Microsoft and Oracle have signed a partnership which enables the supported use of Oracle on Hyper-V and Azure and relaxes the licensing rules on those platforms. Oracle states that it will provide certification and support for Oracle applications, middleware, database, Java and Oracle Linux on Windows Server Hyper-V and Windows Azure. As of today, customers can deploy Oracle software on Microsoft private clouds and Windows Azure, as well as Oracle private and public clouds and other supported cloud environments.For the purposes of licensing Oracle programs in an Authorized Cloud Environment, customers are required to count each virtual core as equivalent to a physical core. This policy applies to all Oracle programs available on a processor metric.
When licensing Oracle programs with Standard Edition One or Standard Edition in the product name, the pricing is based on the size of the instance. Authorized Cloud Environment instances with 4 or fewer virtual cores are counted as 1 socket, which is considered equivalent to a processor license. For Authorized Cloud Environment instances with more than 4 virtual cores, every 4 virtual cores used (rounded up to the closest multiple of 4) equate to a licensing requirement of 1 socket.
Under this cloud computing policy, Oracle Database Standard Edition may only be licensed on Authorized Cloud Environment instances up to 16 virtual cores. Under this cloud computing policy, Oracle Standard Edition One may only be licensed on Authorized Cloud Environment instances up to 8 virtual cores.
Standard named user plus licensing applies, including counting the minimums where applicable. For Oracle Linux purposes, each Authorized Cloud Environment instance is counted as a “System”. “Basic Limited” and “Premier Limited” support is not available for Authorized Cloud Environment instances greater than 8 virtual cores.
This policy applies to cloud computing environments from the following vendors: Amazon Web Services – Amazon Elastic Compute Cloud (EC2) and Amazon Simple Storage Service (S3), Microsoft Windows Azure Platform (collectively, the ‘Authorized Cloud Environments’). Unfortunately VMware is missing so far. Let’s hope this will change. Perhaps VMware will announce a similar partnership with Oracle at VMworld in San Francisco end of august.
Relaxing the licensing model applied to virtual infrastructures would mean a great push towards virtualizing business critical applications build on Oracle databases. It will save customers a lot of money if they only have to license their virtual cores instead of all physical cores in the cluster. But if this deal will not apply to VMware infrastructures it will be quite a slap in the face and might compel customers to abandon vSphere for Hyper-V, even though Hyper-V isn’t yet as mature as vSphere.
After attending the Nutanix Star Trek into Darkness event (thanks again!) and the brilliant animation explaining how Nutanix operates at a high level (http://youtu.be/nSqwAxhFpA8) I kept thinking about the use for building a cloud based on Nutanix technology compared to a more classical approach based on blades. It seems on the one hand that a Nutanix based cloud is a lot simpler to design and implement and looks to be very easy to scale. Need more resources? Add another node.
Of course I still have a lot of questions, like does the distributed storage (presented as NFS) perform better inside the chassis compared to across chassis, or will it perform the same? How many replicas of data blocks are stored across the nodes to ensure high availability? I will have to have discuss this with Raymon Epping, the Nutanix System Engineer in the Netherlands.
But based on the information available (and disregarding pricing for the moment), it looks like Nutanix offers highly scalable data center building blocks, not limited to private, hybrid or public clouds. Their converged infrastructure approach, combining computation/processing and storage in single nodes with a Google-like distributed file system, that in turn are combined in chassis with up to 4 nodes with 10 GbE uplinks is very scalable. After the initial four node chassis, you can scale out per node. So you start small, with four vSphere hosts and add another node to the cluster when necessary. The only downside in my humble opinion is that when you run out of storage, you must add another node consisting of storage and compute resources. Then again, is that a bad thing? Depends on the price I’d say.