Erik Bernhardsson, the previous CTO of Better.com, presumed the idea that “Cloud suppliers will significantly focus on the lowest layers in the stack … [while] [o] ther pure-software providers will develop all the things on top of it.” This is definitely not the cloud world we presently live in, with AWS CEO Adam Selipsky taking the stage at AWS re: Invent 2021 to talk about how AWS will “continue to build more of these [vertical market] abstractions on top of our existing fundamental services,” with Google and Microsoft currently well down this course of market options, not to discuss applications, databases etc. That seems like more development up the stack, not less. (Disclosure: I formerly worked for AWS.) But Bernhardsson makes an engaging point: The cloud suppliers may get spread out too thin to complete successfully with pure-play handled service suppliers. This might be real, however its hard to see the cloud vendors providing up their need to grow, and theres a lot more money in applications, for instance, than running systems. SEE: Hiring Kit: Cloud Engineer ( TechRepublic Premium) The argument
Commentary: There are factors to believe the clouds would be wise to return to the core facilities that made them popular, but there are other reasons to think this will never happen.
According to Bernhardsson, theres money to be made by focusing on what cloud vendors utilized to do: just core facilities. No, this isnt a product business. One of the most crucial announcements AWS made at re: Invent today was the introduction of the Graviton3 processor. As Tom Krazit said, “Graviton is a years-in-the-making cloud-infrastructure moat that is drawing converts focused on one of the most fundamental cloud computing concern: How much will it cost to run my application?” There is real cash to be made, and genuine distinction to be had, at the core infrastructure level. Its also the area where the clouds face the least competitors, which is where Bernhardssons argument makes the most significant dent.” [D] eveloper experience has actually become an attack vector,” he stated, with start-ups like Databricks and Snowflake outflanking their more recognized cloud provider peers through tighter focus and better designer experience. For the cloud providers, which provide the underlying facilities for all of these application/data warehouse/etc. upstarts, theres a lot of money to be made in partnering well, as Bernhardsson argued: Lets say a client is spending $1M/year on Redshift. That nets AWS about $500-700k in gross profits, after spending for EC2 operational cost and depreciation. If that client changes their $1M/year budget to Snowflake, then about $400k returns to AWS, making AWS about $200k in gross profits. That seems kind of bad for AWS? I dont know, we ignored a lot of stuff here. Snowflakes predicted 2022 research and development expenses are 20% of revenue, and their sales and marketing costs are 48%! For a million dollars earnings, thats $700k. Translated back to AWS, possibly AWS would have invested $300-400k for the same thing? Appears reasonable. Now the mathematics unexpectedly amounts to me. AWS essentially ends up with the same bottom line impact, but effectively “contracts out” to Snowflake all the cost of structure software and offering it. That appears like a good deal for them! Bernhardssons mathematics appears roughly affordable to me, and Im a strong supporter of the concept that the cloud suppliers can not and need to not develop a managed service for every area of software application (from call center services to databases to operating systems to … the list is apparently endless). AWS, for example, now has well over 200 services. That alphabet soup of services (a lot of which take on each other) makes it puzzling for clients to know which to utilize for something as simple as running containers. (AWS has 17 different methods to do this.) Heres where the logic, compelling though it might be, begins to break down. The counterargument If youre in enterprise IT, you understand that your invest in applications, databases etc overshadows what you spend on operating systems and storage. Its very important, but the closer software application gets to the customer, and the more that software assists you to provide a much better customer experience, the more youre going to pay. Little marvel, then, that Selipsky, in John Furriers annual interview with the AWS CEO, stressed how much the company plans to concentrate on industry options. However even looking at current AWS services like Amazon Managed Service for Kafka (MSK) or the Amazon OpenSearch Service, its tough to see AWS giving up on big services to retrench around core infrastructure. Its not that Bernhardssons logic is inaccurate, simply put, but rather that theres another logic involved, which is “revenue development.” AWS, Google Cloud and Microsoft Azure dont have the high-end anymore of returning to core primitives like calculate and storage. Not without setting their stock rates on fire. SEE: Multicloud: A cheat sheet (totally free PDF) ( TechRepublic) Perhaps, to utilize Bernhardssons example of Redshift, gradually these cloud giants will find that their native, higher-order services keep losing out to nimbler, more concentrated competitors. Maybe. Thats certainly taking place in some locations currently throughout the huge clouds. But its likewise real that these cloud service providers are at times providing exceptional services. One really standard, but apparent, example is how each of the clouds has an exceptional MySQL offering than Oracle … which in fact owns MySQL. (I can say “superior” with total confidence due to the fact that Oracle doesnt actually offer a MySQL handled service, which is rather baffling. Just spinning that up and tossing some marketing at it ought to deserve a few hundred million, if not a billion. And yet …) But even without losing to their partners, the clouds might discover that they can construct an even bigger company, and drive more development, by much better making it possible for partners. Were nowhere near that decision at any of the clouds, but maybe it will play out in time, on a service by service, and partner by partner, basis. We will see. To see if Bernhardssons argument will triumph, look for a cloud vendor ambushing any of their native services in favor of much better supporting a partner and raising option. It hasnt took place yet (to my understanding), however when/if it does, that will be a sign that Bernhardsson might have led his time. Disclosure: I work for MongoDB, and previously worked for AWS, but the views revealed herein are mine alone.
Image: sigoisette, Getty Images/iStockPhoto
Heres an insane thought: What if the cloud suppliers decided not to continue to move “up the stack” to applications and other higher-order services? Its not my idea, however its an intriguing one.
Cloud and Everything as a Service Newsletter
This is your go-to resource for XaaS, AWS, Microsoft Azure, Google Cloud Platform, cloud engineering jobs, and cloud security news and suggestions.
Bernhardsson makes a compelling point: The cloud vendors might get spread out too thin to contend efficiently with pure-play managed service suppliers. Bernhardssons mathematics seems roughly reasonable to me, and Im a strong supporter of the concept that the cloud vendors can not and need to not build a handled service for every area of software (from call center services to databases to running systems to … the list is apparently endless). AWS, Google Cloud and Microsoft Azure do not have the luxury any longer of moving back to core primitives like calculate and storage. Were nowhere near that choice at any of the clouds, however perhaps it will play out over time, on a service by service, and partner by partner, basis. To see if Bernhardssons argument will win out, see for a cloud vendor scuppering any of their native services in favor of much better supporting a partner and raising option.