To better understand, in this article, by Cloud we will mean the Iaas (infrastructures as a service) offers offered by the key players in the sector such as Amazon, Microsoft-Azure and Google. Then, by Saas we will mean a private cloud offer supported by a business operator built on its own IT. Finally, by on-premise mode the classic acquisition of software licenses deployed on the end customer’s IT and operated by its own teams.
When launching your archiving project, your choice must be guided by the 3 factors described below:
First factor: the volume of data to be archived
If your unit is around a hundred terabytes per year, you should forget about cloud or SaaS mode and opt for on-premise mode. No cloud provider will be able to offer you a competitive unit cost between 0.0006 cts and 0.0012 cts of euros. Only a solution mounted on a dedicated internal infrastructure (in-house IT) as the major billers in the transport sectors do. energy and telcos, will allow you to achieve these unit cost objectives.
Second factor: the duration of data retention
How long does it last? Short term or long term? Or even without a retention limit?
When you plan your digital preservation project for such durations, in an IT world where all the technical components are changed every 3 years, the archiving software solution chosen must be independent of any technological adhesion: OS, hardware, network, storage…but also the cloud. The first experiences of starting an archiving project in the cloud show a strong adhesion with the chosen cloud solution. The trap set by cloud providers is the low entry cost and usage-based billing. At 3 years, estimating the costs of Cloud infrastructure is a matter of forecasting, what about the costs at 30 or 50 years or even 100 years? This is prediction, therefore a real risk of financial slippage in the long term. Digital Preservation is not part of the same time scale.
Third factor: reversibility
Despite the angelic transparency and interoperability displayed by these Cloud operators, exiting their offer will surely be more complicated than entering it! As in the heyday of large ECM solutions, very organized in terms of customer lock-in. To date, Cloud offers offer no guarantee of reversibility without re-development, adaptations or even re-engineering, necessarily inducing dissuasive exit costs.
Cloud, SaaS or On-premise: the best approach for starting your project is great caution and a precise analysis of all costs, including them over the long term. Mindless dives and evaluations into the cloud will inevitably lead to abysmal future billings. To go further, Sarah Wang and James Casido perfectly summarize in their article “cost of cloud: a trillion dollar paradox” the paradox of starting in the cloud: “Cloud: “you’re crazy if you don’t start in the cloud, you’re crazy if you stay on it.
Gaël Bertin