A cloud over the cloud

Cloud computing is revolutionary, but not yet fully reliable

In the last month alone, online services from Apple, Google and Blackberry all suffered outages. These didn't last too long. Google services were down for less than three hours and Blackberry too, was out of commission for under a day. But the disruptions have, so to speak, cast a cloud over cloud-based services. The cloud is useful and increasingly popular but not fail-safe. A large number of businesses and quite a few government departments around the world, including emergency services and defence forces, rely on clouds. Loss of connectivity and data access for even a few hours is crippling.

Business processes have been transformed as the cloud ecosystem has gained traction. Cloud services started with webmail. But now the service ranges from data storage and sharing, to web-based apps, software suites and tools to manage enterprise-wide needs. The cloud model offers a combination of convenience, efficiency and cost-savings. Ubiquitous broadband access and cheap server space allow users to store and transfer files and data seamlessly across multiple locations and devices. They also allow the delivery of software as a service. Instead of installing expensive office suites on hard drives, users can access online software to collaboratively edit and share documents, presentations and data.

New technologies are disruptive and this is no exception. It has forced companies like Microsoft and Oracle, which derive a large chunk of revenue from enabling enterprise productivity, to review their business models. Microsoft for example, has launched the online Office 365 to compete with cloud-based office suites like Google Docs, Zoho and Polaris Office. Even the Indian IT services model of large teams implementing customisation for enterprises has been impacted. Public clouds offer an array of free, open-source, easily customised tools. Even enterprises, which prefer "private clouds" for reasons of data security, can adapt these and reduce the need for manpower-intensive customisation. The business case for moving to the cloud is therefore, quite compelling. But at some stage, most major cloud-based service providers have also seen outages. The cloud-service pioneer, Amazon.com has gone down multiple times. Apple's iCloud has gone down. Microsoft's Azure and Office365 have gone down. So have Dropbox and Google. There have also been issues about data security, with whitehat hackers demonstrating that it is relatively easy to gain access to cloud-based data. Nor can the system administrators of an enterprise diagnose and troubleshoot outages, or plug security gaps on a public cloud with total reliability. Government departments providing mission-critical emergency services cannot risk moving completely to a cloud-based model yet without redesigning architecture to ensure 24x7 availability. Nor can transnational corporations afford to do this. Some TNCs have compromised by adopting cloud-based models as a complement to their old-style data centres and centralised servers. But this is not an efficient solution.

Creating redundancies, and building a high-availability disaster recovery architecture that works with a cloud-based model are entirely possible. For example, it is possible to create multiple inter-connected but separate cloud availability zones. But such architectures require creative thinking and new methods. They also impact cost structures and processes. Clouds will undoubtedly become more reliable. However, as with old-style networks, they will inevitably go down every so often. Users will have to cope with the challenges of these disruptions to extract the maximum utility from this new disruptive technology.