Dimension Data provides further cloud details


blog You may remember that last week we were fairly hard on IT services outfit Dimension Data for issuing what we saw as a media release heavy on hype (mentioning cloud computing 71 times) but light on technical detail. Well, to the company’s credit, it has now come through with the goods, responding to all of our questions on its cloud computing offering in full.

The following responses to our cloud computing questions are from Steve Nola, DiData’s CEO of Cloud Solutions and former MD of the company’s Australian division.

What technologies does DiData’s cloud use (storage, processing, networking, virtualisation, management and so on)?

We use industry-leading technology vendors. Our Managed Cloud Platform (MCP) architecture provides choice for clients:

  • Storage: EMC and NetApp
  • Processing: Cisco UCS, Dell, HP
  • Virtualisation: VMware
  • Management: As part of the Opsource acquisition we have our own cloud management system called Dimension Data CloudControl that provides the automation, orchestration, billing, provisioning and federation. We have an open API set based on REST and it is published and available. This allows third party application providers to write to our cloud management system. Whatever can be done through the web can be done through the API.

Below are two graphics showing the CloudControl system and the MCP technology components:

Where is it located (South Africa? Australia? The moon?)?
Today, we have our public MCPs physically deployed in San Jose, California, and Ashburn, Virginia, USA; Amsterdam, the Netherlands; Sydney, Australia; with Johannesburg, South Africa and Hong Kong to be deployed by the end of March.

How much does it cost to use it?
Our public cloud pricing is available on our web site. Please refer to our Public Compute-as-a-Service pages and our pricing options can be downloaded here (PDF).

How is the multi-tenanted cloud environment set up — what sorts of customers will be using this section?
Our public cloud is by definition multi-tenanted. We also have solutions for clients that what a private cloud service (fully dedicated) or ‘Hosted Private Compute-as-a-Service’ (some level of sharing of infrastructure) and managed hosting service for certain applications. All of our cloud offerings are managed through the same Dimension Data CloudControl system. So regardless if you are using a public or private cloud (or you are ‘bursting’ from private to public in a hybrid) the cloud management system is consistent. All of our public cloud offerings globally are connected into our ‘Cloud Exchange’. This allows VMs is be initiated on any of the public cloud services around the world. We also allow private cloud services to participate in the Cloud Exchange as well, allowing hybrid services.

What is the guaranteed uptime/availability?
We offer four standard guarantees for our public compute services. These focus on availability and uptime, performance and service responsiveness. Dimension Data provides a 100 percent availability guarantee on both server and network, which means we reimburse clients for server or network unavailability encountered on the system as a result of any issues relating to the MCP. Additionally, we guarantee that there is less than one millisecond latency between servers on our cloud platform, which enables clients to design with confidence server-to-server type applications to run on our cloud service. And, Dimension Data’s cloud service has a 24/7 service desk and we guarantee to respond within 30 minutes of any critical incident being logged with our helpdesk.

Which virtual environments are supported (Windows, Linux, Solaris, etc)?
We support Windows Server 2993 Standard and Enterprise, Server 2008 Standard and Enterprise, Red Hat, CentOS and Ubuntu.

How possible is it to migrate workloads in and out of DiData’s infrastructure from and to the infrastructure of other providers?
On our public MCPs, our workloads are publicly addressable, so it’s quite straightforward to physically move a workload from one cloud server to another. Our REST-based open API set gives our clients the ability to schedule or automate shifts in workloads. However, it also requires the other cloud provider to have a published API.

What applications are supported on a ‘as-a-service’ model?
Our initial launch has focussed on our compute-as-a-service offerings, We have a roadmap for additional services which will be rolled out during 2012. In the meantime, our US team continues to support Web-based applications for clients as part of our managed hosting services.

Image credit: Dimension Data


Comments are closed.