Cloud Computing in the European Arena: it’s Time for Quality
Cloud computing is a business model that has delivered its promise, not only in cost and environmental efficiency, but also, and perhaps surprisingly, in terms of stimulating innovative research. After what one might a call critical-mass period, where the market demand and supply developed, and indeed in an explosive way, perhaps it is time to turn to a critical-quality period. It would be great if all stakeholders were to be involved in this new phase, but one must accept a bit of a slow-down with regard to innovation, from cloud providers wishing to leverage their investments. This is why initiatives like the European Cloud Partnership may have an extraordinary importance, by helping improve the quality of experience with cloud computing (CC) in the European arena.
Despite CC being supposedly an open market, there is little doubt that the lack of common standards is hurting further technical progress, at least in two areas that deserve attention: common low-level standard interfaces which may reduce lock-in; and non-functional standards regulating security and dependability issues. The cost of moving one’s IT between cloud providers is often prohibitive, even when the interface is as low-level as Infrastructure-as-a-Service (IaaS). On the other hand, the public-cloud market offer is completely de-regulated in regards to issues as sensitive as privacy or availability.
This being a first step to creating conditions for safe and fair contracts, there must however be technical muscle supporting the critical parts of SLAs (service level agreements), without which the former goals will be mere illusions and prey for legal traps. It seems imperative that Europe should maintain and even increase the thrust on advanced R&D for the cloud computing domain. Since the commodity cloud market is largely U.S. dominated, the offer of quality, added-value services may regain Europe the desired competitiveness.
Two research avenues come to mind as being key factors of success in this effort.
Firstly, solutions enabling low-cost resilient (secure and dependable) operation of CC systems, appeasing some of the current fears of cloud users, specially individuals and SMEs, which cannot afford expensive, proprietary offers. Moreover, this would help public administrations, which must otherwise incur high costs to safeguard the quality of data protection required by their status.
Secondly, solutions enabling the emergence of open, multi-cloud or cloud-of-clouds platforms, diluting the prospects of market monopoly and lock-in. In fact, most, if not all, of the disrupting information and communication technology advances of the past decades (Ethernet, IP, WWW, LINUX, AES, SSL, to name a few) happened through open technologies, or initially proprietary technologies that became open and, oftentimes, standardized. Even some of the technologies brought about by cloud computing, such as big data analytics and peta-scale file systems, which really got prime time in terms of industrial take-up when open versions saw the light, such as the Apache Hadoop MapReduce and File System project.
Wrapping-up, the European Cloud Partnership may have an important role in influencing the layout of policies implementing some of the measures debated above. Regulations defining categories of cloud operation are extremely important and some evolution would be welcome there. For example, rules stipulating incremental levels of resilience (privacy, confidentiality, authenticity, reliability, integrity, availability), to be enforced by different categories of CC services and platforms, might be key to clarifying what stakeholders may or not do with, and may or not expect from, a given cloud computing offer. Likewise, rules imposing different levels of restriction to data and computing location for CC services offered and/or used within the EU boundaries, might help improving trust on cloud computing and meet legislation for critical data manipulation. Interestingly enough, one hand may wash the other, for there are good reasons to believe that, as these needs exist, they will be strong motivators for innovative R&D efforts and the promotion of the industrial use of advanced technologies.
Faculdade de Ciências da Univ. Lisboa, Portugal