2015 wasn't about AWS. It was about everybody getting ready to try to beat AWS One star eclipsed all others in the enterprise in 2015: Amazon. Or, rather, its cloud division, AWS. It somehow became a given that AWS leads in Infrastructure as a Service (Iaas), with the one-time ebook shop now a major IT infrastructure provider. That ebook flogger has become an IT infrastructure provider for the likes of …
@Credas
I agree, if interest rates affect one, it affects all and it does not matter if it is cloud, capital will be more expensive in house as well.
They also seem to have forgotten that in real terms IT spend goes down in price, or you get twice as much for the same cost, whichever way you need to look at it. What AWS does is let start ups in web apps get going and the successful ones, grow in such a way that AWS cost pretty much goes down in price year on year for both AWS and the start up.
Not sure how Norwich Union uses AWS, but Glacier & S3 would seem a no brainer, the cost to do that in house with the durability & availability offered by AWS, provided you can show regulatory compliance, why would you want to do that in house?
Just some late night rambling :-)
"And so it will for every alternative for the cloud."
I don't think that's true. When the cost of idle infrastructure goes up it will put even more pressure on achieving 100% utilization to eliminate idle infrastructure. The cloud shines brightest on cloudy days, because of the higher processor and storage utilizattion of those capital investments.
"All the others are using buzzword bingo and (legal) accounting tricks to hide how (badly?) they are doing in the cloud."
Except perhaps Microsoft - who report raw cloud revenue numbers - now exceeding AWS for the last 2 quarters and growing significantly faster - and remaining profitable whilst undertaking vast infrastructure growth.
I was amused to note British Gas listed as a significant AWS customer - from working there I know they have hardly anything in AWS - the vast majority of systems are currently with HP - and their current growth in the cloud space - as with many corporates - is into Azure...
This 'capex is a threat to AWS' stuff has been printed here a few times now.
What does the author think enterprises build datacentres and buy on-premise hardware with?
If anything the threat of increasing interest rates makes the dominant cloud players more attractive. Amazon and Microsoft have shown a willingness to take substantial short term hits to maintain their growth path and each is now spurred on by the other.
Enterprises, on the other hand, are increasingly capex averse and will definitely factor the cost increases into their financial modelling.
My prediction is that as interest rates go up AWS and Microsoft will take a hit on profits (though probably hide it) and keep service costs level to make their service look even more attractive to bean counters.
Of course, in the long run any enterprise that gives up its inhouse capabilities entirely will suffer the same way clients addicted to Oracle and SAP suffer now, but that kind of long term thinking rarely plays a role in CFO thinking.
The idea that rising interest rates pose a particular threat to public cloud looks financially illiterate to me.
"Of course, in the long run any enterprise that gives up its inhouse capabilities entirely will suffer the same way clients addicted to Oracle and SAP suffer now, but that kind of long term thinking rarely plays a role in CFO thinking."
Exactly. That's just trading one dependence for another. Hell, in some companies /corporations it's difficult enough to control your own IT.
I have no data to back this up, but my gut tells me that at least half the decisions to "move our IT stuff into the cloud" are not based on rational thought and reasoned analysis whether this is a good move for the company, but on the old "everybody is doing it, so we must do it too".
This weeks about turn, admitting their Xbox one cloud computing exclusive feature (azure) isn't that special and could be done on PS4 too.
It seems the dismal sales of Xbox one, and the slow adoption of azure cloud, Microsoft needs its competitors more than they need them.
http://www.cloudpro.co.uk/collaboration/5562/xbox-live-news-ps4-games-can-use-azure-cloud-power-says-phil-spencer
"Dismal sales"? What have you been smoking anti-Redmond hateboi? PS4 is clearly ahead by a very good margin but describing the Xbox sales as "dismal" takes FUD to a new hight. If you really think that the link that you provided justifies that assertion you should change your meds.
As someone who's been around the industry for awhile (and is hopefully not too curmudgeonly), am I the only person not seeing Amazon adding 8,000 new features a week (slight exaggeration) to AWS as a bad thing? For mission-critical stuff, stability is a very nice thing. We study a platform, implement on it, and then weigh the cost / risk / benefit trade-offs of every future change. For a lot of what we do, any new risk is deeply frowned up - of course, there is also risk in not changing things, but that's why we have grown-up discussions about what's the better path. We're battling herd-based mentality from the (generally young, naive) technophiliacs who want to do every new thing and the (generally old, bitter grey-beards) who want everything to be done the way it was back in the 1970s. There are exceptions, of course - young people with decent perspective and experienced CIOs who move services into the cloud more or less because "everyone is doing it" (which makes it politically safe).
But back to AWS. It's nice that they cut prices, but their pricing model already rivals the US tax code in its complexity, to the point where there are dedicated consultants that specialize in purchasing AWS services. And the thought of a bunch of people - no matter how smart - constantly pouring new features all over production systems scares the crap out of me. As an end user, we have no real control over this. We can't spin it up in a lab and say "oh wow, this is great" or "oh wow, this really sucks." (sh)It just happens and you deal with it. If you're running a mostly stateless service in giant clusters of machines (a la Google or Netflix or most big web sites) then that may be acceptable, but for a lot of systems that just doesn't fly.
A few points come to mind.
Firstly, with regard to pricing, I like the way AWS transparently reflects their actual costs. Different products have different combinations of pricing (e.g. S3 charges separately for the amount you have stored and for volumes of GET requests). This is a good thing, as it ensures that application developers are rewarded or punished for their understanding (or lack of it) of the underlying infrastructure effort needed to power their apps.
Our AWS planning always begins with a spreadsheet where we estimate various key loads created by our user base and the resultant costs. This is also a good thing. In the bad old days, you might build your app with little concern about the infrastructure, then other people would drink from the poisoned chalice you'd prepared. Today, you can model the infrastructure costs with a reasonably high degree of accuracy, free from the distortions of the capital spending budget.
Secondly, about those endless new features. It's true that it's difficult to keep up, but that's also a good thing. In our experience, none of the 'old' features has become less reliable or powerful as new features are added - quite the opposite. It's worth understanding the new features so you can count them in your armoury for future projects, but our experience of committing to an AWS feature is one of being impressed by the improvements in that feature over time.
In Summary, this article has it right. There are few reasons why a developer would choose a poorer, less-well-executed set of integrated cloud services than AWS in 2016, unless you were told to, or were constrained by considerations other than what is the best tool for the job.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
