When deciding between the implementation of either hyperscale clouds or enterprise Private Clouds, it is critical to think about all of the key variables that come with each option: availability, cost, flexibility and support. By keeping these variables in mind when choosing an IaaS solution, organizations can be sure they get the right cloud for their needs. Today, we’ll be focusing on one very important aspect – cost.

Hyperscale cloud platforms are public, multi-tenant infrastructures deployed at great scale and magnitude. Since hyperscale clouds are sold by the instance in fixed sizes, shared, and non SLA-backed, their providers may appear as the cheaper solution in the short term. There is a downside though: customization and data movement can result in sky-high prices in the long-run. That’s a surprise that no one likes.

The hyperscale cloud deployment model can also easily aggravate data sprawl-related concerns. As a customer’s data environment grows, it becomes increasingly more cost-prohibitive for that customer to remove its data from that cloud.  That’s because hyperscale cloud providers typically charge their customers to move data. Take for example a fictitious 10 PB customer.  Based on the AWS website direct connect pricing of $.03 per GB, a 10 Petabyte customer will pay $300,000 just in Amazon charges for out-going data. This doesn’t include Amazon’s usage charges or the need for additional 10 Gb ports to move the massive workloads.  In reality the costs to move would be far greater. So, while hyperscale computing might at first seem a low cost option, in the long run hidden fees and big data have the ability to jack up cost and surprise the customer.

In contrast, Faction offers a very flexible data migration path. A customer can migrate ‘over the wire,’ similar to AWS, but at a lower and often-times free direct connect price. Or, customers can bring their own device and co-locate it in the Faction Cloud for a temporary timeframe at a nominal price. In this example they would also use 2 x 10Gb connections to move all the data to/from their own asset. The 10PB migration could be accomplished for an order of magnitude less than the hyperscaler’s price.

Faction’s flexibility doesn’t stop there. Its consumption model is based on resource pools.  Faction sells cloud by the Bloc. Each Faction Cloud ‘Bloc’ consists of a pool of CPU, RAM, IOPs, network, firewall and a 100% SLA. The Blocs are fine-tunable in any combination of Unit increments and further customizable to become privatized, protected, or reserved. At first glance, an enterprise-class cloud can appear more expensive than a per-instance solution. But using and expanding an enterprise-class cloud comes with zero or very few fees related to scaling and data movement, making costs much more predictable. In other words, what you see is what you get.

No one deployment works well for everyone. For firms in need of a low-cost, non-SLA backed option, a per-instance, hyperscale option may be your best bet.  If your organization prioritizes availability, customization, and flexibility, enterprise-grade cloud is the way to go. By looking beyond just the consumption model and also inspecting an offering’s availability, scalability, cost, flexibility and level of staff support, organizations can be certain they select the best IaaS deployment model for their unique needs.

Don’t forget to check out our Cloud Consumption considerations, Availability Should Never be an Option and We’ll Bend over Backwards for You!