The winners and losers of infrastructure clouds revealed: AWS, Microsoft, Google and Alibaba get fatter AWS remains the biggest provider of infrastructure clouds with a revenue haul of $34.6bn in 2019 but for the first time it has pulled in less than double that of nearest rival Microsoft, which inhaled $18.1bn. That's according to the latest figures from research company Canalys, which said worldwide cloud computing spending …
Is anybody genuinely surprised that the big are getting bigger? The cloud business is ideally suited to favor economies of scale. And like enterprise software business in the days of on-prem software, it's almost inevitable that you'll get 3 (in this case 4 because of geography and geopolitics) big winners and a host of distantly-trailing niche vendors. Look at the evolution of the markets for relational databases, Linux distros, application servers, ERP/CRM suites, and any number of other software markets...
The big problem with cloudy services is that it's like "crack" for CTO's.
They know it costs money, but because it's so scalable, they keep dumping new services on it and can justify it by saying there's no CAPEX.
I wished I'd got into this racket a decade ago!
So true. Then we are going to hit what I have heard people call "Cloudshock", when people suddenly realise how much cloud is costing and pull some things back in house (test environments being an obvious choice). I have heard stories of cloud companies that managed to very quickly spin up due to cloud but suddenly find 40% of their turnover was going straight to AWS.
People are quite naive about the cloud. I have heard a senior exec say in a meeting with all this competitions (between AWS, Azure and Google) things will just keep getting cheaper...Wow.
And an accountant saying to me, will the cloud deployment be always cheaper than our present On Prem one...We are simply swapping CAPEX for OPEX really was my answer.
This is just the latest version of what the Jargon Dictionary used to call the "The Cycle of Reincarnation". Their example was
"function in a computing system family is migrated out to special-purpose peripheral hardware for speed, then the peripheral evolves toward more computing power as it does its job, then somebody notices that it is inefficient to support two asymmetrical processors in the architecture and folds the function back into the main CPU, at which point the cycle begins again.".
This is just a new cycle of the mainframe -> distributed systems -> cloud (mainframe), with a different hat on.
Absolutely no-one is going to be bringing their cloud test environments to an on-prem solution. The ONLY time that might be viable would be if they have kit sat idle which is already using money however the additional engineering costs would almost always outweigh the expenditure recuperation and it wouldn't be 'live-like' and wouldn't be a suitable fit.
If we're talking nearly 40%, the 'others' category is a bit too big to just lump them together without naming them. Who are they?
If I want to look into a relatively serious cloud player but not an American (no, an American firm with a DC in Ireland doesn't count) or Chinese company, where do I go?
It's like crack for system architects, too.
A certain regional water company has relatively recently started shoving all of its telemetry and SCADA data into Azure data lakes. All of it. Terabytes per year. Now, Azure data lakes are very cheap - it doesn't cost much for the data to just sit there, so fair enough. However, when the time comes to do something with that data, that's when they've got you by the shorties.
There will be so much data that it will need to be partitioned in some way before you have at it with something like Hadoop (or whatever the MS equivalent is), which will mean:
* An IT project in its own right within the organisation;
* Hiring of big data analysts/specialists;
* Hadoop/Map Reduce specialists;
* Additonal cloud resources the churn the data, partition it, validate it.
And then, when you realise how much it's going to cost you also realise that they've got all your data. I mean seriously, what are you going to do? Ask them to post it to you on a USB stick?
I thought the idea of cloud was you hired the infrastructure for the time you needed it, churned, got your results, and span it down. Just like it was in the 60s and 70s. But modern IT departments seem to be falling into the trap of 'the cloud' - presumably because it is easy to use, and can be very cost effective. But you need to strictly control what you're using it for, and so far I'm not seeing that. Therby it becomes a false economy; they're merely shifting the costs into a different column.
Well, it being cost effective is *very* difficult to judge upfront (partly due to lack of transparency). The decision to 'move to the cloud' are usually made as part of 'digital transformation' projects that hardly ever explore the downsides of the cloud (everyone is doing it, so why shouldn't we?).
That and obviously there is arguably a conflict of interest, the big names mentioned in this article wouldn't gain much by making the cloud *too* cost effective, what would be in it for them, really? On the contrary, as an example, they are convincing people to pay for things that were previously free (by approximation) such as executing functions (Azure Functions, AWS Lambda).
Maybe the regional water company you mention and government in general are much better off building their own cloud? (which I think they have done already).
I see too much cloud as being like too much imports. The US eliminated large chunks of its in-country manufacturing capacity because "doing it in China is cheaper". Which is fine until something comes along that makes importing the stuff back impossible. Things like COVID19, and the need to eliminate the burning of fossil fuels. Then you are stuck not having control over your essential resources.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
