microtodd writes:
In the wake of Cisco's announcement of entering the cloud market, there are several business case analyses that provide insight into whether the cloud is a good thing or not. Of course there are always competing factors between management and IT, which usually boils down to short-term vs long-term cost and financials vs technicals. What do the Soylenters think? Is the cost savings worth the security risks? Are the technical benefits of reliability worth delegating some administrative control?
This discussion has been archived.
No new comments can be posted.
Migration to Cloud? Just marketing buzz or useful?
|
Log In/Create an Account
| Top
| 1000 moderator points
| 25 comments
| Search Discussion
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
(Score: 5, Insightful) by WizardFusion on Tuesday March 25 2014, @11:34AM
Disclaimer: I work for one of the largest IT companies in Europe. This view is my own and not of my employer.
I am sure there are specific cases for using the Cloud, but I don't know any of them. My company is pushing hard for cloud offerings, including working with the governments (which we already do).
For the marketing people it's all "the could can do this..., the cloud will let you do that...", but it's still a bunch of servers somewhere that need to be maintained by someone like me. Patches/upgrades, hardware faults, and all the usual stuff that needs to happen with servers - physical or virtual.
I see it as just moving the problem from one place to another, and paying a huge amount of money for the service.
The next issue we will soon see is what happens when you fall out with your provider.? Of your provider goes bust, gets taken over, or even just raises its prices just because it can.
Brian Madden has a write up about having an exit strategy for a DaaS provider. You should read it.e /2014/02/10/what-are-you-going-to-do-if-your-daas- provider-pulls-the-plug-you-need-an-exit-strategy. aspx [brianmadden.com]
http://www.brianmadden.com/blogs/gabeknuth/archiv
(Score: 5, Informative) by isostatic on Tuesday March 25 2014, @12:44PM
"Cloud" was taken over by marketing droids to mean a hosted service somewhere, be it a plain vm, or incorporating services too like email, or file hosting.
Originally it talked about building your own applications that scaled across data centres, and that you cold add more processing power on an hour by hour basis. If you need lots of processing power on the last day of the month, you court rent it for a few hours rather than keep lots of servers around. Personally I don't have much need for that type of application, but I can see those that do.
(Score: 2, Insightful) by mvar on Tuesday March 25 2014, @12:50PM
I wouldnt go so far regarding "what happens if..." scenarios, usually a service provider doesn't go bust next day. What i'd be more worried would be loss of connectivity to the provider due to some cable issue etc. Most companies that I've worked with that consider "moving to the cloud" haven't even thought of this simple case. For the average enterprise (medium to small), this risk plus the monthly cost for an FO & backup line doesnt justify (yet) the cost savings from moving to the "cloud"
(Score: 1) by aclarke on Tuesday March 25 2014, @01:15PM
What do you have on the "not cloud" side?
- Hosted web applications already on someone else's hardare?
- Physical servers in your own data centre?
- Physical servers in someone else's data centre?
- Virtualized environment on physical servers, or just physical servers?
What do you mean by the cloud?
- PaaS
- straight-up hardware you rent by the hour, old-school AWS style?
- Something else that probably isn't the cloud but your marketing department thinks it is because you don't own it?
You probably know all this, but I want to put it down for other readers. If you are moving from your own physical, non-virtualized environment to the cloud, you have different systems management tasks, and you probably have fewer of them. You still have to answer the question of "what do I do if the network connection disappears", or "what happens if a server goes down", but your answers will be different. Often your answers with the cloud will be easier and/or cheaper.
With a PaaS model, you will also have less or no work regarding patches/upgrades or hardware faults. You will move your application onto their environment, and they will handle all that for you.
(Score: 4, Informative) by GungnirSniper on Tuesday March 25 2014, @11:34AM
My company uses Amazon's cloud for downloads since they have huge capacity and we can overwrite the files if we ever need to do so. Everything else, except Salesforce.com, is on an internal network.
Tips for better submissions to help our site grow. [dev.soylentnews.org]
(Score: 5, Insightful) by The Mighty Buzzard on Tuesday March 25 2014, @11:39AM
123
456
789
(Score: 5, Interesting) by Vanderhoth on Tuesday March 25 2014, @11:39AM
I didn't RTFA, but "Just marketing buzz or useful?" it depends on what you want to do.
"The Cloud" is most definitely a marking term used by managers to sound like they're in the know. "We're going to move our acoustical data the cloud." Basically cloud is just BS for let someone else deal with it.
In some cases that's good. My phone automatically sends pictures to dropbox where they can easily be shared with mine and my wife's parents so they can keep tabs on their granddaughter. It can also be really bad, like my wife taking a nude selfie to send to me that ended up in dropbox, her poor father... Joking aside, the bad part is actually because if the cloud host you're using decides to shutdown, or is shutdown or has a data breach it's like keeping all your eggs in one basket that's precariously perched on top of the empire state building. Someone farts the wrong way and you lose everything.
"Now we know", "And knowing is half the battle". -G.I. Joooooe
(Score: 3, Insightful) by ElderGeek on Tuesday March 25 2014, @02:23PM
(Score: 4, Funny) by MrGuy on Tuesday March 25 2014, @11:40AM
No, really - that's what they call it:
I'm sorry, I can't take anything with that name seriously.
(Score: 0) by Anonymous Coward on Tuesday March 25 2014, @12:28PM
Intercloudz, then?
(Score: 5, Insightful) by mrider on Tuesday March 25 2014, @11:43AM
I wish the damn pundits would just take a chill pill...
As things evolve, we will see a blending of services and methodologies. Some stuff makes sense in the "cloud" (FSM how I hate that term), while others don't. Some things make sense to have a touch interface, others don't. Some things work well on tablets, others don't.
Why does it have to be so black and white?
What is this "slashdot" I keep hearing about?
(Score: 2) by elf on Tuesday March 25 2014, @11:50AM
It all comes down to security / latency and cost.
Is it more cost effective to have a cloud provider host your data?
Do you need to read this data in a performant way and do you need to read it often?
Do you care about the many articles of people being hacked and the chance that you could be next because your provider had a lot of peoples data all in one place. And I guess do you rate your provider secure
To me cloud computing is a buzz word, I think cloud is good for portability (where you want to access the same data from many devices) but apart from this I don't see it as something that will catch on in the larger companies. Cost will make it beneficial for small companies but as a company grows the other conciderations take over.
(Score: 2) by VLM on Tuesday March 25 2014, @12:42PM
"the chance that you could be next"
In public, all employees of all companies are rockstars at the peak of ability.
Reality is of course more depressing in that 50% are below the median.
The managerial advantage of security issues on the cloud is that if you flip a coin and/or actually figure out you're below median, then its safer for your career to jump off the bridge with everyone else in the cloud than to have questions asked, like how a company staffed entirely by top 5% of skill rockstars none the less hired below the median resulting in a security breach.
This will eventually be figured out by the market, where security fools will tend to congregate into cloud services and a selling point will be local hosting of security related stuff rather than cloud hosting of security related services.
Obviously a major competitor, speaking as a guy who worked at cloudy SAAS type places for most of my history, a customer's uptime and data is not worth a penny more than the cost of sales to acquire a replacement customer and/or some kind of NPV of the revenue the customer represents. And thats not very much, I assure you. On the other hand, if you think of a locally hosted mail server, suddenly the CEO can demand the head of everyone in IT, from CFO on down, if its not fixed, at least in theory.
This also applies to things like scalability and uptime and all that, not just security. Cloud will eventually equal incompetence in the eyes of the general public, at which point it'll mostly go away, at least as a fad.
(Score: 3, Insightful) by MrGuy on Tuesday March 25 2014, @11:51AM
Even the case studies referenced are indicative of the problem that there are multiple definitions of what "the cloud" means, and whether an idea is good depends on which one you mean.
Two of the leading definitions of "the cloud" as I've seen them are:
* Distributed managed hosting, where you rent a virtual server from a "cloud hosting provider" and pay for metered CPU, memory, networking, etc. Not that dissimilar from buying managed hosting located in a single physical datacenter - just that it can't get hit by a meteor.
* Software as a service, where rather than hosting and maintaining a running copy of software, you buy it from a vendor, and pay for usage, hits, whatever, so all you do is configure the thing (probably via a web interface). The vendor may in turn be using a distributed hosting scheme, or may own their own datacenters. The vendor now "runs" your app for you.
These are radically not the same concept.
Cisco appears to be going for Option A. They're becoming "yet another Cloud hosting provider," competing with Azure, EC2, etc.
Personally, I find it's hard to get super excited about "yet another entrant into an already crowded space," especially when the article itself seems to imply this is something of a "me, too" move, as opposed to a genuine innovation. ("The whole IT industry is moving to a world where more applications are running in hosted environments, and Cisco doesn't want to be left behind....")
(Score: 5, Insightful) by Buck Feta on Tuesday March 25 2014, @12:10PM
If they want to sell a lot of hammers, they make every problem sound like a nail.
(Score: 5, Insightful) by VLM on Tuesday March 25 2014, @12:34PM
Some historical analogies of "do everything in a vague manner" were ISDN and .net
The likely outcome based on previous fads is we'll end up keeping one feature, and that one feature will likely become the definition or analogy of the name which originally meant a lot of things.
For example ISDN was meant to be a top to bottom network suite. It ended up meaning 2B+D really slow broadband home internet access, and also meaning PRI trunk signalling right before VOIP destroyed traditional TDM trunking. But all the other weirder stuff ISDN was supposed to do never took hold.
Likely, in a decade, we'll still be cloud computing, but rather than meaning everything/nothing it'll mean some small aspect, like cloud computing = shared webhosting billed by bandwidth rather than more traditional billing. Thats all it'll mean at that point. The trick is getting ahead of the curve to see where its going to be.
(Score: 4, Informative) by aclarke on Tuesday March 25 2014, @12:59PM
Just yesterday I completed a major migration from physical servers to the cloud. There were quite a few reasons for this. The company is small and they don't have in-house networking experts. That's not their focus, so why should they divert attention from their main products to handle issues on which they're not experts anyway? On that vein, the servers were running operating systems which were no longer even supported. The servers were in a different country than the development team, which made physical access prohibitively difficult. There was no backup system in place if any of the existing hardware failed or there was any other problem of that sort. Plus, it's hard to tell how much now, but costs are going to be reduced significantly. One estimate was up to 75% in cost reduction; I don't think it's going to be anything that dramatic, but it's going to be noticeable.
One question was about security. In my opinion, the new system is more secure than the old one. Patching servers will be much much easier, as patching the old servers were actually impossible due to a number of issues. Since at least part of the system is on a PaaS model, we don't even have to worry about patching there. Plus, since nobody on the team is dedicated to managing this sort of thing, we are basically in a sense hiring the cloud provider to help look after some of this. Changing systems has introduced new security risks, but overall I believe the product is much more secure than it was.
I generally would say I'm a proponent of starting on the cloud, and then possibly moving to one's own hardware. In this case, the move went in the opposite direction and I'm very confident it was the right move. Many people think "cloud" == "hardware you don't own" and that's it. That's part of the answer, but the answer also includes the ability to scale resources up and down as needed. This just isn't possible with physical hardware, at least not in the sense that I can pay US$0.17 for an hour of server time when I need it.
(Score: 2) by dotdotdot on Tuesday March 25 2014, @12:59PM
From Wikipedia [wikipedia.org]:
(Score: 3, Informative) by trimtab on Tuesday March 25 2014, @01:34PM
Umm. We were all drawing "clouds" on whiteboards in the 80s to represent "networks that had access, but were undefined" at meetings. Some marketing guy must have seen a whiteboard. ;-)
"Cloud" is pure marketing. It allows the customer to define it as "what they want" in their own head. It's purposely vague. In reality, it is just a re-marketing of "client/server computing" with virtual machines.
Sun was doing "the network is the computer" almost 25 years ago. The difference now is a better network and virtual machines.
(Score: 5, Insightful) by everdred on Tuesday March 25 2014, @01:54PM
On its surface, it's a less politically fraught term for "outsourcing"... at least for now.
We don't take no shit from a machine.
(Score: 4, Interesting) by NCommander on Tuesday March 25 2014, @02:46PM
Well, I'm somewhat a fan of having the cloud as a set of hosting VMs/physical machines. SoylentNews's infrastructure would be considerably more difficult if we had to have setup dedicated machines in advance of go live, or when we need a new machine to do something (with the concept of not overloading one machine). We're hosted on Linode, but with a cavat that sooner or later, I expect us to go to dedicated hosts (likely co-lo or something similiar). If we need a new machine, its a few clicks and BOOM, its up.
The important thing about the cloud is you need to make sure you know what you need, and to really make sure what you get is better than what you can build with being cheaper. There's a nice aspect that at least as far as backend infrastructure/networking etc is handled for you. At least right up to the point you try and do something that gets in the way (this happened to us when we tried to connect our off-site backup to our cloud; essentially bridging two clouds. The VPN configuration to make this work was special and its still not great).
Still always moving
(Score: 0) by Anonymous Coward on Tuesday March 25 2014, @03:34PM
That's a stupid direction to go in (cloud-based hosting -> dedicated hardware). You're pretty behind the tech curve NCommander.
(Score: 2, Interesting) by Anonymous Coward on Tuesday March 25 2014, @03:38PM
For almost any business, there is zero business case to spend money on your own infrastructure when you can outsource, if you think like a manager. Migrate to "the cloud", fire all your sysadmins, and give yourself a bonus. Why have your own hardware to repair and upgrade when commodity computing is so cheap? The economy of scale kicks in - a "cloud" data center needs fewer people and had better automation than your homegrown stuff.
The next bloodbath in the computer field will be SQL Server admins. I can see it coming. If you're an SQL Server admin, better find another job. A lot of vertical market packages are built on SQL Server, and as soon as SQL Server instances are easily available in Azure, managers are going to slaughter SQL Server admins like you've never seen and outsource. The only thing stopping them is that cloud providers don't quite have it available. Pre-tuned SQL Server instances for different vertical market packages are coming at some point.
(Score: 0) by Anonymous Coward on Tuesday March 25 2014, @04:30PM
This is just one more example of companies trying to sell shit we don't need.
The PC "revolution" was exciting because we got away from the rooms with the
gurus and the elevated floors and tape drives and keypunch machines.
And now "they" want us to go back to remote storage ? Transmitting over fiber or
cable so it can be intercepted and possibly compromised ?
Not with my data, not now, not ever.
Apple tried to ram iCloud down users' throats by killing off local USB sync of
iOS devices. But the users raised hell, and the next rev of OS X is bringing back
local sync. If Apple had not done this I'd never buy another thing from them,
because I will not be coerced into using a product I don't like.
.
(Score: 1, Funny) by Anonymous Coward on Tuesday March 25 2014, @05:42PM
Yes he can
It is called retromancy
http://tvtropes.org/pmwiki/pmwiki.php/Main/Whateve rmancy [tvtropes.org]