Up-and-coming cloud computing provider OpSource has bettered its chances for enterprise adoption by doubling its maximum virtual server size. The OpSource cloud is built atop the VMware hypervisor, and a recent switch to vSphere 4 means customers can now deploy virtual instances composed of eight cores and 64GB RAM. Large instances are critical, especially for cloud providers targeting enterprise users with complex applications. OpSource is among a small number of providers offering this much performance per instance.
Presently, Amazon Web Services is the only other cloud provider publicly advertising such large instances. Joyent offers eight-core, 32GB instances, and several others provide 16GB instances, but there’s demand for large instances. For example, RightScale noted a nearly fivefold spike in “extra-large” instance deployments among its customers, noting “larger servers typically reflect support for more users, larger databases, and additional services being made available.”
OpSource touts its enterprise-readiness thanks to variety of security and compliance features – including advanced Virtual Public Cloud and VPN options, as well as SAS 70 compliance – so the next step seemed to be offering virtual servers large enough to get enterprises porting more users and bigger applications to the cloud. Because of its strategy of partnering with — and reselling through — trusted middlemen like telcos and managed service providers, OpSource certainly could attract a healthy enterprise customer base.
However, higher-performance servers don’t necessarily equate to high-performance computing, a market that AWS currently has locked down. The big difference is high-throughput, low-latency networking and high-power processors. For example, AWS Cluster Compute Instances offers less RAM than its standard High-Memory Quadruple Extra Large Instances, but utilize a 10 GbE network and Intel Nehalem processors.
To illustrate the difference in workloads, AWS announced its GPU Instances on Monday, and already today, Elemental Technologies is offering a video-encoding service atop the new infrastructure. Previous HPC workloads running atop public clouds were limited to batch-processing and basic number-crunching, which, technically, OpSource might now be able to attract.
Related content from GigaOM Pro (sub req’d):