And both think it is time for you to build or buy a cloud
I thought the whole point of Cloud Computing was that you rented space on someone else'?
So the CEO is hearing all about clouds now and the financial director is looking at his pile of beans and as usual wants you to do more with less. And both think it is time for you to build or buy a cloud. Where do you start? The answer is by being brutally honest with yourself and your bosses about everything around you. A …
Maybe, maybe not. Those selling computing-as-a-service want you to rent their facilities and, ideally, become locked-in to them through API issues and/or the difficulties of moving TB/PB of data from one service to another.
They have the advantage of easy expansion and the press of a credit card, and you should not have to worry about general maintenance. You really should consider your backup/migration strategy, and of course who can access your data in some far off land if that gov demands it.
Building your own should give you the advantage of data sovereignty and little or no lock-in (beware of licences etc that allow the vendor to turn the screws with no easy alternative) and no problems of ISP bandwidth for data-hungry jobs.
Of course, you still need your own hardware, UPS, admin staff, etc, but all of the should cost less due to it being easier to manage and better utilised.
I've always taken "cloud computing" to be more akin to a thin-client paradigm, where something remote to the user does the heavy lifting.
Funny, when I started at Uni in 1984, series PC computing was a nascent science, and 90% of work was done via terminals which connected to a mainframe (even the BBC micros were used as VT10x emulators ;) ).
Then everyone had their own PC and ran apps locally.
Now they use their PC to connect to the cloud and a remote server runs the app.
Plus ca change ...
I guess I mean "virtualised infrastructure" where you have a medium-large infrastructure with redundancy, etc, you would expect that to have built in, but critically the ability to carve out smaller chunks that are mostly OS-agnostic (maybe even already with an OS) to run each of your applications on.
However, it is an open question whether that should be a small part of some 3rd party's (like Amazon) or your own hardware.
Of course, that is very much like the old mainframe model where you bought time+resources as needed, and not yet another server to be looked after each time you needed something new/bigger/without stupid OS/version interactions.
Indeed. Or, as I first wrote here: http://forums.theregister.co.uk/forum/containing/1452371
I have a double-handful of clients with usable corporate computer systems spread across most of the continents on this dampish, muddy rock that we call "The Earth". Several are Fortune 500s.
Not a single one of my solutions contains the word "cloud".
"Distributed", yes. "Cluster", yes. "Grid", yes (one legacy system). "Centralized", yes. Even "Peer-peer" and "stand alone workstation with network availability on demand". Etc. But no "cloud", not anywhere.
IMO, "Cloud" is a reference to the old ISO OSI-model textbooks that used a "cloud" image to try to hide the actual networking layers below the so-called "presentation layer".
Enough non-technically-inclined managers "took a course" with outdated textbooks that this "cloud" nonsense entered the corporate vernacular when said managers moved into marketing (hint: If your school is teaching the OSI model, you're obsolete before you paid the course fee).
The first trouble with your scenario is that the "OSI model" only has 7 layers. The second is that the layers start at level 1, with the physical layer ("wire's wire" as one of my old mentors used to say). The third is I was referencing an actual textbook I was "taught" out of called something like "Networking for Management" that put everything below Layer 6 ("Presentation") in the "cloud". As an already degreed Network Engineer, working on my MBA at the time, I constantly corrected the obviously clueless idiot "teaching" the class, until the Dean called me aside & asked me to grit my teeth & stop correcting "one of my best Professors".
The school? Stanford University.
'IMO, "Cloud" is a reference to the old ISO OSI-model textbooks that used a "cloud" image to try to hide the actual networking layers below the so-called "presentation layer".'
We always used a cloud-like squiggle in systems diagrams to represent 'the internet' and / or 'internal infrastructure that is necessary, exists already, but needn't be specified and could be changed without affecting the system we're interested in' - which is quite close to what people mean by The Cloud now.
I don't have a problem with it being called the cloud and if I was offering a solution that merited it I'd happily use the term - once you've given the customer what they need, there's no harm in giving them what they want as well.
It sure is an attractive idea to outsource menial daily maintenance and overall worry to Cloud providers. But I rarely glimpse any insight into the impact of latency? Can someone offer some real-world enlightenment? Do all companies take at least one critical app, port it to the cloud, and then do benchmarking before migrating?
and asked if they were considering cloud technologies. The tech lead who was interviewing me said they were prevented from using it as there was a risk data could be housed overseas, in different jurisdictions from Australia. I attempted to explain the use of Private Cloud, where you could own the hardware and therefore control the location, and use the new cloud technologies internally. After the second failed attempt to explain this, I realised I did not want the job that badly to work for someone that thick.
I don't see how a "private cloud" helps simplify things. If you own the hardware and control the location isn't that just a corporate network running cloud software. It's one more thing to manage on top of the other technologies outlined at the beginning of the article. I'm really sick of the word cloud, outsource is better.
A big part of this is the "trust" issue which is very dependent on the temperament of whoever makes the decisions. Do shops that are as mixed as mentioned at the beginning of the article outsource the hosting of their website? If they do, they are already "in the cloud" as far as I'm concerned and everything is fair game for sticking stuff on hardware other than one's own which is all we are really talking about. Rackspace for storage, hey your files are already in the cloud.
The only question is how much does one want to offload. can it be done and how much does it cost.
And how paranoid are you.
is the logical development of delivering functionality that deals with information, in the same way "logistics" is the functionality of transporting things from A to B.
Someone in business wants to send something from A to B, they drop it into "logistics" and it happens. What is now happening is someone in business wants to do something with data, they drop it into "the cloud". In both cases the neither know, nor care, *how* it happens. I doubt many logistics managers worry about what tyres their trucks use, or whether they own or lease the trucks, or whether they actually manage the trucks at all, but retain a 3rd party (DHL, etc) to provide the service.
This is part of the reason behind the Windows8 situation ... unless it does something Windows 7 can't, then it's meaningless at the corporate level. Jane needs to crunch some figures - all things being equal* whether she does that locally, or in a cloud app is immaterial.
*Of course this is the question ;)