and is a risk to business continuity.
In one of the worse cases of RAID array failure we’ve seen, a company came to us after being told specialist parts from HP where expected to take up to 4 weeks to be delivered! It all ended well though as we migrated their infrastructure and exchange service into the cloud and had them up and running within a few hours.
Granted hindsight is a wonderful thing but if their infrastructure was already in the cloud then the server would’ve migrated over to a different physical server in the cloud whilst the disk was being replaced. So, no noticeable downtime.
With on premise severs there are other scenarios which often occur which require a level of physical investigation, such as mother board failures and RAM issues – these can take days to resolve.
In the event of a disaster where you need to replace your on premise server(s) or if your production box has failed, it could be a week.
Could your business cope if this happened to you?
You’ll be glad to hear there are solutions, and a lot of companies mitigate against these very scenarios with failover kit. However, with on premise servers this proves to be expensive as upfront costs are doubled. It’s also inefficient as you have servers sitting idle doing nothing until its needed.
Why not have your fail over server in the cloud, and pay a small reservation fee each month ready for when disaster strikes? If for whatever reason you need onsite tin, there’s nothing to stop you replicating it into a datacentre so when your onsite infrastructure fails, cloud takes over and your company can still be productive.
We're here to help
We’re always available for assistance, guidance, and advice. If you want us to provide an impartial evaluation of your infrastructure and offer solutions in the event of failure or a disaster, then just complete the information and we'll get back to you.
We appreciate that no two businesses are the same and there are many factors to consider when moving your infrastructure to the cloud; are you looking for replication and fail over, are you looking to move some services / elements or your entire infrastructure.
Once again, we’re here to help with anything from scoping, building, testing, monitoring and tweaking resources to make sure you get the performance you want within the budget you can afford.