Who Has Your Data? In The Cloud, It’s Not You…

By Guy Bunker | July 5, 2010 | 1 Comment 

Guy Bunker

The news last week was that EMC was closing its Atmos cloud storage service with immediate effect – you can keep using it for developmental purposes but that’s about all.

Why did it close? Industry analysts said that it never took off, and recent surveys show that it is still a way off becoming mainstream.

So… what if you have data in a cloud service provider and it decides to shut down the service? EMC isn’t the first to do this, and it won’t be the last – you do need a contingency plan. In this case, there is a grace period where you can get your data out – but, as a simple task, workout how much data you have and how much bandwidth and figure out how long it would take to get the information out. This is a simplistic view as everyone else will also be trying to get their data, so bandwidth is going to be under severe strain (the equivalent of a run on the banks…) Do you have enough local storage to hold it all and if you have data being processed by an application, then will you be able to get your hands on the application as well?

Let’s assume you did manage to get your data out, then how long will it take to get it reloaded onto another service provider’s cloud and get the application back up and running?

Business Continuity / Disaster Recovery needs to take into account outsourced (out-tasked) IT services and have contingency plans for service outages and shutdowns – planning should start now… after all there’s no time like the present.

Guy Bunker

When More Is Too Much

By Guy Bunker | September 8, 2008 | Leave a Comment 

Guy Bunker

The London Stock Exchange suffered an outage today and while the details are a little sketchy, it seems that excessive trading brought the system down, or at least some part of the system – and here lies the problem. Systems are increasing in complexity and the interrelationships are becoming less and less understood. The result – an outage in a minor application can bring the whole system to a standstill.

Companies cannot afford to overlook the infrastructure on which their IT runs. If the IT becomes more complex, for example using virtualization or Service Oriented Architecture or third party service providers then these have to be included in Disaster Recovery / Business Continuity plans. DR/BC plans are often not tested frequently enough – with one of the reasons being given that ‘it will effect the customer’… well, at least in a test you get to choose the date and time. Friday’s at 8pm is a good time to start, or Saturday at 10am after the end-of-week tasks have completed… that way, when it doesn’t go to plan, you have another 24 hours to sort it out.

So… once more, revisit those DR/BC plans and schedule a test and make sure it includes all the 3rd parties as well. How well would your infrastructure cope if you had a 100% increase in demand… twice the business or none at all?