Due diligence around datacentres and cloud may increase post-Sandy
Although its services such as StatPro Revolution were not put offline by Sandy, the storm that hit North America, Neil Smyth, marketing and technology director said the difficulties that some web services have reportedly encountered may result in more due diligence by users of cloud services in future.
Sandy was a rare event, but, like others such as terrorist attacks it is something that needs to be taken into account during planning of web based services that run off servers requiring power, and in certain cases people, to keep them running.
A key issue is architecture of systems, which Smyth said was the primary way to build in better redundancy and failover responses (when services are automatically switched to run off another server if the first one goes down).
“We have a data centre in Toronto [another city battered by Sandy]. We have not seen any interruption of service for ourselves. I am aware of some sites going down, but nothing that has actually impacted StatPro. Our view on it is, well, if you have a single data centre in a single location without any architecture for failover or resilience, then that is not the best strategy. That is not something we would do in terms of delivering services to clients.”
“If you do not have a failover strategy or [have one] that has not been tested or put into plan, then you are vulnerable to things like this, especially in areas of the world subject to extreme weather. In London we do not have extreme weather, but there are things like terrorist attacks or anything that could [lead to] what we classify as a site disaster – when a site literally become unavailable in its entirety for a period of time.
“But, it is down to the architecture. We are planning an extension to Revolution called Revolution Plus, that is all hosted in the public cloud, and we are using Amazon for that. Amazon has got three North American data centres – one on the East Coast, which may have had one or two issues, but we have not noticed anything, and two on the West Coast. They have got a European one, two in Asia and one in South America. So, they have got seven sites, and you can hook them all up, be very dynamic and proactive about how your services move across facilities and sites with Amazon. So, these issues can be resolved with minimum downtime, you are talking maybe minutes while the system fails over nicely and your clients carry on. I do not see it as something that is putting me off this way of working.”
Despite the technology, it has also been apparent that the human element has thrown up some surprises. Not least among these was the decision to close NYSE and Nasdaq. Smyth said he was surprised that trading in markets was shut down completely, given the developments that have been put in place since 9/11. The trading could have continued 100% electronically, he said.
“It is strange, especially for a market like Nasdaq that is pretty tech focused. The first I heard of it they said it would be electronic trading only, then half an hour later it was a complete halt to trading for two days. That is a bit weird, and I am sure there will be a lot of questions about it.”
“These things need to be tested. It can be difficult to full scale test, but that was a strange one – why it had to completely close. But if people physically cannot get to where they need to be, then if you still have that human element of people needing to move around then these types of scenarios really do hinder that.”