Virtualising servers, purchasing space in data centres and utilising applications hosted and managed by third parties can have some undeniable advantages: they can increase efficiency, decrease IT-related costs, allow greater mobility and also represent a greener alternative for organisations. But as the popularity of cloud computing grows, so do concerns regarding the unclear implications of the new technologies.

If the initial worries were mostly about security of data stored at a provider, now an even bigger question is arising: what would happen if an organisation wanted their data back to bring it in-house as they grow, or to transfer it to another provider as part of a merger, or a cheaper and more efficient provider for some of the services (e.g. only email or back-up)? Although it is possible to retrieve and migrate data, it is not an easy and straightforward operation and the costs involved might represent a barrier, causing the organisation to be locked-in with the provider – and accept any price and conditions they might decide to impose.

The problem with the newness of cloud computing technologies is that there are yet to be set standards for data formats and APIs to allow interoperability between infrastructures. Cloud computing providers are already working on how to improve portability and reduce latency during data transfers, but only within services and platforms hosted on their own, proprietary infrastructure.

Migration to another vendor can instead be a lengthy and expensive procedure – apart from possible end-of-contract penalties, organisations will be charged both for format conversion and for the transfer, including additional charges for bandwidth usage which due to the high latency, might altogether amount to a very large figure. Migration costs can be prohibitive when dealing with a large amount of data, therefore even if it might seem easier and convenient to have only one vendor providing all services, storing the entire organisation’s data within one infrastructure represents a threat which might obstruct growth, structural changes and the search for more cost-efficient and bespoke solutions.

Experts reckon it might be a few years before data and service portability within vendors will be possible, but organisations need not put off a move to the cloud – they just have to apply some smart thinking. The key to avoiding lock-in, it seems, is to not have all the eggs in one basket. The wisest organisations are already using this technique, which sees them cherry-picking various vendors for different services: one provider for email, another for back-up and another couple for applications and VDI. There are a few criteria for choosing, not necessarily based on the cheapest offers: ideal vendors have to first of all provide modular packages, use popular formats for data and services and be transparent on regulations and fees applied to data transfer.

Many benefits can be achieved with this strategy: for instance, organisations can create a bespoke and flexible solution, and choose the best offer for each service. In some cases the overall price could be higher than the cost of a single provider for all services, but if it is lengthy and economically prohibitive to switch vendor, then price is inelastic and can be increased at any time, leaving the organisation no choice but to pay. It is also essential to take into account the risk of a provider going bust: the recent security attacks on Google and the dotcom meltdown have taught us that no company is too big to go out of business.

To avoid data and financial loss, the only solution is to use more than one vendor. It is only through a game of pick and choose that lock-in risks and their consequences can be avoided, while still enjoying the cost-efficiencies made possible by cloud computing.