"Any aaSes in your organisation?"
The occasional MCSE, as everywhere else...
Cloud. It's more than renaming the old stuff isn't it? It's a way of working in which you don't really want to know anything about the nuts and bolts. You just want it to work and deliver the service you've signed up for. In other words, it's a black box with a service agreement stuck on it and, with a bit of luck, a bunch of …
While I understand the point of the article - that cloud computing isn't the second coming - I respectfully disagree. I think the fact that El Reg's core readership is in the IT space skews the results significantly.
While I would hardly say I've drunk the Kool-Aid - I don't agree that cloud computing will ever completely displace enterprise apps - I do believe that SaaS has much more potential than most IT folks are willing to give it credit for. Particularly because it can provide a company with tremendous agility to adapt its front-end tools to reflect changes in business processes - changes that heretofore have taken months, if not years, to implement with on-site applications.
With most SaaS platforms, much of this can be done by the business in a matter of weeks, rather than by IT in a matter of months - so I do think that IT should sit up and notice, or else the business will move on. Rather than the typical IT bits being at risk to outsourcing - helpdesk, development, telcomm support, software maintenance - the areas of IT that have typically been secure will be at risk - IT project management, business analysis, etc.
As a former sales guy who, via operations, has come in sideways to the IT world, I believe it's important to get different perspectives on things like this - any operations folks out there what to chime in?
Cloud computing is a strange beast and I'd say it makes a lot more sense if deployed as a consumer product rather than a business architecture, Take the Cherrypal model as an example, a low power (2watts) little black box replacing the conventional home desktop and the user could pick and chose which apps they want in a similar way to the iPhone saving them the cost of an overbloated OS and delivering what they need? That and Amazons Mechanical turk could be a good example of cloud business?
As for cloud working, surely it all comes down to pipe resilience, for example if the infotubes get all clogged up with UTP freetarding and Profile Pharming then playing spider solitare could become a real ball ache!
(Paris because I've not had enough sleep and feel used and dirty)
Totally agree with you, I have embraced the whole SaaS/PaaS/IaaS model as I do think it is the future but only when the "cloud" can truly be connected to back office applications. This gives us the flexibility to offer customers what they need in the most cost effective manner.
There are a number of problems with this:
1. Broadband in this country is, well, crap! ADSL2+ isn't very good and Be* offering bonded 24Meg and saying its 48 is just ridiculous. I'm 3 miles from the exchange and have 24Meg. I sync at 6.
2. People are used to LAN speed connectivity to Email and files and to a certain extent some apps. Until point 1 is address this will never change. Imagine running sage in a cloud!?
3. So we end up at primarily web based technologies, basically what we already have but with the "Google" model attached to it.
I think one of the big problems in systems architecture is data storage and retrieval. Legacy apps are difficult to change because they have been designed from the database up. If your database schema needs changed, or the whole model requires re-normalisation, your legacy app is finished, time to start again.
Then again, with M$ we have SQL services in the cloud. Is this not just the same thing, but on a bigger scale and when things change its a whole new bigger problem to deal with.
The only way the "Cloud" will fly is if we embrace the Google model, flat and scalable is king, Native XML storage is the future and someone needs to realise it soon!
We are a small company providing a brand/sales collateral storage and distribution service to large companies. Our biggest problem is storage management. Storage as a service provides us with several benefits. Firstly we don't have to make large up front investments in storage infrastructure. Second we can scale the storage up and down instantly. Thirdly we can make use of the storage provider's caching capabilities to hold content near the consumer.
There is simply no other way to achieve our goals. Our business would not be viable if we had to build up a global distributed storage infrastructure ourselves.
What about reliability? Most of the storage providers have an SLA, but these don't amount to much. We therefore use several storage providers and data is replicated across them. Our software applications will try multiple locations in rotation when retrieving a given file. This mitigates the risk of a single storage provider being unavailable.
One way or another, software and infrastructure as a service will transform IT over the next few years. There are ways of retaining control and the cost benefits are too huge to ignore. Those companies that don't make at least some use of the technology will suffer or fail.
"There is simply no other way to achieve our goals. Our business would not be viable if we had to build up a global distributed storage infrastructure ourselves."
Do you honestly believe this gibberish? Are you in management or marketing?
*I*, personally, have designed several global distributed storage infrastructures, with built-in redundancy and all of the other bells and whistles for reliability. It ain't rocket science. But then I'm not trying to sell the fortune-150 companies that use it on outsourcing ...
Of course it isn't rocket science. Technically it would be no problem at all.
My point is about capital outlay. Do it yourself and you've got some significant up front costs. By using storage providers, our storage costs become a revenue item i.e. a direct cost of sale as opposed to a large fixed overhead.
As a small business, particularly in times when loans are hard to come by, we could not afford the capital outlay required to build our own global infrastructure, therefore without the service providers our business would not be viable.
I tried replying to you last night, but apparently I upset a moderator. Few of the posts I made last night were allowed ...
ANYway, the gist of mine was that basically, you are selling a third-party solution where even a second party is not needed ... which I might try with a couple of my more clueless clients. Why the heck not, if they are daft enough to give me money for no work on my part ... especially seeing as these particular folks are a pain to deal with ...
In a nutshell, with disk & RAM and CPU as cheap as they are, and leveraging the Internet and next-to-free bandwidth along with better-than-good freely available operating systems and encryption, why would any global company need a second party for a distributed storage system? Much less a third party?
They already have the distributed locations, they already have the bandwidth, just add cheap disk & RAM and CPU as needed to the geographically diverse locations. Throw in any one of a number of solutions to keep stuff synced up (I'm thinking N+1 redundancy here, no need for more in most cases), and you've got a one-time cost of hardware+installation, and the monthly electricity and bandwidth and admin costs that you are already paying.
Keep it in-house. No added security problems. Fewer reliability issues. Cheaper.
Works for my clients. Your mileage may vary, etc.
 Nobody, and I mean NOBODY, uses all the bandwidth available to them.
 If you're paying per-byte instead of per-pipe, renegotiate the tariff ...
 Well ... Maybe teenage boys downloading porn ...
Biting the hand that feeds IT © 1998–2021