Industry analysts have cast doubts on the ability of cloud service providers to meet and maintain the IT industry's much-vaunted service availability requirement. This is the so-called 'five nines' (99.999%) of uptime, long seen as the 'Holy Grail' of uptime targets by all end-users of hosted services.
Is there any reason for the service delivery of an application residing on a server in a data centre to be rated higher (or lower) than a cloud-based application? Does this mean 100% uptime is the benchmark for the new generation of cloud service and application providers too? And is it achievable?
Firstly, it's important to note that no service provider can guarantee 100% uptime. There are many opportunities for 'gremlins' to invade and for something to go wrong. Just look at the cellular telephone service, which regularly drops calls and often delivers questionable voice quality, to realise the harshness of the challenge represented by the maintenance of 99.999% uptime.
Between the lines
Although many service providers are at pains to point out that their 99.999% uptime assurance caters for a wide range of 'force majeure' eventualities, such guarantees need to be examined and the fine print analysed carefully.
Trust is one of the most important aspects of a successful service provider-client relationship.
Martin May is regional director of Enterasys Networks.
Before glibly making a 99.999% uptime assurance - allegedly achieved globally by only around 10% of the organisations that claim the figure - cloud service providers need to establish a business model that explains how their service level agreements (SLAs) are structured.
While a 99.999% uptime SLA is not impossible to achieve, it must come at a premium price if the service provider is truthful. After all, 99.999% uptime translates into just 0.4 minutes (24 seconds) of downtime for every month of operation, or only five minutes per year.
Such a guarantee has to involve multi-level redundancy, including physical data centre diversity. Importantly, redundancy will be required all the way up the delivery chain to include power supplies, networks, distribution philosophies and storage configurations. Fulfilling these requirements is not a cheap exercise or the work of a moment.
I have long been an advocate of cloud service provider transparency, highlighting its importance in terms of security for all new cloud technology adopters.
Candid approach
While cloud service providers need to provide a transparent means for users to monitor the flow of corporate data and its storage locations at all times from a security perspective, it is also important for them to receive an honest appraisal of the service provider's uptime records.
Ensuring a vendor has a strong uptime history, clearly understood SLAs and a solid data protection narrative are all key to a successful cloud deployment.
In this light, trust is one of the most important aspects of a successful service provider-client relationship, and the process of setting and meeting expectations around service delivery must be seen from this point of view.
Users might well be more comfortable if the cloud service provider would offer an honest trade-off between lower - more realistic - uptime guarantees and lower costs.
Maybe a 99.99% uptime guarantee, representing four minutes of downtime per month (43 minutes per year), or a 99.9% guarantee (43 minutes per month, or eight hours and 46 minutes per year) might be attractive if appropriately priced, policed, documented and accurately - and honestly - reported.
The fact is, many SLAs are written in such a way as to exclude downtime (replacing it with 'maintenance') and are peppered with disclaimers, prerequisites and conditions so that in the event of unexpected downtime, the service provider doesn't have to pay penalty fees.
When reports in the press appeared in which the research group Gartner criticised Amazon's 99.5% SLA guarantee, spokesperson Kay Kinton responded, saying Amazon [unlike some of its competitors] does not exclude actual downtime from its calculations.
“Our Amazon Web Services health dashboard gives customers complete, constant access to how the services are performing, she stressed. “What matters most is demonstrated performance, and ours has been strong.”
Should users insist on 99.999% uptime from their service provider? According to Andrew Hiles, president of consulting company Kingswell International, it's a futile exercise. “There are more dangerous threats we can address cheaper,” he says. “A slow or poorly designed Web site or an insensitively thought-out interactive voice response system may have far more - and longer-lasting - effects on customer loyalty and market share than 10 minutes of downtime a year.”
Follow Martin May on:
Facebook: http://www.facebook.com/martin.may.enterasys
Youtube: http://www.youtube.com/user/mrmartinmay
Twitter:@MartinMay001
Linkedin:http://za.linkedin.com/pub/martin-may/19/609/415
Share