Slashdot Mirror
In Three Years, Nearly 45% of All the Servers Will Ship To Cloud Providers
dcblogs writes "IDC expects that anywhere from 25% to 30% of all the servers shipped next year will be delivered to cloud services providers. In three years, 2017, nearly 45% of all the servers leaving manufacturers will be bought by cloud providers. The shift is slowing the purchase of server sales to enterprise IT. The increased use of SaaS is a major reason for the market shift, but so is virtualization to increase server capacity. Data center consolidations are eliminating servers as well, along with the purchase of denser servers capable of handling larger loads. The increased use of cloud-based providers is roiling the server market, and is expected to help send server revenue down 3.5% this year, according to IDC."
In a three years, nearly everyone will send their own data to the NSA without even having to be asked!
Sent from my ASR33 using ASCII
[I am a developer not an admin]
It takes us weeks to months to get a new server provisioned and ready for use where I work. We did a MAJOR project years ago with the promise that it would take less than half an hour to do so, but that is never the reality. They put in huge servers with virtualization, a SAN, and everything else they asked for to do this, but they just don't. It has turned our workplace into slow IT because of admins not because of development. We can develop a solution in days and then take months to deploy.
Now we can within an hour have our server set up in Rackspace, have our network admin make a firewall rule for it and it is all set up within the day. Our admins are making themselves irrelevent and they don't even realize what they are doing.
BTW, I am 100% against using "the cloud", but am having a very difficult time justifing that position with what I see on a daily basis.
Almost 100% of all coal is shipped to electricity providers. Reliability and Economies of Scale.
That agenda is pushing dumbass CIOs into making bad decisions. Cloud Services, Co-Lo Hosting and the services wrapped around them are good tools to have at your disposal but like any tool if you don't know how to use them you can leave your organization high and dry. IDC and Gartner have a vested interest in selling Cloud and their associated third party service vendors to businesses since they're market makers. They're no different that your stock broker calling you up trying to sell a stock that's on their "hot sheets" to drive revenue. Companies pay these idiots for their "research" which is usually some guy sitting down and reading Internet articles and going to conferences where they hear long sales pitches from CSC, Rackspace and Amazon. None of this replaces a good set of people and an Enterprise Architecture strategy that the organization needs to develop and own.
What IDC misses here is two of the big cloud players, Google and Amazon, are growing their own servers so IDC's true "insight" should be that HP, Dell and IBM are going to lose server revenue more not from larger bulk deals with cloud providers but the fact that the bigger players are just going to buy components. Also companies aren't writing blank checks to their IT organization anymore. This means those big budget projects where you roll in racks of servers will be pushed more and more to virtualization. There's also the aspect that there are a lot of businesses who will never let their data or their customers data fall into the hands of any third party, even a hosting provider and they will still need servers and disk and products because year after year their existing footprint gets older and you need more capacity and to refresh your infrastructure.
Harrison's Postulate - "For every action there is an equal and opposite criticism"
Where some PHB does the buying of your cloud system so you can be stuck with low end systems, small bandwidth , small web space and so on as the PHB picked bob's cheap cloud space.
This. They seem to not take into account that part of the reason that companies want servers it to put private information there. And some countries have strong privacy policies for protecting their citizens information. That should ban a lot of companies for storing their servers into US based clouds at the very least.
My prediction would go into the opposite direction of cloud servers, toward personal/home servers, increasing the use of p2p/mesh encrypted networks and services, at least in the countries not actively cooperating with the NSA. That should be bigger than emerging country specific clouds.
One thing that /. readers often fail to take into consideration is that many companies may find that it's easier to outsource to a company with a solid reputation for hiring good people than to try to hire good people on its own. For smaller companies in particular, there's a hiring bootstrap problem here. They have to hire the right people who will be able to identify the candidates to build a solid IT team. A lot can go wrong, and many companies may in fact benefit from outsourcing to a reputable company who they can sue the hell out of if there is an issue and a highly paid consultant can point the finger at them cutting corners to make a few extra bucks.
If you don't have "Cloud Provider" in your services portfolio, you're like, so totally last century. Nobody provides server hosting or IT services these days. Everyone does cloud, man. The same old IT department at your employer is now a Cloud Provider.
If you have a server in your mom's basement . . . congratulations, you are a cloud provider!
It's all so everyone can claim that they are doing Cloud.
Schroedinger's Brexit: The UK is both in and out of the EU at the same time!
There are two items at play here...
1) Server consolidation - when I was at AMD a few years ago, I saw a series of roadmaps showing the predicted consolidation based on hypervisors 300 servers to 30. The immediate thought that went through my mind is "the cost of enterprise CPUs" need to go up otherwise there will be blood in chip market. Servers were the cash cow for the market.
2) Migration to cloud - this is really consolidation mk II. Move to the cloud and rely on focused efforts to migrate, load balance, spin up and spin down services. All with the economy of scale that large datacenters provide. This has hit the OEM manufacturers (HP, Dell, etc) since the larger players in the market can go direct to China with the volumes they need and
Ultimately it is a question of reducing unused capacity. According to some stats (google "datacenter utilization"), 1st party utilization is around 5-10%, cloud utilization is around 20-30%. The two items above really deliver a 1-2 punch to the Server and Chip industry.
I dunno how paying the admins he was complaining about is not continually paying for installation.
and really, if we're realistic, they weren't running realibility tests on it..
world was created 5 seconds before this post as it is.
No longer be in control of their data?
You don't get it, do you? The NSA doesn't have to obey the data privacy laws inside or outside the USA. Next, it is the very country's intelligence agencies that are providing the take (intelligence collections) that give the NSA their data streams in violation of those same country's privacy laws. And, to make the deal all the sweeter, the NSA provides a nice sweet intelligence package on those country's citizens and the NSA just so happens to get a nice sweet intelligence package on US citizens in direct contravention (my opinion that I used to enforce with weapons) of our Constitution.
And that's if they don't do things messy, like GCHQ did to Belgian telecoms systems.
"[I]t is a wise man who admits the limits of his knowledge or skill, and that pretending either causes harm." --Terry Go
Paying the admin is continuously paying for maintenance that he included, not installation.
When you cant win, ad hominem.
See, if they spy on Americans, they could get in trouble. See, as an intelligence agency, there are limits on what they can do wrt Americans, and if you ahve an American server and an American person of interest, then you have to do a bunch of paper work and go to a secret court and it's just a big pain in the ass.
BUT if you ship everything overseas, then it's fully within plausible deniability in harvesting all of the information from a source controlled by a foreign national. Once it goes off shore, the drag net gets to sift through everything. The NSA's mission is to sift through every scrap of data they can get ahold of. The only people who would want non-American servers are non-Americans, because they have no protections whatsoever. Americans *should* want American servers as there's a whole judicial process involved once everything is under US jurisdiction. That won't stop the NSA from "accidentally" combing though your stuff, but if they screw up even a little bit then a good lawyer can have it all thrown out.
Is it just my observation, or are there way too many stupid people in the world?
I'm so sick of "the cloud the cloud the cloud." Everything is a freaking cloud now. It's stupid marketing horseshit and that's all there is to it. When I'm doing a consultation for a business and they ask me about "storing things in the cloud," the first thing I do is tell them what that word really means.
"The cloud" just means you're putting all of that data on hard drives owned someone else you don't know.
When I change the context this way, businesses suddenly start to think twice. I also like to point out that Dropbox has been found to open your documents for some unknown reason as a recent example to show that you don't know who is going through your stuff when you push it off onto another person's computer. Then I bring up the point that if law enforcement decides it wants to look at your data for whatever reason, you have less control over that because it's stored on someone else's systems and the warrant or subpoena could potentially go to that provider instead of you. Then there's the fun part when a cloud provider makes a mistake and accidentally gives your account to someone else you collaborated with, or deletes your account without a trace or any notice. Don't even start on the NSA end of this mess. Trusting "the cloud" is a stupid idea.
Most companies don't like the idea that when they move their data into "the cloud" when the possible repercussions are put into perspective and the marketing gimmick is stripped away.
You people complain when the stories are old, and you people complain when stories arrive from the future.
There's just no way to satify you people, is there?
the only difference is you are assuming the "cloud" provider has ALREADY DONE THE WORK.
You're not 'assuming' it, you've written it into the contract in the form of SLAs. In most organizations I've worked with, there are rarely SLAs between IT and the departments they support, or, if there is, they are ignored. Not the case with an SLA between a cloud provider and an organization.
I'd like to know what your reasoning is? I think we are simply witnessing the movement of certain base-level IT services into the commodity space. This has happened in many other industries once they become mature. For instance, unless you have some critical, unique, proprietary capability, you probably farm out your manufacturing. Why have capital equipment and specialized employees unless they are going to be utilized 100% of the time? A well-run contract manufacturer will be doing just that. The same thing is happening with IT. Why run in-house email or public-facing services? If the internet is down, you won't be getting email anyway, and no one will be able to see your web page. If you have multiple locations or telecommuting employees, you are already at the mercy of internet speeds and availability. Certain services are mature, and unless you have some specialized need "the cloud" works just as well or better than your in-house solution.
W..w..W - Willy Waterloo washes Warren Wiggins who is washing Waldo Woo.
Just yesterday I turned up a new "cloud" based service per corporate directive.
Holy crap what a mess! Granted, I don't have to deal with any hardware and the project was completed in 5 hours. But, what a convoluted and mind numbingly complex series of integrations and cross connections.
Project: A PHP site/app for mass-mailing and list management through AWS.
I have 9(!) new and unique userIDs and passwords for setup and administration. IDs for service providers, servers, applications...
I have no clue where anything is physically located.
I have no idea how secure it actually is because I have no idea what or how many systems it all uses and how well they are secured.
It will be a nightmare to troubleshoot this beast if it stops working!
There is a lot of opportunity for service interruption due to billing issues at any of the numerous different services. Domain, DNS, Hosting, App subscription, AWS...
I could re-implement this entire mess on a single LAMP server(provided a big enough pipe) in an hour, post OS install.
Do they count your own cloud like my synology server too?
It's one of the best ways to avoid NSA.
Not the case with an SLA between a cloud provider and an organization.
I really don't know where you get that idea from. Cloud SLA's are not worth the paper they aren't written on.
Both technical and political solutions should be taken, developed and adopted if we want to get back our privacy. There are several countries that are in bed with NSA (that will do everything in their hand to prevent people to try to keep their privacy, including the ability to have home servers), and several that not. And if well things could be out of hope in US, england, australia, sweden and some more, in others meaninful actions could be taken.
The point is doing what is within our possiblities. If we know for sure that the data in the US cloud will be inspectioned by the NSA and even passed to potential competitors then is not wise to store things there. We can work in protecting countries or home networks, things that should be under our control, if we are not up to the challenge at least we tried.
Although I don't proclaim to be able to predict what will actually happen in the future. I the past in the computer industry has bounced between server "cloud" centric and client centric for years. There are advantages in having both, In your example of email while it is true you can't get new email while the internet is down you can still read old emails. If the emails where stored only on the server then this would be inconvenient. Also there is a difference your connection to the internet going down and your email cloud provider going down. It is one more point of failure.
Also don't underestimate the value of having control over your data, you do not want to be reliant on some random person/company being up, not go bankrupt, or change its terms and conditions on you. Also people like having the impression of ownership, I think its something inherent in our nature, how many things do you own that you use only use occasionally, that would much be a much better allocation of resources if it was shared?
Why write the SLA on paper when you can store it in the cloud?
"They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety."
Cloud SLA's are not worth the paper they aren't written on.
I work for a SaaS company. Our customers hold us strongly to the SLAs. If your providers aren't, then you need a different provider, better lawyers, or both.
Out of curiosity, does somebody already offer a service where you can put your Windows Shares and even Domain Controller to cloud? Then you would use them transparently and users could also log in to them at home. Is this possible?
Reasoning? The mainframe is a "cloud in a box". We are almost up to 1970. The mainframe died. The "private cloud" (think 1989 Citrix) has seen its growth and decline.
We host, we insource, we host again, and repeat. Rather than the challenge of "why will this one fail like all others before it have" ask the question the other way, "why do we think this one will be permanent, when all others before it failed?"
Learn to love Alaska
There's just no way to satify you people, is there?
I thought they meant In Three Years, Nearly 45% of All Servers Will be Obtained from Bankrupt Cloud Providers.
So if you hold on to your existing autonomous infrastructure today... in three years you will be able to upgrade your server very cheaply!
If we can convince everyone to hold on to their existing autonomous infrastructure starting right now... we won't even have to wait three years! Those sad little cloud service pound puppies will start hitting the market in months.
In light of this I have decided to hold on to my own autonomous infrastructure for one more day. Now it's your turn.
<blink>down the rabbit hole</blink>
Sooner or later, companies will realize that they can save money by taking back ownership of their IT infastructure. It's simply a cycle very much in the tradition of pre-Abrahamic societies that viewed life in general as a never ending cycle.
Corporate beaurocrats need to re-arrange the deck chairs in order to make it look like they are doing something productive. Sooner or later, they will change things even if there isn't any real reason to.
A Pirate and a Puritan look the same on a balance sheet.
No you fool! Forget privacy, there's a bigger danger! If these trends continue, we'll upload the last existing server to the cloud and shut down the server, only to realize that the cloud was on servers! THE INTERNET WILL JUST DISAPPEAR!
GOOD ONE. But it's already too late. The last of the content disappeared years ago. Everything is being served from Squid proxies. If you don't believe me check the Last-Modified time on this page. See how it is, like, this very minute? That means there is a coverup in progress.
<blink>down the rabbit hole</blink>
100% of all servers will ship to companies whose executives have used the "cloud" buzzword to promote the company.
just a ghost in the machine.
The cloud isn't technological. There are a few enabling technologies like virtualisation, but the cloud itsself is a business model. It's just a new, upmarket term for 'outsourcing to a specialist contractor.'
I think people "like the impression of ownership" not simply because it's some sort of quirk of human nature, but because it equals control of what's owned. If you think about it though, when it comes to most things of large value - we don't really own what we say we own. A lender does.
I don't know very many people in the U.S. who own their homes, free and clear. Most people I know with relatively nice cars have a loan on them, too.
So why would we be so eager to make those arrangements? Well, there's still the promise that at the end, when all the payments are complete, it truly becomes yours. And just as importantly, as long as you pay on time, nobody ELSE out there has any say so or ability to borrow/use what you're paying for.
That's my problem with a lot of these cloud based services. They offer a number of benefits, but you give up some control in order to use them. I think some people are so used to payment arrangements as part of a purchase, they feel like they're still in control of what they put in the cloud. "I get my very own unique username and password, and I can log in and do whatever I like with the service at any time as long as I keep making my payments on time!" Problem is, there's no end to those payments when the service becomes "yours". You're just a renter of the service, and the law isn't even very clear as to what the "landlord" is obligated to do with your data if you're evicted from the system.
I'm more cynical: I believe that the vast majority of people couldn't care one bit about internet privacy until it affects them personally and directly. The only people using mesh networking and encrypted p2p are pirates and enthusiastic activists.
That was kind of painful. Worse than this, even.
Better headline: IDC expects current trend to continue, extrapolates linearily despite thousands of years of evidence that few things scale in a linear fashion.
Like all trends in tech, this hype will hit a saturation somewhere and then something else is hot. We've seen this a dozen times before, why do we always look at the newest trend as if we're newborns seing the sun for the first time?
Assorted stuff I do sometimes: Lemuria.org
Our uptime is guaranteed; we are required to adhere to certain maintenance windows. If we don't, customers can (and have) penalized us. It's in the SLA - That's our "service level."
>There are advantages in having both, In your example of email while it is true you can't get new email while the internet is down you can still read old emails
That has more to do with your email client. Not where your service is located.
>Also there is a difference your connection to the internet going down and your email cloud provider going down. It is one more point of failure.
Or one more point of redundancy depending on the design of the mail architecture. Your local email can go down too.
>Also don't underestimate the value of having control over your data,
You can have both, if you use the right software.
> Why pay someone huge fees per month when your staff can easily do whatever it is.
Depends how much staff you have. Paying your staff isn't free.
>gmail simply does not do all that Exchange does.
There is cloud hosted Exchange these days.
>cloud prices - fixed, fixed fixed lol
Do you even know what you're talking about? Cloud prices, such as AWS have been dropping year by year.
Also don't underestimate the value of having control over your data, you do not want to be reliant on some random person/company being up, not go bankrupt, or change its terms and conditions on you.
On the Ts & Cs, you have a point, but for the rest of it, I ran my own mail server for years. My uptime never came close to matching gmail, and I'm far more likely to go bankrupt than Google or Amazon. Gmail's spam filtering is better than anything I ever achieved, too.
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
Look you anonymous coward, if you just want to be confrontational I'd suggest you go take some anger management courses before you get back online. Suffice it to say we have contracts with our customers and if you're unable to negotiate contracts with your providers you need to go find new ones.
Seriously Linux was taking over, then Apple was taking over. I hardly ever see a serious game changing IDC prediction that pans out. Oddly enough back in 2000 and maybe earlier I heard execs from Sun saying the network is the computer. Cloud is just a few companies trying to claim this is a new idea rather than the latest push to let others own your data. To ensure it works they have spent the last 10 plus years convincing everyone that privacy is dead and not required.
The difference between the cloud services we see these days and typical servers from the past is that cloud services are typically designed to scale easily by adding new servers on the fly or servers in multiple locations. Of course plenty of traditional services are jumping on the cloud bandwagon which isn't helping with the confusion.
Nah, they won't really go bankrupt, they'll just use all their servers to mine bitcoins, once they hear that Gartner or IDC expects bitcoin to be at $11.7k in three years.
Does it mean that 45% of all servers will be used to heat up water ? Or maybe something about weather forecasts.
Mainframe? You had to shell out big money to buy a mainframe. Cloud services are rented. In the 70s you could certainly buy time on a mainframe, and a lot of people did. But small companies never had an in-house mainframe as an economically viable option. If your company is so big that your IT guys are all specialized and their time is mostly occupied in their specialty, then the cloud doesn't make any sense. If you are smaller and you don't need a full time Outlook guru and your server sits idling most of the time, then you might very well do well with outsourcing your email.
W..w..W - Willy Waterloo washes Warren Wiggins who is washing Waldo Woo.
A mainframe vs a minicomputer turned the mainframe into a business model. You'd buy CPU cycles or I/O, not unlike today's cloud (other than the Original Cloud often billed on use, not capacity). "outsourcing to a specialist contractor" has been around since IBM was founded, and EDS, and piles of others. "The Cloud" is nothing new, and has been around (without the fancy title) for longer than most of us have been around (longer than that if you include non-computer clouds).
Learn to love Alaska
So it's just like a mainframe, where more CPUs could be added (or connections wired to other mainframes interchangeably), so that capacity increased with no action on or change to the customer. Again, I see nothing that indicates the cloud is any different than a '60s mainframe.
Learn to love Alaska
Why would you buy a mainframe? You bought a 64k leased line and connected to the "cloud" mainframe. You'd give instructions remotely, and the mainframe would calculate and return answers. Same as Citrix. Same as the Cloud. You were only charged for what you used.
The Cloud is a new term to obfuscate the fact is is an old idea.
Learn to love Alaska
Functionally, there are vague similarities to a mainframe (both are used for computing), but there are many differences.
A mainframe is a centralized computer, cloud services are typically distributed among many machines.
A mainframe is designed for reliability, cloud services are designed for easy fallback to an identical machine
A mainframe is made from expensive customized computer parts, cloud services are average servers
A mainframe is limited in how much it can scale, cloud services can just keep adding servers to meet customer demand
Yes, if you were small you leased time. That is exactly what I said in my post. The tasks have changed since the dominance of the mainframe, however, and certain common tasks are now mature. It is possible that email is your competitive advantage, but it is almost as likely that electricity generation is your competitive advantage. That is to say, it probably makes as much sense to run your own email server as it does to run your own generator. Of course there are exceptions - you might be big enough to employ a full time Exchange guru and you might use a significant portion of your server capacity, just like Alcoa uses so much electricity that it owns hydroelectric dams.
W..w..W - Willy Waterloo washes Warren Wiggins who is washing Waldo Woo.
If the emails where stored only on the server then this would be inconvenient.
Why would you do that? Keep using your current client. If you want to jump in and use a webmail client, then there are solutions for keeping that local as well.
I agree that people should keep their data in a form that is easy to recover from the loss of a provider, just as they should keep their data in a form that is easy to recover from the loss of their local data center.
W..w..W - Willy Waterloo washes Warren Wiggins who is washing Waldo Woo.
A mainframe is a centralized computer, cloud services are typically distributed among many machines.
A mainframe is a distributed computer in a single (or small number of) chassis. A cloud typically distributed within a single (or small number) of buildings.
A mainframe is designed for reliability, cloud services are designed for easy fallback to an identical machine
A mainframe is designed for uptime. A cloud is designed for uptime.
A mainframe is made from expensive customized computer parts, cloud services are average servers
Why do you care about the cost of the hardware, when it's the "service" you talk about?
A mainframe is limited in how much it can scale, cloud services can just keep adding servers to meet customer demand
A mainframe can scale to infinity, but in practice is considered limited because people didn't buy infinity, but are buying it now, so we are seeing it now.
Looks like your points of differentiation are driven by ignorance, not logic. They are both remote computing platforms. They both were sold with the same vague wording. They both failed, as services in-sourced. Oh wait, the cloud hasn't failed yet, because it's still undefined marketing speak, and not an actual thing.
Learn to love Alaska
If you abstract it enough, then of course they're going to look the same. A PC is the same as a dumb terminal if you don't care about what's behind the keyboard. Both let you type stuff in and then the screen prints out results. All the points of differences you basically agreed with, then pointed out that if you look at it from farther back then it's the same.
The main point was that "cloud" has been done before and failed. So have terminals. So why are so many people so certain this is the one cycle that will last forever?
Learn to love Alaska
I wouldn't say it failed, I'd say different variations have come and gone being replaced with better ones as technology progressed. This one likely will be replaced by the next advancement in technology.
Yes, "replaced when obsolete" is close enough to "failed" for me. Everything not currently in use has been replaced when obsolete, and everything in use will be replaced, just a question of when. So I'm not sure why so many swear by the Cloud, as if it's the first thing on the planet that will never be obsolete.
Learn to love Alaska
I work for a SaaS company. Our customers hold us strongly to the SLAs
Of course you're going to say that. The reality is very, very, very different. SLA contracts have holes you can drive a truck through in them.
If your providers aren't, then you need a different provider, better lawyers, or both.
Ahhh, yes. I need to spend money on lawyers..........
Whatever. You've been confronted and called out.