Is pay-per-hour the future for IaaS?

Does it strike you as fair that if you want to run a query on your big data that requires vast computing power, you have to pay for more power than you actually use?
Of course not, but that’s what is happening when you purchase Infrastructure as a Service (IaaS) on a monthly, weekly or even daily basis. Yearly contracts are plain ludicrous and are an anachronism in modern business.

We believe pay-per-hour is the most sensible way forward – here’s why. 

Big data demands big power
We’ve recently blogged about the myth of the cloud being a low-cost option for businesses, but that isn’t to say that there aren’t significant financial benefits to be had from using the cloud. This doesn’t necessarily mean cheaper, more a flexibility of payment, paying only for the processing power you actually need.

We live and work in an era of big data, terrabytes of information created each and every day that contain vauable insight about an organisation. But processing this big data requires big power, especially if your database holds billions upon billions of entries.

95% of enterprise data is unstructured, which means that older systems on the whole, cannot cope with it. Hadoop can, and is in many ways the perfect method of extracting value from big data. However, it is well-documented that Hadoop, whilst an invaluable tool in delving deeper into an organisation’s data, is an extremely power hungry application.

A flexible pricing infrastructure
The situation is further complicated by the fact that more often than not, an organisation will only need the vast processing power required by Hadoop for a short period of time. So a year-by-year contract is patently unsuitable if you want to run test queries on your data and a month-by-month investment is not much better. More flexibility is required.

That’s why we offer our bare metal infrastructure on a pay-per-hour basis, with no minimum commitment. So you can literally buy an hours’ worth if so desired, making it incredibly easy to get started on working with your big data.

Predictable cloud pricing
With flexibility of payment should come predictability of pricing. A majority of cloud providers make it virtually impossible to accurately predict pricing. The hourly rates seem low, but monthly bills always seem to end up more than you expected. This is due to the inherent difficulty in predicting I/O in advance and the poor performance of many VMs, which are in effective at processing big data compared with bare metal.

That’s why we removed the hypervisor, ensuring that you get full performance, 100% of the time. Both our infrastructure and pricing are transparent and the performance you buy is the performance you get, without exception.

Predictable pricing and flexible payment is a potent combination when it comes to IaaS – why would you choose another option?