Bringing the Cloud back down to Earth
Many factors determine the overall reliability of particular cloud services: the service provider's operations team and processes, the servers, the software, the list goes on. I still believe we are in the early days of cloud computing and we are still seeing a tremendous amount of evolution in the industry and the technologies.
This rate of change often means that particular technologies are abandoned or adapted by service providers at their will (and to consumers' benefit) over the life of a cloud contract. In fact this ability to upgrade at ease can be one of the benefits of cloud computing. Not so easily upgraded of course, especially without some form of disruption, is the underlying datacentre. As cloud services grow in particular datacentres, the cost and complexity of trying to move those services also grow.
Hence, the location and risk profile of a cloud provider's underlying datacentre should be important factors for enterprise customers in making their cloud choices. Indeed, whilst servers and in fact most technology components of a cloud solution can be judged to a large degree by SLAs or the like, the risk profile of the underlying datacentre is not so easily judged.
Thankfully, skepticism toward cloud computing seems to be going the same way as climate change scepticism and in a curious twist the potential impacts from climate change are making datacentre risk more topical and potentially more important to the overall cloud computing equation. Evidence to date of this has been in the increased storm activity and its direct impact on IT services and datacentres in particular. Hurricane Sandy for example directly impacted the hosting performance of high profile sites such as the Huffington Post, Gawker and Buzzfeed.
As a cloud advocate, it was pleasing to see that the majority of IT services impacted by Sandy were not cloud services per se,which are typically hosted in purpose-built datacentres, but rather mainly non-cloud services, often hosted in basement level server rooms on low-lying Manhattan island. The potential impacts that climate change could have on datacentre risk however are not just in regards to increased storm activity but also include factors such as the potential for rising sea levels over time and even to the availability of adequate water supplies that are typically used for datacentre air conditioning.
Of course it is not just the potential impact of climate change that needs to be taken into account when looking at datacentre risk. Other risks include political stability, natural disasters, workforce quality, reliability of power supply and even the ease of doing business in a particular country. All these factors are therefore considered in theData Centre Riskparameter,one of ten conditions mapped out in the annual Cloud Readiness Index (CRI) by the Asia Cloud Computing Association (ACCA). The CRI, with the 2013 edition due in November, provides excellent guidance on how ready Asian countries are for cloud computing. It makes for a particularly interesting readwhen the different parameters are broken down and weighed against each other.
This is how the CRI has been designed,– with the knowledge that different businesses and industries prioritise different aspects and therefore have to approachit from their own perspective. Some organizations will need to apply their own weighting to factors such as Data Centre Risk or to other factors that may be more important to them such as International Connectivity or Data Sovereignty. One aspect that is interesting to me when something like the Data Centre Risk parameter is broken down is that it becomes obvious that just because your datacentre service provider obtained the cheapest rent or best tax break for one particular location, it doesn't necessarily mean it is going to be the safest place to entrust your company's vital data.
Within this context, the upcoming release of the ACCA's Cloud Readiness Index 2013 actually produces some surprising results. Some of the smaller profile countries in the APAC cloud industry are highlighted for their safe geographic location and potential advantages for fostering cloud computing.
Likewise, it also shows that some of higher profile countries may be able to offer a lot in terms of business incentives and cloud government grants but you can't easily change the location of a continental fault line or a country's single power grid. Not that business incentives for cloud and government grants aren't appreciated - they are - but as the Index shows,each country need to consider many different factors to encourage the cloud computing industry. This is somethingend-usersneed to consider when selecting a cloud service provider and location. With cloud computing being such a rich and complex combination of processes and technology systems it is really only as strong as its weakest link. None any more important than the fundamental bricks and mortar building block of the whole stack – the physical datacentre...despite what those vendor PowerPoints may have you believe!