Microsoft Will Be Largest Infrastructure As A Service Vendor By 2019, Says Morgan Stanley Survey (geekwire.com) 75
An anonymous reader writes from a report via GeekWire: According to Morgan Stanley's 2016 CIO Survey of 100 CIOs (75 CIOs based in the U.S., and 25 based in Europe), Microsoft's Azure will overtake Amazon Web Services (AWS) by 2019 to become the largest Infrastructure as a Service (IaaS) and Platform as a Service (PaaS). The survey finds that roughly 31 percent of the CIOs will be using Azure for IaaS, compared to roughly 30 percent using AWS. Today, roughly 21 percent are using AWS and 12 percent are using Azure. According to the survey, Azure is already leading AWS in PaaS, used by about 18 percent of the respondents, versus AWS's 16 percent. Azure's lead will grow slightly by 2019, growing 9.8 percent versus 6.4 percent. Nearly 30 percent of all applications will be migrated to the public cloud by the end of 2017, up from 14 percent today, the survey said. On-premises apps will decline to 58 percent, from 71 percent today. Predictably, hardware vendors, including conventional and flash storage makers, will continue to suffer as their market is eaten by the cloud. Hardware spending growth is down this year to 3.2 percent, from 3.4 percent last year. Microsoft recently announced it will be entering the legal marijuana industry. It will partner with Los Angeles-based startup Kind on a system for tracking the legal growing and sale of marijuana, with Microsoft powering the software through its Azure cloud computing service.
Just so I understand (Score:2)
There is no hardware in the Cloud®? All my data is stored in water droplets?
Re: (Score:2)
There is no hardware in the Cloud®? All my data is stored in water droplets?
I used to think this way, if the hard drives aren't in my private server, they have to be in the cloud? How could the hardware vendors be hurting, right? I think it's because the cloud providers use resources so much more efficiently. Here is what I mean: Let's say you have 1,000 businesses, each with a 1 TB SSD in their private on prem servers. Let say each business on average uses, 250 GB of storage. Now let's say those 1,000 businesses, with 1,000 SSD's in the field were to move to the cloud over
Re: (Score:1)
No, you were thinking exactly as I was. Data needs hardware, and storage is used much more efficiently "up there", so yes, sales will be down. The question is, with everybody saying how the market is priced by supply and demand, are hardware prices collapsing? If not, then the market isn't doing so badly after all. A 0.2% decline in growth doesn't sound very impressive one way or another. Yeah, it could be billions, but it's spread petty thin.
Anyway the joke went over like a lead balloon with the moderators
Re: (Score:2)
Maybe the storage vendors will continue with their trend of consumer grade and enterprise grade stuff... the gap in
Re: (Score:2)
Storage vendors could pull what the government in Ontario, Canada did to us.
No they can't. The government of Ontario is a monopoly, so they can unilaterally raise prices. Disk vendors compete. If one raises prices, it loses market share. If they all raise prices, they may be prosecuted for price fixing, and will also encourage other companies to enter (or re-enter) the storage market.
Maybe the storage vendors will continue with their trend of consumer grade and enterprise grade stuff
Cloud companies do NOT use "enterprise grade". They use consumer HDDs. "Enterprise" HDDs have no reliability or lifespan advantage over consumer drives. They just have an extended warranty, and
Re: (Score:2)
Disk vendors compete. If one raises prices, it loses market share. If they all raise prices, they may be prosecuted for price fixing, and will also encourage other companies to enter (or re-enter) the storage market.
This is bullshit. There's TWO vendors: Seagate and WD. Nobobody can enter that market, they own it.
SSD is a different story because the technology is still evolving, but it's a tiny market. About 15% of new laptops sold have SSD, and in server/SAN it's even less.
Re: Just so I understand (Score:2)
Three. You forgot Toshiba.
Re: (Score:2)
True. Still it's quite a depressing pie chart...
Re: (Score:2)
"They use consumer HDDs"
And I assume software RAID? Consumer HDDs attached to a hardware RAID adapter is a very risky proposition, even when using OBR10
Re: (Score:2)
And I assume software RAID?
Correct. AWS uses software raid only [amazon.com]. Azure likely does the same.
Re: (Score:2)
Storage vendors can also expand their products to more niches as well:
Fault-tolerant drives come to mind. I remember one drive maker having a model with two active/active independent sets of drive heads. If one set failed, drive throughput would be slower, but the data can be pulled off.
Different drive shapes. It was mentioned a few weeks ago about having taller 2.5" drives so platters could be stacked higher. This would be useful for arrays.
Hybrid SSD/hard drives, where the HDD itself does the autotier
Re: Just so I understand (Score:2)
Same for server hardware and network gear in the form of virtualization.
Re: (Score:2)
Cloud providers have vastly more redundancy requiring much more raw space.
I'd be interested to see any real numbers if anyone has them, but my guess is it's a wash.
Re: (Score:2)
I would almost guess that the need to accommodate growth spikes plus redundancy would make cloud providers at best a wash with on premises storage, even accounting for thin provisioning by cloud providers.
The thing that could possibly give cloud providers a real edge is deduplication across storage, but that has a potential performance penalty and a real CPU and I/O cost.
Keep dreaming... Azure is super sketchy.. (Score:5, Informative)
Azure storage services has got some awesome consistency guarantees.. But it won't scale like S3, not in terms of requests, throughput or features.
Microsoft has an abyssal story for authorization and access policies... There is nothing like IAM that crosses all services, some services have policies, other services you get shared secrets (to be shared between all users)..
Honestly, they can't even figure out to make a consistent naming policy between different storage services... Figuring out what characters is allowed in names of fields, resources, urls, etc. is a nightmare... Even with azure storage services (queue, table and blob) they have vastly different restrictions... It's a joke.
The only thing interesting with Azure is their table storage service, price and simplicity wise it's a joy. But given how bad everything else, I'm tempted to move my stuff to AWS and pay a bit more for dynamodb...
Re: (Score:1)
Haven't had any opportunity to so past couple of years..
Re: (Score:2)
Re: (Score:3)
How is Azure sketchy?
Did you read the GP's post? He gave examples right there...
Re: (Score:3)
I also have a few with Amazon too. No trouble there either. How is Azure sketchy?
An azure storage accounts have a single secret key shared between all users... If you have two servers/apps/persons using the same storage account they MUST share the same secret key. You can issue temporary keys, but you have to build an manage an authorization system that issues such keys. The user management in azure does not extend to cover storage accounts other than all or nothing, and all users share the same secret key. This is insane! Unthinkable in any non-trivial deployment.
Re: Keep dreaming... Azure is super sketchy.. (Score:2)
AWS IAM is great and so many ways to establish identity. I have only a little experience with Azure, but it does seem a lot less flexible.
Re: (Score:2)
Re: (Score:2)
Try DynamoDB for a while and see what happens. It's a truly horrible product. Impossible to predict performance or price, it's like the worst the cloud has to offer, all in one product.
Re:Keep dreaming... Azure is super sketchy.. (Score:4, Insightful)
A bigger war chest doesn't guarantee victory or Microsoft would have eliminated ALL comers 15 years ago.
Scratch that - the only player in any industry attached to IT would be IBM.
Google has a *HUGE WAR CHEST* (tm) too
Re: (Score:2)
there's 3 years of development time between now and 2019, and with Microsoft's deep pockets
Microsoft is stupid... Sad but true. They are not developing consistent services. Throw whatever money you want after it, if you have no single user manage, authentication and authorization system covering all APIs you loose. If you have different arbitrary restrictions on what ASCII chars is allowed when naming resources for different services (just in azure storage service, not counting everything else), it's going to fail...
AWS is not perfect, but it is fairly consistently designed... As in IAM users a
Re: (Score:2)
Other than cost (where Amazon gets you coming, stashed, and going), what is so bad about S3? OpenStack's Swift is maturing rapidly, but S3 is still ahead of it when it comes to features.
Re: (Score:2)
I know of only a few companies using Azure - and all of them are Microsoft consultants.
Most of the enterprise shops I deal with are on private cloud using Redhat or VMware, or Openstack with Rackspace or Dell or are on AWS.
"Terminal Stupidity Is On The Rise.. (Score:1, Troll)
Re: (Score:1)
They make 20 billions in profit per year. This means that in a single hour they make more money than you will do over you entire life.
So maybe you should take the smugness a few notches down.
Re: (Score:3)
FFS, did you just the compare the revenue of ONE fucking guy to company that's 40 yrs old and has 100,000+ employees?
Perhaps you can dial down your own smugness and tell us how M$ is playing catchup to VMware & Amazon, the latter being a online version of your general country store.
It's like VW losing to Walmart in automotive innovation
Re: (Score:2)
FFS, did you just the compare the revenue of ONE fucking guy to company that's 40 yrs old and has 100,000+ employees?
Warren Buffet makes 1.5 million per hour. That's four times more than Twitter, and he's just one guy. He didn't inherit that money, he earned every penny.
Strangely, Warren Buffet doesn't make smug comments about the need for highly profitable companies like Microsoft to "die quietly".
Re: (Score:2)
Warren Buffet is an old school investor who's long admitted he doesn't understand tech so feel free to find another example.
Or explain just how a book-and-toothbrushes online store became the biggest name in The Cloud, even though they've never really courted the bread-and-butter enterprise market and heavyweights like M$, HP, IBM, Oracle & Google are playing catchup 10 years on.
Re: (Score:2)
That book and toothbrushes online store invented cloud computing, but they're not making money with it.
Re: (Score:2)
I beg to differ - http://www.recode.net/2016/4/2... [recode.net]
Makes me wonder (Score:2)
they're doing it wrong (Score:3)
Azure is a lot more expensive than AWS, and they tend to increase price and decrease services. Meanwhile, AWS is on a mission to constantly lower prices.
The logical outcome would be for AWS to see their market share grow, but the opposite is happening. Why? I think it's because they don't give a shit about enterprise customers. Even if you spend $55,000 per month on AWS you'll never get incentives or personalized service from AWS. On the other hand, enterprise customers get VIP treatment from Microsoft, wth steep discounts and effective, even proactive support.
A while ago maybe it mattered to get rock bottom price for cloud computing. Now people want more than the Walmart experience when they expand their infrastructure in the cloud, and that's why Azure is growing faster than AWS.
I still use AWS for my own projects, but I'm spending tiny amounts of money compared to the two clients I've recently helped move from AWS to Azure. And their main reason to switch was that they want to rely more and more on the cloud, and AWS couldn't be bothered to answer emails while Microsft sent PEOPLE to meet and wow them and make it enticing to consider Azure.
Re: (Score:1)
Azure is a lot more expensive than AWS, and they tend to increase price and decrease services. Meanwhile, AWS is on a mission to constantly lower prices.
WTF? I think you have been listening to AWS marketing BS, price is actually one of the key benefits that is making Azure grow so fast. Azure almost across the board is similiarly priced or cheaper. We run both AWS and Azure and are gradually moving from AWS to Azure over time purely based on price as functionality wise they are pretty similar for us.
Re: (Score:2)
No. Retail price for VM is consistently 25-40% higher on Azure.
Ex: a 1.5GB RAM instance is $35 on Azure, while an instance with 2GB is $20 on AWS. Both run Linux.
Re: (Score:2)
No, what you don't get is the big picture
Re: (Score:2)
what are you talking about? the free tier on AWS means that for 1 year you get some services for free, it doesn't give a discount beyond that. And what the fuck is an Azure VPS instance?
It's obvious that you're not an Azure or AWS customer and that you don't even know what cloud computing is, so stop trying to peddle your guesses as facts and go play with your linode.
Re: (Score:2)
Azure offers to match AWS prices to clients with an ELA, but you'll have to provide a link to prove that AWS does the same because I've seen no situation where it's cheaper to use Azure. That's like saying that Walmart will match Pier 1 prices.
Re: (Score:2)
Also, bear in mind that with Azure*, you can make a private or Hybrid cloud that works, acts, is controlled and is programmed exactly the same way as a public one.
If you use AWS for your public cloud needs, and want to also deploy private/hybrid cloud, you need to handle those in a totaly different way at all levels...
* Also with OpenStack, but OpenStack is WAY HARDER than Azure.
Re: (Score:2)
OpenStack's biggest advantage is the F/OSS nature of the product. For companies and organizations that might not have money for licensing, but have lots of people and man-hours to throw at OpenStack, it might be a solid solution, although there are things like VMWare's Fault Tolerant VMs and HA items which Nova really needs for it to be more enterprise friendly.
Re: (Score:2)
Fault tolerance in VMware is neat and a lot better in v6, but it requires a hell of a lot of network traffic for sync. A four CPU VM running FT eats 60 MB/sec sync, and that's just a bog-standard Win2012r2 VM running updates. It throws warning flags running it on 1 Gbit links.
Even with 10 GBit Ethernet, I'd be wary of trying to rely on FT for a server with any appreciable CPU and disk workload, and even then its recommended to have a second datastore for the FT VM, so the resource requirements for best pr
Full Circle (Score:1)
Microsoft was founded on the principles of local computing and in effect 'anti data center/mainframe'. Now people ( companies ) are seeing that it really was the right idea after all.
Have these polls ever been right? (Score:4, Interesting)
I don't have hard data, but my memory tells me they get it wrong a lot more often than right.
Re: (Score:1)
>A new, international survey of 100 CIOs
These surveys only tell you the mood of other CIOs. Which if you are a sales guy is useful. Or if you are a CIO and you want to make the "safe" choice, you at least know what that is. "Everyone is doing it" goes a long way with the boss.
>Roughly 31 percent of the CIOs will be using Azure for IaaS, versus roughly 30 percent using AWS. Today, about 21 percent are using AWS and 12 percent are using Azure.
This says to me that if you aren't an early mover, you are mo
Re: (Score:2)
This says to me that if you aren't an early mover, you are more likely to answer "Microsoft" to any questions about vendors.
Also the term CIO is heavily biased toward big slow movers. I'm working on a large project with a lot of small developers, none of which have CIOs. And they are all 100% AWS.
Azure disk/ssd IO performance (Score:3)
Last I checked, Azure had possibly the slowest disk and ssd IO performance of any cloud service. Amazon crushes them in this space.
MIaaS (Score:2)
Market share doesn't equal size (Score:2)
Setting aside the accuracy/sample bias of a survey like this, they seem to fundamentally misunderstand the business.
How much infrastructure does the average customer on each one actually use? I mean if it's 30 and 31 percent, but the average customer on AWS is using five times the resources, Azure is still much, much, much smaller.
microsoft could also be obsolete too (Score:2)
when it comes to speculation there is two sides to that coin,
I believe it (Score:2)
"Microsoft Will Be Largest Infrastructure As A Service Vendor By 2019"
What with them changing Windows to a subscription service and the insane success of their Azure stuff, I wouldn't be a bit surprised if this turns out to be true in a year or two.
I don't know what it is about Azure that's so compelling, but it does seem like a shitload of businesses are going for it in a big way. MS can't build datacenters fast enough to keep up with demand and that's gotta mean something.