Hardware 1.0 vs. Hardware 2.0

Hardware 1.0 vs. Hardware 2.0

I’ve seen a lot negative reactions to the terms multi cloud and hybrid cloud. I would argue, that’s because we are framing it wrong. It’s really Hardware 1.0 and Hardware 2.0.

In Hardware 1.0 Dell, IBM and HP all tried to sell us fairly similar servers. We typically bought these servers, but we could finance them or lease them. I loved the DL380, they were bullet proof. We used to call them the Boeing 737 of servers (until the roofs started tearing off ruining the analogy). The Dell R710s were everywhere. The IBM Blade Centers were all the rage.

Each of these vendors always tried to add value to their x86 servers, in the way of extra software. Sometimes it was drivers and management add-ons, but they also tried to “climb” the stack with databases, back-up software, Java and eventually even full-on line of business software (HR, CRM, Supply chain, etc).

Cloud is really no different. The similarities are striking. The business models are strikingly similar. Instead of buying servers, we now rent them by the hour or minute. So, hardware 1.0 was buy, hardware 2.0 is rent. And, the same progression is happening (faster this time). The cloud providers are climbing the stack into business software to try and remain sticky. That’s fine and dandy.

But, there’s a critical difference between hardware 1.0 and hardware 2.0. In hardware 1.0 you could move from IBM to Dell, or from HP to IBM because the Intel instruction set protected you. Each of those vendors (and many others) also offered CPUs from AMD. You could keep your Oracle database no matter which hardware 1.0 vendor you moved to. Heck, you could keep your IBM DB2 database and change the hardware to Dell.

With hardware 2.0, you can’t swap pieces in the stack, it’s all or nothing. If you leverage Amazon SQS, RDS (could be changing slightly with on premise) or DynamoDB, you can’t swap out the underlying hardware. You can’t even swap out the software. There’s no upstream, open source community driven project behind it to protect you. There’s nothing. You will never be able to run SQS, RDS or DynamoDB on Azure or Google infrastructure. Period.

You might say to yourself, “well it’s fine to align your business strategically to one vendor to move quickly and leverage innovation.” For an experiment? Sure. For long term infrastructure? No. Here’s why.

It’s way more than access to innovation. It’s disaster recovery (DR), it’s pricing, it’s strange edge case bugs that can’t be fixed, it’s actually innovation lock-in. What happens when Amazon has a huge, cascading outage? What happens when Google Compute has a faster, cheaper database? What happens when you introduce a new feature in your code base and it hits a bug with the cloud provider’s service – but, it’s a use case they don’t view as strategic? You probably don’t spend enough to be strategic to them either.

You think Service Level Agreements (SLAs) will protect you with refunds? Let me tell you the dirty, little service provider secret – refunds protect service providers, not you. SLAs get you to sign the contract, make you feel good and protect the service providers legally – this trick is as old as the telco providers. Sure, they will give you your $50,000 back, but you lost $3M during the outage.

Sounds like a great deal, right? No, if you are making money, you need DR. If you are a startup, “strategically align” away and get your access to innovation – you’re not making any money anyway.

Here’s another secret. If you think you need a cloud provider to get easy access to innovative technology, you actually have a people problem. There is a ton of innovative, open source technology available in easy to consume container images.

So, no hybrid cloud, and multi cloud are not an end. They are a means to an end. The ends are age old, they’re hunting new top line revenue, and preventing too much bottom line spend. They’re giving your innovative software engineers access to innovative software so they can add value. They’re giving your operations team leverage when negotiating bottom line spend (cloud contracts).

If you are making money, have innovative employees, and leveraging Hybrid/Multi Cloud + Linux + Kubernetes + Open Service Broker, you will have easy access to your own ecosystem of cloud neutral software. You will achieve your ends.

2 comments on “Hardware 1.0 vs. Hardware 2.0

Leave a Reply

Your email address will not be published. Required fields are marked *