Add to Technorati Favorites
Last week I spoke about this theme at the CON.ECT conference in Vienna with the same title. I found it a pretty representative sample of the present perception and market situation. Here are a few highlights.
There is a very live interest in Cloud Computing. This event managed to convince some 40 C level and Team Managers to spend most of a day learning more and getting acquainted with this topic.
Most of the effective projects are essentially hosting of traditional applications in a virtual hosting environment, driven essentially by cost savings and a measure of on-demand infrastructure features.
Google, Amazon and Microsoft were perceived by the audience as the key players in the foggy space of Cloud Computing. My presentation of effective Cloud based core applications and implementation examples, and the technical architecture required for such, raised a lot of interest and surprised many attendees. I was in turn surprised by the almost total unawareness of the architectural/technical issues related to the deployment of Cloud applications (as opposed to Cloud Infrastructure).
Another intriguing and highly demanded topic was the legal angle. Data Protection and Privacy legislation impose quite specific responsibilities and liabilities on IT operators and providers, and the multi-tier settings of data storage in the Cloud create potential liability exposures that must be assessed and dealt with before a business implements such a solution.
My conclusion – there’s still quite a lot of fog and hype surrounding Cloud Computing, and the industry must continue to educate and communicate in order to enable the potential it holds.
I’m keen to have more opinions on this.
I used to think that multitenancy was a key incentive for Cloud Computing. The experience of the last couple of years leads me to a different view. The Cloud makes IT more ubiquitous, but multitenancy is far from being ubiquitous – so most of the services offered in the cloud are either not multitenant or use some kind of shared-hardware scheme (virtualization). So far there seems to be more cloud computing capacity than there is a demand for it – so the improved resource usage aspect is not so important. I also see that many enterprises adopt the Cloud advantages for their own software, for example rolling out core enterprise applications as RIA’s to their extended ecosystem partners and thus streamlining the supply chain and reducing costs.
My conclusion is that multitenancy is more of a future feature that would emerge when the economy would mandate it. So my recommendation is to adopt application platforms that enable multitenancy and develop future applications as such, and in the meantime take advantage of the Cloud using pragmatic approaches.
The proliferation of “everything-as-a-service” acronyms is often confusing, and merits an explanation and simplification. The VMware acquisition of SpringSource is an excellent illustration of the architecture. At the infrastructure level we find Operating System resources, which VMware encapsulates and virtualizes, offering shared hardware multitenancy but very limited elasticity. In order to increase the resource elasticity – which is the key factor of cost savings – virtualization needs to extend to the application level. That is the next layer, and I would expect that a tight integration of SpringSource with VMware would in fact provide this for Java based applications.
This evolution has a lot of similarities to Microsoft’s move with Azure. Whereas Azure offers Cloud enablement for .NET applications, VMware+SpringSource would do the same for Java applications. However, in both cases this applies rather to newly developed applications – existing applications need to be redesigned in order to take advantage of the virtualization and resource abstraction features.
As I have noted in other posts, ISV’s who want to extend their portfolio and take advantage of the growing demand for SaaS need to work across multiple deployment models, where development and maintenance costs can double if they need to create the same application in more than one format.
So the main challenge for most ISV’s is to manage an extended solution portfolio, continuing to service their current customer base with current deployment models while driving growth through Cloud Based deployment. VMware+SpringSource will facilitate this for Java oriented ISV’s, as the announcement states support for both traditional JEE deployments as well as Cloud based deployments.
An alternative to the bottom-up approach of system infrastructure vendors such as VMware or Microsoft, comes from some Application Infrastructure vendors such as SalesForce.com or Magic Software. These vendors provide for some time already PaaS and SaaS/Cloud Enabled Application Platforms (SEAP), which deal with virtualization and elasticity by abstracting system resources from the applications, so that XaaS can be achieved at conventional data centres.
Today I came across an account of a UK based ISV whose been there and done that – successfully, even in the present economy climate. Take a look at the story of FactoryMaster and how they manage take advantage of the new platforms.
Posted in Mindshare, Opinions & Comments
Tagged Application Infrastructure, Application Platforms, Cloud, FactoryMaster, ISV, JAVA, Magic Software, Microsoft, Multitenancy, PaaS, Platform as a Service, SEAP, SpringSource, virtualization, VMWare