Lesson 6: The future of infrastructure is hybrid

In recent years, the way that companies approach infrastructure has felt a little black and white. They’re either a cloud fan or in the bare metal camp.?

The cloud hyperscalers - today identified as the big three, Amazon Web Services, Microsoft Azure, and Google Cloud - have played a big part in creating that divide. Not deliberately, but they promised scalability, efficiency, and lower infrastructure costs at a time when bare metal was still considered traditional and inflexible. Google is even doing billing by the second these days. Understandably, companies flocked to benefit from these promises.?

Unfortunately, those promises haven’t been able to quite match the reality of what it’s like to work with cloud hyperscalers. As I mentioned in my first article in this series, I’ve been a customer of all three hyperscalers - both as a highly valuable customer and as a small, limited spend customer. The difference in the experience of working with these companies as both types of customer was enormous.?

I’d like to say that as a highly valuable customer, the experience was ok. It was certainly better but I still battled with complex, almost indecipherable 1000-line Excel documents for invoices and standard support of break-fix only, with any additional guidance provided by knowledge base articles. If I wanted better support I had to pay for it and I was already spending huge amounts of money with them.?

The other option open to companies were the old school bare metal hosting suppliers. They traditionally offered long term contracts and server provisioning could take 1-2 months. But, they were more cost effective and in recent years have become far more scalable and flexible.?

Both have their benefits and their limitations. Which is why I believe smart infrastructure buying today should take a hybrid approach. I’ll explain my two key reasons for this:??

Reason 1

Every company will have peaks and troughs in the daily curve of their usage. The hyperscale cloud providers will convince you that your cloud infrastructure needs to follow those curves up and down to ‘optimise your spend’. But cloud that follows the curve is never going to be as cost effective as bare metal. In fact, it could be as much as 10x the cost of doing it in bare metal.

The problem is that bare metal still can’t quite offer the level of flexibility that the cloud can. The best and most cost-effective solution is to put in place a baseline of bare metal with cloud auto scaling. This means that scaling can still take place but if the peak stays consistent, it can be backfilled by bare metal, reducing reliance on the cloud and therefore reducing spend.

There will also be parts of the workload that don’t need to scale and therefore don’t need the scalability and associated costs of the cloud. To host these on bare metal makes much more sense.?

Reason 2

Increased resilience. If you host your entire infrastructure with one provider and that provider goes down, your infrastructure goes down with it, along with your ability to provide products and services to clients. As I mentioned previously, even the mighty AWS has been known to go down. By splitting your infrastructure across different vendors from both bare metal and cloud, you are introducing a form of built-in disaster recovery.?

For example, if your bare metal provider has an outage, you can move those workloads over to AWS for the time that the outage is taking place and vice versa. By taking a multi-vendor, hybrid approach you can massively de-risk your business.?

Have you tried a hybrid approach? How have you found it? I’d love to hear your experiences in the comments below.?