Oracle Offers Subscription Pricing For On-Premises IaaSOracle Offers Subscription Pricing For On-Premises IaaS

7 Dumb Cloud Computing Myths

Oracle has repackaged a set of familiar products into what it calls a new infrastructure-as-a-service offering. Unlike Amazon or Rackspace IaaS, the Oracle products, announced Jan. 15, get installed on the customer's premises and are operated by the customer.

But Oracle will charge for the package by monthly subscription, instead of the usual lifetime license. Oracle will also take responsibility for integrating and configuring the package's elements, making them easier to deploy on a customer's premises. And it will shift what had been a customer's capital expense into a monthly operating expense.

The products involved are the already familiar engineered appliances Exadata Database Machine, Exalogic Elastic Cloud, Exalytics In-Memory Machine, Sun ZFS Storage Appliance and the SPARC SuperCluster, an appliance running Solaris 11 on up to 16 SPARC processors with eight cores each. Oracle in effect has given the "infrastructure" name, often applied elsewhere to plain vanilla x86 servers and storage, to a set of its highly engineered appliances.

Oracle IaaS will also offer the option of an elastic compute capacity as a service, called Capacity on Demand. That's another way Oracle IaaS differs from its standard packaged software product. It includes the option of elastic CPU scaling, or compute capacity on demand, which regular, sold-by-the-CPU packages wouldn't allow. If a customer decides to invoke compute capacity on demand, there is an extra service charge for that month. For retailers, that would be a valuable option in the months of November and December, during peak shopping periods. But there is no charge in months when it is not invoked.

[ Want to learn more about how Exalytics provides in-memory business intelligence? See Oracle Exalytics: Is It A Must-Have for BI? ]

When providing the appliances as IaaS, Oracle configures them so that it has about 25% of the CPU capacity in reserve. If the customer invokes compute capacity on demand, that reserve becomes available to handle an increased load, said Juan Loaiza, Oracle senior VP of systems technology, in an interview. So a system normally running at 75% of its capacity has the equivalent of one-third of usual operating power available in reserve.

"The idea is to bring the benefits of the cloud to on-premises infrastructure," Loaiza said.

Oracle customers want to do that, for a variety of reasons, rather than place their workloads in a public cloud infrastructure, he added. "Customers want control over their system. They want to keep their data on premises. They want to keep their encryption keys inside the firewall," he noted.

The IaaS package includes 24-by-7 premier Oracle support, and the systems themselves have been configured and are operated for strong reliability, security and performance characteristics, said Loaiza.

By operating Oracle appliances as an "infrastructure" set, customers don't have to buy upfront the extra capacity they normally include with a license purchase. For customers who don't use the extra compute-capacity-on-demand feature, the monthly subscription price will be lower than the regular purchase price for up to five years. If they do use the elastic compute capacity, it will be priced lower than the purchase price for four years, or longer, Loaiza claimed. "It depends on how much elastic compute capacity you use," he noted.

"When something unexpected happens, the customer can rent compute capacity on the fly" and pay for it as needed, instead of having to overprovision his equipment from the date of purchase, he said.

Loaiza said Oracle has built out its cloud customer base to 10,000 customers, a claim repeated in Forbes online on Jan. 18.