What’s happened to the cloud?
13 July 2012 | 0
One of the major selling points of cloud services, be they software (SaaS), platform (PaaS) or infrastructure as a service (IaaS), is resilience. Those giant, specialist providers can, due to economies of scale, provide a greater level of resilience than you, mere mortal, and so your application, service, web site (delete as appropriate) is better off in the cloud where it can enjoy near infinite uptime on some elastic base that can cater to its every whim.
So why are so many cloud services falling over?
Amazon, yet again, had a major outage on its service mid-June, that affected such services as Netflix, Instagram and Pinterest. The same failure point, a Virginia data centre, experienced problems again on 29 June. One tech journalist, John Koetsier of Venture Beat has said that customers are defecting as a result.
And the godfather of ‘as a service’ delivery, Salesforce.com experienced a major outage recently too. Rachel Ramsey, web editor, TechZone360, reports that “Salesforce’s cloud-based CRM systems suffered a systematic data center failure that disrupted both normal operations and backup systems, affecting nearly all 68,000 Salesforce customers in the most recent outage.”
Add to this the Office365 outages, the O2 mobile network fall over and the France Telecom experiences and, all of a sudden, cloud and major network providers don’t look nearly as able to provide the basic advantages of a cloud service.
Why is this? In most instances, what appears to have happened is that some element of infrastructure, be it a network switch such as in the famous Blackberry case, a server failure or power outage, as with Amazon, fails and causes a major outage. How can such points of failure support what are supposed to be resilient, redundant services?
It would seem that cloud as a philosophy does not have the depth to extend beyond the data centre, when a stiff breeze seems able to dispel it. All of the advertised advantages ring very hollow when what are publically explained as apparently simple failures can blow such holes in cloud services.
While many service providers appear to have designed their primary service platforms according to the best practice in cloud services, many appear to have delved no further in applying these practices and so appear unaware of weaknesses in infrastructure that can potentially threaten cloud services.
The terrible truth is that there results a Titanic Syndrome. There are millions of cloud services delivering what they promise every day around the world but as cloud services are everywhere described as “unsinkable”, in the service provision sense, it takes only one major outage in a prominent service to capture headlines with obvious results. The problem is that there have been so many outages in the last year or so since cloud became a ubiquitous term, were they to be ill-fated vessels, there’d be a pile up mid Atlantic by now!