It is often said that, central to cloud computing, is the notion of location independence. You don't need to know where the infrastructure supporting the cloud sits, and you don't care. As long as the cloud delivers computing and storage capacity to the end-user as well as access to application software and databases, all within an expectation of a high degree of reliability and redundancy, the end-user no longer has to think about the data centre, says Lex van Wyk, Managing Director of Teraco Data Environments.
Lex van Wyk, Managing Director of Teraco Data Environments.
Following the ITWeb Cloud and Virtualisation Conference, held in Johannesburg in July, it was obvious to me that too little time is spent thinking about the back-end infrastructure when it comes to cloud. Cloud offers the virtual delivery of data centre components: virtual servers and storage and virtual networking. The decision to house your cloud services operation, or actual cloud, within a data centre should involve consideration of connectivity and downtime, and the ability to match capacity with the requirement of the cloud.
In many respects, the concept of cloud and that of data centre co-location are alike in their cost saving benefits to the organisation. Accessing business applications through the Internet reduces IT expenditure and resource costs in the same manner as outsourcing your data centre requirements. The cost of designing and building a data centre is one saving, but the benefits of vendor neutrality in a data centre and having access to numerous connections out of the environment is, in most cases, not possible for most in-house data centres.
Key to cloud offerings is connectivity. Connectivity to international providers like Seacom and WACS, mobile operators; Cell C, Vodacom and MTN; terrestrial networks including Telkom and Broadband Infraco; and fibre infrastructure provider, Dark Fibre Africa will ensure cloud providers can reach their clients and that the client, in turn, has choice. In a June 2012 article, the International Working Group on Cloud Computing Resiliency (IWGCR) was quoted saying that a total of 568 hours of downtime at 13 well-known cloud services since 2007 had an economic impact of more than $71.7 million. Cloud reliability is not covered enough, and specifically not in reference to moving core applications into the cloud.
Outcomes from the ITWeb Cloud and Virtualisation Conference included the fact that cloud is now a mature technology. Reports stated that it is no longer about whether companies need cloud services, but rather about what to deploy into the cloud first. I would say that companies need to first ask who will provide the support and infrastructure behind the cloud or cloud services, and what is the connectivity available to such infrastructure to ensure limited downtime?