Verizon Launches New Cloud Compute and Storage Service

Verizon plans to offer a new cloud Infrastructure as a Service (IaaS) that will allow enterprise customers to provision compute and storage capacity.

Already a major provider of cloud IaaS with the Terremark service it acquired over two years ago for $1.4 billion, Verizon is poised to offer a more configurable set of services with its new offering. The company laid out plans to roll its new Verizon Cloud Compute and Verizon Cloud Storage services on Thursday at the Interop trade show in New York.

The new services will allow enterprise customers to configure any combination of compute, memory, storage and network, said Verizon Terremark Chief Technology Officer John Considine during a keynote address at Interop.

"Instead of forcing users to compromise and choose the preselected settings of CPU, memory and storage, the Verizon cloud allows you to independently set these parameters and build the machine you actually want," Considine said. "You set these parameters and only pay for what you use. This means for the first time you can run your applications in a multitenant cloud and not feel the effects of multitenancy."

In private beta since last December, the services are now available for public evaluation. The company plans to make the services generally available later this quarter. The services will operate out of seven datacenters in the United States, Latin America and Europe, with plans to expand into Asia next year. Customers can configure 21 different compute environments including Windows Server, CentOS, Ubuntu and Red Hat Linux.

Kevin Clarke, Verizon Terremark's director of cloud engineering, acknowledged the company is phasing out the Terremark brand for its enterprise cloud business. "Our brand going forward is Verizon," Clarke said. "We are leveraging all the assets Verizon has to establish ourselves as a credible alternative to other cloud providers in the market."

Clarke explained how Verizon can make such a claim. "What we've actually built is a distributed switch and we've unified our networking and storage protocols as a flat layer 2 network that we're running through traffic shapers to offer performance and quality of service," he said.

While Considine said in his keynote that the storage service will support Amazon Web Services S3 storage API, he didn't describe the Verizon Cloud Compute and Verizon Cloud Storage services as open or portable in any other ways, such as support for the growing OpenStack platform.

Clarke said Verizon over time will be able to address that. Up the stack, the service's new orchestration layer is completely based on home-grown Java code, Clarke explained.

"The object model we are embodying in terms of the implementation of the code itself is divorced from the API set," Clarke said. "So we have the ability in the future, and it's on our roadmap, to offer other API flavors, including OpenStack, Amazon, CloudStack, what have you." In addition to S3, the Verizon Cloud Storage service supports the OpenStack SWIFT specification, Clarke noted.

Verizon has its own API set, which expresses what's operationally possible in the system, and the company is offering developers an SDK to implement that API set, Clarke explained. It will support other APIs next year.

From an infrastructure perspective, the new service is based on the Xen environment with support for OmniVision, which supports multiple virtual environments. Initially, customers can run Xen images or VMware images with its support for VMDKs, Clarke said. The company plans to support Microsoft's Hyper-V in the future, he noted.

For now, Verizon is emphasizing the configurability of the new service. "Control and configuration is important in the cloud," Clarke said. "That model of one size fits all is frustrating to the user community and it doesn't necessarily give them what they need. We want to offer flexibility and the ability to control that flexibility dynamically."

Posted by Jeffrey Schwartz on 10/04/2013 at 11:42 AM