Welcome to Mostly Cloudy! Today: Uber makes a strong commitment to cloud computing, GitHub declares its auto-coding helper is ready for the enterprise, and Twilio makes another big round of layoffs.
Photo: by Dan Gold on Unsplash
Your instances are waiting
Uber built a formidable array of its own information technology assets and engineering talent as it stomped its way around the world. This week marked an abrupt shift in that strategy, after Uber signed two separate seven-year cloud deals with Google Cloud and Oracle.
Uber has been using cloud services from AWS and Google for several years, but 95% of its workloads ran in traditional on-premises data centers and the hardware supply-chain disruptions caused by the pandemic finally started to expose the flaws in that strategy, Uber’s Kamran Zargahi told the Wall Street Journal.
Exact details of the split between the two providers were not disclosed, but Uber plans to run the mapping services that route drivers to passengers on Google Cloud, and move Uber Freight workloads to Oracle. Likewise, we don’t know how much Uber plans to spend across the multiyear cloud deals, but I’d have to imagine both companies were willing to extend healthy price discounts to secure a deal of that length.
In a blog post last year, Uber outlined how it wanted to “reimagine our infrastructure stack for a hybrid, multi-cloud world,” tweaking the way it launches new cloud availability zones and manages existing deployments. That post is a great example of how much internal work companies need to do to support multiple cloud providers at scale.
“Crane has arguably been one of the largest engineering endeavors in which Uber has ever engaged,” the authors wrote. “It represents a large cultural shift, from an organization defined by bespoke, manual toil to one defined by heavy use of standardized automation. It required collaboration from many teams.”
Each cloud provider has a slightly different way of accomplishing similar tasks, and Uber engineers built an abstraction layer over cloud APIs to help its developers launch and maintain projects across those providers. Companies that attempt to do multicloud without some kind of abstraction layer can wind up in silos, with internal AWS experts or internal Azure experts playing gatekeeper to launching new apps because nobody else understands how to make it work.
Assuming its system works as designed, Uber could make up for some of its corporate sins by open-sourcing that system to help other companies manage multicloud complexity. With even AWS now on board with the concept of multicloud computing (and the shine wearing off Kubernetes month by month) something will need to emerge to fill that gap.
GitHub takes the wheel
While one of Microsoft’s generative AI projects has crashed back down to earth after the eye-rolling hype over the relaunch of Bing, another one of those projects is gaining steam. GitHub announced this week that Copilot for Business is now broadly available for $19 per user per month, extending enterprise-friendly features such as policy management and license management to the original Copilot product.
While there are still a lot of issues to be solved when it comes to the widespread use of an AI assistant for coding, it’s clear that software developers appreciate the help. Lost amid all the “learn to code” jeers from a certain segment of the population is that a bigger-than-you-think part of coding involves repeating simple tasks over and over again, as opposed to harnessing the power of computing with incredible intellect and skill behind every keystroke.
Given how central GitHub is to modern software development, it’s far more likely that Copilot will change the status quo much faster than Bing. And it won’t even try to make you feel bad.
Around the enterprise
Twilio laid off 17% of its employees months after an 11% cut, and two high-level executives have also left the company.
Residents of Northern Virginia are starting to complain about the relentless pace of data-center expansion in their back yards, one month after AWS announced plans to spend $35 billion in new construction and renovations in the area.
With quantum computing still a distant dream, one of its most promiment startups, Righetti Computing, is in deep trouble.
Former Googler and AppSheet founder Praveen Seshadri’s Medium post about life inside Google right now is a must-read for anyone trying to understand the company’s predicament nearly 20 years after its IPO. Bonus points for all the shade thrown at Who Moved My Cheese?
Thanks for reading — see you next week!