AWS on track to be bigger than IBM by Christmas
theregister.com
theregister.com
These days, no one gets fired for choosing AWS, and people get hired for knowing how to work with AWS.
Instead of you paying IBM to run your "mainframe" infrastructure, you pay AWS to run your "cloud" infrastructure. It's the same kind of service, with the same kind of economics, but under a different guise. AWS even offers an "on-premises cloud" solution for large customers.
As was the case with IBM at its peak, once you start using AWS's alphabet soup of ancillary services that work only on AWS, switching away from AWS becomes about as easy and enjoyable as replacing all the plumbing in a skyscraper.
Azure is the main alternative. There's always a #2.
Now I feel like the pendulum has swung waaay too far. And I say this as an engineer who once worked for AWS. Of course you can still use EC2/ECS to do this sort of setup, but it seems everyone is oh so eager to throw everything onto Lambda and Fargate and completely abstract away the machine to the point where you have no insight or control as to what is running your application. It's absolute madness - it's like when my mother was programming on mainframes in the 70s/80s except instead of a batch of punch cards it's zipped files of python and js
At this point I find even running your own Kubernetes cluster to be more palatable than serverless solutions. With that at least you can still have some control into what's going on.
I see them both equally safe in terms of "won't get fired" at this point, although Azure maybe was risky as recent as 3 years ago, they seem stable now.
MS seems to have more of that super sticky engagement with customers that IBM had. Our MS tech teams seem unable to use technology from another vendor. Those on AWS like it and prefer it since it's familiar, but seem more adept at leveraging other technology commercial or open source. That super lock in MS has seems very IBM like to me?
It's a close race for sure.
Maybe they are both IBM, at different points in IBM's history?
the next wave of innovation could come from compnies like iExec and FileBase who offer storage and lambda-like utility.
the cost, allocation, and delivery is governed in a decentralized manner.
the value accrues to the network where you uave a vote.
interesting times ahead when thinking of a post-cloud-lockin world
You can argue with price or the peculiarities of a service or that AWS is taking over the world, but I have never seen them use contracts to force effectively companies to comply.
It's those megacorp whales, isn't it?
Which would make it a strange comparison (but ok, it would just emphasize the size of AWS)
IBM largely makes their money supporting legacy software, doing services and consulting work, and through Red Hat's software offerings
Although I believe mainframes are sticking with IBM. Those are still quite profitable, and you could imagine a "mainframe cloud" for clients that are still dependent on them.
I wouldn't say it's only megacorps.
Universities have internship courses with IBM, and IBM have both normal and research contracts with the universities. Even tiny universities with less than a hundred students.
All the banks have contracts with IBM. All of them. There are hardware contracts, software contracts, and customised development versions of both of those for the banks.
Most of the schools also have IBM contracts. Not generally customised development, just deployment of prebuilt hardware and preconfigured software. Even the kindergartens do.
All the ISPs have support contracts. They tend to be more secretive about what the contracts are, but from what I hear it's mostly prebuilt hardware and very slightly customised software, and the support contracts are where IBM make the most of the income from the various contracts with the ISPs.
All the hospitals have contracts. Customised hardware contracts, but boring and preconfigured software. However, the support contracts are fairly busy and IBM have a really high turnover in the teams that support the hospitals.
The government buildings, of which we have local, state and federal, all have IBM hardware in them. Some of which has been under active support for three decades. Software wise, though, the government is edging away from IBM due to some recent fuckups that have seen IBM blacklisted for new contracts.
Same hubris as Steve Ballmer about the iPhone.
I strongly believe that if company share prices and PE ratio levels reflect investor confidence in the future of the company, then isn't IBM trading at 23 PE an overvaluation?