In enterprise IT, there are pretty clear dividing lines between on-site and cloud infrastructure. The benefits and drawbacks of each are numerous. If I wanted to be grossly simplistic, I’d say on-site trades high capital expenses and lower utilization for lower latencies, easier compliance, and a gradually declining cost curve. Cloud infrastructure meanwhile has a fixed cost structure which can get expensive and higher latency, but you get the benefit of only paying for what you use, with nearly limitless scale.
Ideally, it would be nice to get the hands off, easy scaling, and on-demand pricing of the cloud, while still retaining all of the advantages of having your infrastructure on-site. HyperGrid thinks they’ve found a way to have your cake and eat it too.
On-site cloud may sound like a contradiction in terms, but HyperGrid spells out how it’s possible. It’s essentially a managed cloud service that sits in your data center. You provide power, cooling, and a top of rack network, HyperGrid can manage the rest.
Now, a managed service that provides equipment on-site isn’t new. What is novel is how HyperGrid charges for this. Instead of a fixed contract or lease term, they’re innovating by using an actual on-demand pricing model. It reminds me a bit of what Zadara Storage is doing with storage, except this is a complete managed infrastructure service.
Just as with public cloud providers, HyperGrid customers can specify the attributes needed for a given VM, and pay monthly for as long as that particular resource is needed. If needs change, so does the cost. Because this is a managed service, when a customer hits 75% capacity, HyperGrid will contact the customer to setup new node shipments if needed.
Normalizing Cloud Costs
HyperGrid is making the argument that this setup is much more cost effective than using public cloud providers exclusively (although they fully support hybrid cloud operation and have a VM migration offering to the cloud as well). In an interesting wrinkle, the company recently acquired a company called XOcur to prove this. XOcur developed an engine to normalize cost and performance across the various SKUs of cloud offerings to easily show how expensive a given workload would be in real dollars. HyperGrid is offering a simplified version of this at CloudMadeClear.org. This is limited to comparing VM cost across Azure and AWS, but it gets the point across.
This isn’t just a static model either. All data is pulled in real time using real instances, so the results should be pretty spot on, outside of any agreements a particular company has with AWS or Azure.
HyperGrid is announcing this pricing model today, and it’s a pretty unique offering. But this just fits into their overall managed service portfolio. These include containerizing legacy apps, allowing you to run serverless offerings with on-site hardware, and providing complete monitoring.
HyperGrid is trying to offer themselves as a feature complete public cloud competitor that lives at home. It’s an ambitious goal. This new pricing model puts them on parity with one of public cloud’s biggest advantages.