Opinion: Microsoft Blames Outages on Overheated Servers

15th March 2013

Microsoft recently came forward with an apology for the recent Outlook and Hotmail outages which left customers disconnected and unable to work, blaming the downtime on overheating servers.

Microsoft's vice president, Arthur de Haan, took to the firm's blog with a written explanation of events. The company had been automatically updating the firmware on one of its core systems, a process which had been done successfully on many occasions in the past, when it failed in an unexpected way. The failure caused a temperature spike in the datacentre which triggered safeguards and the shutdown of several servers.

The overheating issues resulted in over 16 hours of downtime and required both software and human intervention to fix.

Maintenance and Support matters...

The reason for the downtime highlights how important server maintenance is and how automatic updates are no replacement for a real engineer if you want to keep systems safe. If you're looking for expert help with maintaining your business servers, why not contact VPW your local Newton Abbot IT and computer support today?

The outage is yet more proof, along with other incidents from major companies like Amazon and Google that really goes to demonstrate that there's no real safety to be found in using the "large" companies when you want to adopt Cloud and Hosted Technology over a smaller provider such as ourselves. In fact, there's a good chance a well setup, professional company (like us, naturally) would be more beneficial to you. Unlike the major companies, we understand the needs of small business and adapt our services to meet your needs. There's an incentive for us to do this since we have to work harder to gain customers and a good reputation, and ultimately we are better geared to helping you - after all you're a customer, not a statistic for us, and if you have an outage, you can be sure we want it fixed and give it the same urgency you would.

Size doesn't mean safe...

It's a common, but misguided viewpoint that somehow there's safety in going with large players - as if they're immune to issues, or would be able to resolve them any faster. This incident alone at 16 hours shows just how easily a problem can escalate and despite the enormous resources of a company like Microsoft, how long an outage could last. The truth is that technology breaks, and is complex. Companies need specialist, capable IT Support to ensure things work well - and are fixed when (not if) they go wrong. We put a lot of effort into our services to make them reliable, but nobody is immune to issues, big or small. When something does go wrong, we look at every possible option, work with affected customers and since we know you better than a large company does can often help you mitigate some of the issues to reduce the overall disruption too.

Even if you move away from IT for a moment, large companies can, and do, fail - like the mass of retailers we've seen such as HMV in recent months, or even multi-national companies like Enron too. Safety in cloud computing exists through working with capable partners who help you utilise the technology, let you know what the actual risks are AND then help you address the potential risks too!

The Cloud Computing is "risky" concern

Back to Cloud Technology specifically - you might see these outages and think you should avoid cloud computing - indeed we often speak to customers who like the idea of the benefits offered, but are worried too about the potential problems. It's a reasonable thing to consider (and very wise too!). The truth is that cloud computing has risks, just as in-house or on-premise computing does. The risks are different, but they're not gone.

The reality though is less scary. As in our experience, the vast majority of our customers have very poor, and sometimes non-existant backup strategies. They have outdated and frustrating systems that don't let them work flexibly and have lots of hidden maintenance costs. Choosing good cloud technology helps you address high priority risks such as backup and improve productivity and efficiency, so moving to cloud computing often gives you huge improvement in these areas and reduces some of your risks. Companies like us then work with you to ensure you can deal with the rest.

What we commonly see is situations where a customer has an existing IT service - normally from an on-premises server, which hasn't been too bad in reliability stakes, if not productivity and flexibility. When they hear and see cloud, they feel negative about it as they've seen he headlines of outages. What they haven't considered is that every day, they're less productive because they're missing out on some functionality that would save staff time day in, day out. They're also ignoring the ticking time-bomb that is that unloved server sitting in a corner. The blunt reality is that some day it does go wrong (and we help customers every day with this type of disaster), and when it does, it's often very costly, and causes huge disruption - far worse than any cloud issue where lots of the basics are covered and resources to repair and resolve are plentiful.

Conclusion

Cloud computing is no magic bullet, outages like this weeks with Microsoft do happen. They're not every day, and they're not as bad as often hyped in the press. Cloud computing is often a good choice for small businesses - with the right implementation and risk mitigation. Sometimes customers would benefit from a hybrid approach - getting the most from cloud and on-premise systems to address risk. Before you choose any new IT infrastructure, make sure you take a fair and balanced view. Talking to companies like us here at VPW is a good choice since we offer both cloud and on-premise computing and help recommend the right choice for YOUR business. After all, we're here to make your IT easier.

Just call us on 01392 950 950 to find out how you can benefit - there's no heavy sales pitch, we promise.

mage Credit: IntelFreePress (flickr.com)
http://www.flickr.com/photos/intelfreepress/6722295999/