Don’t get me wrong. The cloud is amazing. It has made it so much easier to start new data projects, experiment with machine learning, and quickly move storage infrastructures into production. The flexibility alone has provided businesses with the much-needed resources for short-term, immediate projects (like a spike in workloads) that on-prem often just can’t provide, making the cloud an indispensable, true cost saver. But when it comes to long-term storage needs, like steady workloads and data growth, that flexibility diminishes, and the cost greatly increases. But why?
Cloud providers have highly efficient software-based technology and data centers that can deliver storage and compute more efficiently, in a more cost-effective and scalable way than traditional enterprise IT. However, for some reason, they don’t share that advantage with their customers. As Sarah Wang and Martin Casado illustrate in their recent article, “The Cost of Cloud, a Trillion Dollar Paradox” for a16z.com, cloud providers use this to funnel a sizable portion of your business’ revenue to them. And for many companies, cloud usage grows linearly with their revenue. Effectively shifting a good chunk of your revenue to the cloud providers.
But this isn’t just about saving a few (million) dollars. There’s also the cloud tax to consider. The cloud tax is based on the idea that organizations using cloud services are essentially leasing access to a cloud provider’s hardware and software. Taking this percentage or ‘tax’ from your revenue significantly reduces your margin, which directly impacts the market cap of any given company, causing massive drawbacks in the long run.
Companies like Dropbox have demonstrated how a repatriation of workloads and data from the cloud to their own data center have limited the cost of infrastructure from growing linearly with their revenue. The a16z article shows this nicely how Dropbox has improved their margins by bringing the data back to their own infrastructures.
Even though it has become a tool that nobody wants to miss, and most can’t live without, we have to be realistic about the long-term. So, the big question is: How do you bring the large amounts of data back on-prem without losing the flexibility? You need an infrastructure that is just as efficient as the cloud. You need real software storage.
We created Quobyte, the real software storage, after working for Google and seeing the benefits of a gigantic software storage infrastructure firsthand. Real software storage runs on any commodity server, that way you can benefit from cost-effective hardware instead of being ripped off by appliance vendors. It scales linearly so you can grow to 100s of PB by merely adding more servers and all it takes to start is 4 servers.
Start today by downloading our FREE edition here.