This page contains a Flash digital edition of a book.
HPC 2013-14 | Te cloud


Cause for concern? Te industry is certainly becoming acclimatised to the idea of cloud, yet there is still a level of hesitancy. One of the first barriers, touched upon earlier in this article, is one of cost. Andrew Carr, CEO at Bull UK and Ireland, has observed that cloud computing has been pitched as a more cost-effective solution, and, he says, in the consumer world it has been, but in the HPC industry this is not always the case. Tis is partly due to individual demand as the cost of renting generally does not compare to the cost of ownership when it comes to the sustained and consistent use of resources. And there are other, more pressing factors to contend with. ‘True cloud is about using any available capacity from a network of data centres, but because of the sensitive data, simulations, outputs, and algorithms people are running on HPC they want the reassurance of knowing where their data is at any given time,’ Carr explained. He believes that, while the cloud model does


enable organisations to extend their processing power and have access to the type of compute that historically only those with the deepest pockets were able to take advantage of, an on- demand model rather than cloud-based services is where the market should be heading. Again, the distinction here can be somewhat hazy. ‘As a user, I should be able to store all my


data in the cloud securely and not have to worry where it is stored. As long as there is a service level agreement wrapped around it that guarantees the availability of the data, as well as back-ups and restore if necessary, to a certain extent it doesn’t matter where the data sits as I will simply be buying a service like any other,’ he said. ‘Tat’s where the cloud is very powerful – but also where it’s something of a concern depending on the nature of that data. An on-demand model comes with the requirement and delivery of a small amount of customisation prior to usage. Put simply, on-


Selecting a service


Infrastructure-as-a-service (IaaS) Infrastructure as a service (IaaS) is a standardised, highly automated offering, where compute resources, complemented by storage and networking capabilities are owned and hosted by a service provider and offered to customers on-demand. Customers are able to self-provision this infrastructure, using a Web-based graphical user interface that serves as


16


an IT operations management console for the overall environment. Application programming interface (API) access to the infrastructure may also be offered as an option.


Software-as-a-service (SaaS) Software-as-a-service (SaaS) is software that is owned, delivered and managed remotely by one or more providers. The provider delivers


software based on one set of common code and data definitions that is consumed in a one-to-many model by all contracted customers at any time on a pay-for-use basis or as a subscription based on use metrics.


Platform-as-a-service (PaaS) A platform-as-a-service (PaaS) offering, usually depicted in all-cloud diagrams between the SaaS layer above it


and the IaaS layer below, is a broad collection of application infrastructure (middleware) services (including application platform, integration, business process management and database services). However, the hype surrounding the PaaS concept is focused mainly on application PaaS (aPaaS) as the representative of the whole category.


As defined by Gartner Research


‘Depending on the workload profile, the cloud still has a lot of latency variations and as such is currently far less predictable than running applications within a data centre,’ he explained, adding that smaller HPC organisations that may not have the financial means or human resources to run HPC clusters in their premises, however, will be very drawn to a cloud approach. ‘Having the flexibility of resources allocation


without the need to incur capital equipment costs does have a lot of advantages for the scientific community.’ One further, and increasingly critical,


“While cloud computing does provide an attractive infrastructure, it’s going to take quite a long time before the large scientific institutions and HPC centres move their workloads in this direction”


demand is a more engaging process between the customer and supplier.’ Carr went on to say that he does not foresee


a time when HPC capabilities will be 100 per cent on-demand or cloud-based: ‘Tis isn’t just due to concerns about security, or how people provision the capability. Rather, it is the need to convince independent soſtware vendors to adopt a cloud-based licensing model that is on- demand driven than perpetual. Tat’s currently one of the main inhibitors to take up.’ Jean-Luc Chatelain, chief technology officer


and executive VP of Strategy and Technology for DataDirect Networks, agrees that while cloud computing does provide an attractive infrastructure, it’s going to take quite a long time before the large scientific institutions and HPC centres move their workloads in this direction.


advantage in cloud’s favour is its ability to facilitate collaborations and the sharing of research between organisations and institutions. Te emergence of academic clouds, for example, fits in with Chatelain’s prediction that the HPC community will begin to adopt professional, multi-tenant clouds that are configured to meet the needs of, and foster relationships in, specific industries. While collaborative clouds are on the rise, Chatelain admits that we are unlikely to ever see national laboratories taking this approach given the nature of their work and how it relates to national security as, despite the fact that cloud providers have put a lot of resources into the security of their offerings, some organisations will always find it difficult to overcome security concerns. Nevertheless, collaborative clouds are


being embraced – albeit at a steady pace – and this approach has led to the development of a new type of storage. According to Chatelain, object storage, which ensures each piece of data has its metadata – thus making it ideally suited towards collaborations – will become the underlying technology for storage within cloud environments in the future.


The crystal ball So what will the cloud computing landscape look like over the coming years? Dave


Skymax/Shutterstock.com


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36