Opinion Enterprise IT infrastructure has consistently given us worthy investments to make and jobs to do in the last 20 years.
In the early to mid 2000’s, server virtualization made hardware fleets more efficient, improved utilisation rates, made it easier to manage applications, and saved money. It was a no-brainer, everyone adopted it happily and enthusiastically, and felt good about it afterwards.
Increasing CPU core counts helped make server virtualization and consolidation possible. What wasn’t to like about processors that increased the capacity of servers without growing hardware fleets or requiring more rack space? Again, little encouragement was needed to invest and upgrade.
The same enthusiasm greeted the late 2000s debut of affordable solid-state disks because they sped things up nicely. Sure, SSDs made network bottlenecks more apparent. But by the early 2010s networks needed an upgrade anyway in part because networked resources had become more important, be it shared storage housing those SSDs or neat stuff online like SaaS.
SaaS blew away the painful parts of operating enterprise software. Public cloud tamed infrastructure. IaaS offered access to scale like ever before, and addressed the elusive issue of elasticity to handle peak capacity without wasteful overprovisioning. By doing so it enabled whole new classes of businesses such as streaming video operators.
Public cloud was the perfect companion to mobile apps, which as smartphones became ubiquitous emerged as the digital storefront every business had to have but couldn’t afford resilient infrastructure to support when user numbers were small. The iPhone made a market. The public cloud made the market thrive.
Public cloud wasn’t for everyone, or every workload, but it was undoubtedly a welcome innovation and alternative. Few adopted it begrudgingly.
Its success saw it remixed into private clouds that delivered next-generation server virtualization by further improving server utilisation rates and taking IT shops from ‘we can get you as server next week’ to ‘here’s the self-service server provisioning portal’. Which rather impressed those who like to complain IT is not responsive.
Hyperconverged infrastructure simplified enterprise architectures and paved the way for software-defined datacenters that spurred the big networking vendors to improve their offerings.
Containers and Kubernetes changed the way apps are built, probably for the better. Terabyte-class hard disks made storage at almost any scale affordable. GPUs lit an afterburner under plenty of workloads.
All were embraced enthusiastically.
Which brings us to 2022 when most of the discussion I’ve heard about the big new thing is … hybrid multi-cloud. Which excites nobody.
The discussions this journo hears about hybrid multi-clouds are that it arrived without anyone having planned it or wanting it. Instead, it appeared thanks to a little bit of shadow IT, a whole lot of paranoia about being locked into a single cloud, and plenty of cherry-picking the best of different clouds.
Before long, organizations ended up with lots of different workloads and data scattered across diverse private and public clouds.
The results looked a lot like old school silos of applications, data, and policy. Thankfully without each tied to a proprietary Unix, but still with enough uniqueness to complicate operations.
Nobody likes a silo because they’re painful to integrate – which always happen eventually – and require discrete teams with obscure skills that don’t play nicely with others.
Silos are also risky: inevitably, one of them misses out on an important piece of security or governance and suddenly you’re asking the PR team to implement its crisis comms plan to explain a data breach.
The tools that have emerged in 2022 to tackle hybrid multi-cloud try to make their purpose a positive by saying they’ll help buyers to build a “supercloud” or a “smart cloud”.
But they’re really multi-platform plumbing and management overlays to make sure that cloudy silos can be made to behave consistently in line with an enterprise’s security and governance standards. Getting that working is remediation, not innovation.
In theory you’ll emerge from such efforts in a better state to keep on cherry-picking the best of public clouds. And also better-equipped to repatriate workloads from clouds as you, and many others, realize clouds were oversold and sometimes under-deliver.
But getting better at multi-cloud won’t be satisfying, because we should have known better than to end up with the mess we’re in.
If your 2023 is about taming a hybrid multi-cloud, good luck. I hope your efforts are appreciated.
On the upside, there’s more transformative fun on the horizon. Compute Express Link (CXL) might make the notion of composable infrastructure real and achievable. SmartNICs could improve networks and servers.
I hope you get to play with those, and tame your multi-clouds, in 2023 and beyond. And feel good about it while you’re at it. ®