Tech Lover 9

It’s not surprising that interest around private clouds and their implementation inside corporate firewalls is increases as many enterprisesreckon that public clouds are not secure or trustworthy enough for migrating their IT systems and/or data to. VMware, HP and others have been introducing private cloud solutions lately that are intended for enterprises that want to obtain the primary benefits of cloud computing, like auto-provisioning, self-service, cost granularity, breakdown pricing and pay-for-what-you-use, within their own premises or data centers. Some solutions, like the IBM WebSphere, also offer “bursting” of temporary resource needs into a external cloud – a type of a hybrid cloud, instead of building your own excessive infrastructure to handle peak-loads.

From an enterprise perspective, public clouds often do not address the specific needs and requirements that enterprises make. Especially is this true when it comes to issues related to data privacy, data residency, governance, compliance, security as well as other issues like trust factor. Many enterprises have requirements that simply prevent them for deploying certain processes or applications within a public cloud provider environment. For many of them, the obvious step would be to consider establishing a private cloud. However, as has been discussed in length by many thought leaders, private clouds are probably only a viable alternative for larger enterprises as well as often being controversial when it comes to deriving all the benefits of true cloud computing, e.g. as often repeated by the SalesForce CEO, Marc Benioff (Private cloud is not a real cloud). Whatever is being said, it’s a fact that public clouds do not sufficiently manage many of the security and data privacy concerns that enterprises need to comply to. For example, an enterprise will not keep its PCI compliance for resources or processes it migrates to a public cloud provider. Also, local privacy legislation and data jurisdiction regulations make it even illegal to adopt public cloud services or migrate data to the public cloud!

As addressed by Terry Wolozsin in an interesting article, it seems that certain cloud providers, e.g. Oracle, have started to provision some of their cloud solutions specifically from within individual territories or jurisdictional boundaries in order to address and fulfill the regulatory requirements that local enterprises need to comply to, thereby mitigating some of the problems associated with public clouds. Thus, with the idea to enable enterprises to easily adopt the local cloud services without worrying about some of the public cloud issues, e.g. uncertain data residence. Establishing one large data center in each continent to serve enterprises from multiple geographies and countries, a la Amazon AWS with it’s data center in Ireland for the European market, does not sufficiently address concerns of enterprises.

However, as in the case of Oracle, the cloud solution (Oracle CRM) is being being provisioned by a local hosting provider, i.e. a “semi-public cloud” solution. From Mr. Wolozsin’s article:

What this means for Oracle customers is that they can adopt Oracle’s CRM cloud solution, while addressing their data residency requirements – the data never leaves the shores of Australia.  If successful, Oracle plans on similar offerings through partners in other geographies. They are able to offer this because each customer gets their own instance of Oracle CRM on-demand, but it’s delivered to the end user as a cloud solution.

Then does this approach extract the benefits of private and public cloud computing into one cohesive pool – removing the barriers that enterprises face when it comes to the public cloud?
Enter the notion of the “quasi-private cloud”. The quasi-private cloud is a cloud that provides a middle ground between hosting your own applications in a private cloud, and using a true public cloud application/service completely managed by the provider.

Despite being an interesting concept, the quasi-private cloud has its own challenges. For example, the responsibility can be unclear between the cloud service provider and a hosting provider. Also, it seems this model misses some of the essential benefits of public cloud computing, namely the economies of scale obtained from massive data centers. It’s hard to see that smaller, local hosting providers will be able to retrieve such benefits – presumably mirroring in their price offerings.
Nevertheless, it will be interesting to see if the quasi-private cloud will find its way as the middle ground between private and public cloud computing.