After three years of largely virtual events, Las Vegas was back in full force for at least two major technology conventions this month. Cascadeo was an invited speaker at re:Invent 2022, by far the biggest cloud trade show on Earth. The following week, we attended Gartner’s Infrastructure, Operations, and Cloud Strategy (IOCS) at the same venue. The contrasts between the two conferences were remarkable, but the common theme in both cases was clear: Let’s get back to business.
Any remaining COVID-era restrictions were cast aside, with tens of thousands of people from around the world mixing indoors for days at a time at the conferences and vendor parties after-hours. History will determine if this return to near-normal was premature, but the audiences during both weeks were confident and eager to return to pre-pandemic life. It was amazing and energizing to see the world putting itself back together, and companies getting serious about resuming normal operations after an extended period of disruption.
While re:Invent focused exclusively on cloud and its emerging concepts, the Gartner presentations provided a reminder that many – perhaps most — companies and governmental agencies in the North American market have not yet meaningfully adopted cloud, and some have no real intention of doing so. Working exclusively in the cloud industry, it is easy to forget that roughly 90% of IT workloads have yet to move to public cloud—and that many companies and workloads never will. From inside the industry, public cloud feels fully mature and the obvious answer to nearly any IT problem. Despite the irrefutable benefits cloud offers around security, scalability, physical infrastructure operational challenges, and CapEx-vs-OpEx, a lot of companies still have not found their way to the public cloud for a variety of reasons, many of them legitimate. Legacy equipment vendors and technologies are not going away anytime soon.
At times, the contrast reminded me of Toyota’s steadfast insistence that battery electric vehicles (BEVs) will not replace internal combustion engine vehicles despite the overwhelming advantages that BEVs offer in terms of reliability, operating cost, tailpipe emissions, diverse upstream energy sources, mechanical simplicity, and overall energy efficiency. Toyota is not wrong, and neither are the legacy industries; ICEs will be with us for a long time, but like dial-up internet and the horse-and-carriage era of ground transport, the market for these technologies will peak and then endlessly decline. Just as it will take humanity decades to wean itself off of fossil fuels in a best case scenario, legacy non-cloud IT will be with us for a very long time. It simply is not possible to replace the fleet of deployed vehicles or datacenter infrastructure overnight, even if everyone agreed that this was the right path – clearly not the case, and provably untrue in many corner cases.
Cascadeo remains firmly convinced that cloud wins in the end, and that companies who are reluctant to embrace it face significant and mounting limitations from a business agility standpoint. With that said, I personally still own and operate ICE vehicles on occasion because there are real scenarios like long-distance towing where BEVs are not yet up to the task. Similarly, I think a legitimate case can be made that public cloud is not the right answer to every conceivable problem or use case. 37Signals, the makers of the popular Basecamp app, recently published a blog announcing their return to conventional infrastructure. This company is exceedingly well-respected within the engineering community and far from Luddites stuck in the era of SCSI RAID arrays and AS/400s. For very-large-scale, steady-utilization workloads, it may well be true that buying and operating data center infrastructure is objectively cheaper than public cloud. I do not, however, believe the case can be made that it is better or faster in any meaningful regard other than cost. Sometimes cost is the most important factor, especially when that cost is in the millions of dollars annually. All but the largest users will have a difficult time making the cost argument because of the inherent efficiencies of public cloud vs. building and operating private infrastructure.
The divergence in focus between the re:Invent and IOCS was perhaps the most notable insight I took away from the conferences. Different audiences, agendas, ideas, and assumptions made for a striking contrast across the two consecutive weeks. re:Invent operates with the fundamental assumption that cloud is the future and AWS is (for all intents and purposes) the only cloud that matters. IOCS offered some content around cloud and included a few cloud vendors on the trade show floor, but primarily seemed to cater to the pre-cloud or never-cloud crowd. Gartner also holds Azure in high esteem. The cloud-centric or cloud-curious attendees at IOCS seemed outnumbered by more traditional IT managers and buyers.
At times, it was a bit like two parallel universes, aware of each other’s existence but not really that interested in or bothered by one another. Perhaps a better analogy would be religion: people immersed in one often implicitly assume that it is the correct answer, and that others are incorrect, irrelevant or misguided. The re:Invent crowd, like the BEV crowd, understands that the world has fundamentally changed and that companies that fail to adapt will fall increasingly behind the competition. The Gartner IT community, like Toyota’s insistence on an ICE future, has the enormous mass of the installed base, decades of history, and approximately 90% of the market. Convincing one camp that the other is right would likely be about as futile as trying to convert the devoted to a radically different faith. Neither is right or wrong – they are simply different solutions and approaches that suit diverse organizations and use cases.
Despite the enormous number of people encountered daily and the inevitable COVID exposure that carries, I returned home to Seattle exhausted but healthy and thrilled. Being in a crowd, at an after-party, or presenting live before hundreds of people brought back the energy and excitement of pre-pandemic conferences, and I would expect subsequent years to be even bigger and better than this year’s magic. A (very) few masks dotted the sea of people in the corridors of the Venetian, but that was the sole reminder that the pandemic had ever been an issue. Technology adoption presses always toward the future; its forward motion is self-sustaining through the most extreme of circumstances. Some companies will adapt quickly and others will resist. Those that keep pace with cloud evolution will have the capacity to weather whatever crises, large or small, that may arise. Cloud will become the predominant computing approach not because it’s cheaper, but because it’s faster, it’s better, and it provides a critical capability—business agility—that is essential for virtually any company to survive in the long term. Cloud is the key enabling technology that allows for maximum flexibility and endless adaptation to an ever-changing world and market landscape.
With that said, Toyota will be building ICE vehicles for years to come, just as Dell will be making servers and EMC will be making storage equipment. Gasoline and VMware will still be sold for decades. The transition from the massive installed bases of legacy IT infrastructure and ICE vehicles will not happen overnight – even if everyone agreed, the cost and scale of replacement is immense. The key is to chip away at the problem, scaling up the next generation of technology while respecting the fact that the old technology is not going away anytime soon and still deserves investment, attention, and respect from all involved parties.