I'm having real difficulty figuring out whether this is the next big thing or the next big flash in the pan. I'm also unclear on the (nebulous you might say...) concept, papers on the subject don't really carify it for me.
My questions are specifically:
Software as Service: How is this really different from a well designed SOA archtecture over, say, WCF, other than where the code runs
Who owns the code?
Who writes the code?
Where does your code reside and how do you test it? How "transportable" is the code?
Are you locked into a vendor? If so, the dangers this raises surely outweigh than any efficiency gains you might find.
How are security concerns addressed?
Is this really different than outsourcing a company's hardware function (other than the frameworks involved?)
# Software as Service: How is this really different from a well designed SOA archtecture over, say, WCF, other than where the code runs
# Who owns the code?
# Who writes the code?
# Where does your code reside and how do you test it? How "transportable" is the code?
# Are you locked into a vendor? If so, the dangers this raises surely outweigh than any efficiency gains you might find.
# How are security concerns addressed?
# Is this really different than outsourcing a company's hardware function (other than the frameworks involved?)
I don't think it is.
On Microsoft Servers as I understand it. Published I believe via VS2010.
Not very well last time I read about it. They're addressing it, but it still needs work.
I don't think it is. The idea is that Microsoft takes care of all that stuff for you. You buy the level of service you need and the number of instances you want. It gets very contractual. Think their licensing agreements.
I've stayed away from it for the time being because I can't see any sizable company wanting to allow Microsoft to have control and hostage taking power over their data and livelihood. It sounds like a real downer.
That's called seagull management (or sometimes pigeon management)... Fly in, flap your arms and squawk a lot, crap all over everything and fly out again... by _Damian S_
Thanks for the reply, after one week, an answer! This isn't meant as a criticism of you, it just screams volumes about the amount of work going on in the cloud I think.
People keep screaming "next big thing" on the cloud, but when I look at it, the security concerns and the "ransom effect" will prevent business such as banks (who tend to be the drivers on such things) from ever using it. I hadn't considered Miscrosoft's Byzantine pricing structure, even for something as "simple" as VS2010 we had to get an external consultancy in to find the best deal for our company.Their cloud pricing must be horrific. Some people are really fired up about this, and I've been wrong about technologies in the past (Windows devs are resisting WPF, which I think far superior to traditional Winforms), but I think this is a dud, at least in the medium term. My post was really a last-gasp should I take a look at this tech, or should I do something else question.
Cloud Computing might have a use for SMEs though: the security provided will probably be better than can be acheived in-house,but I suspect Google is a better option.
I was having a beer with one of the MS technical reps for cloud computing in SE Asia a couple of months ago, and asked him this exact question. And you are right the primary target is the SME exactly b/c of the security that MS can bring to the service.
Apparently they have some extraordinary service agreements designed to reassure SME's their data is safe and remains theirs. He did admit that the big end of town were not prospects and banks won't touch it, they have enough trouble controlling their (our) environment without more vapourware thank you.
There you go - 2 replies.
Never underestimate the power of human stupidity
I am a student doing masters in computer science engineering.I wish to develop a Multitenant data analytical application in java.Could any one suggest me the procedure to develop a simple multitenant application.. Or else please suggest me any useful links to proceed further.
I have decided to develop the application using Netbeans..
I wish to develop a Multitenant data analytical application in java.Could any one suggest me the procedure to develop a simple multitenant application.
If you do not know the procedure ho do you expect to develop it? I would suggest you try something more suited to your skills, or do some research into the subject you wish to work on. Google is always a good place to start, or possibly your University library.
Just say 'NO' to evaluated arguments for diadic functions! Ash
I find it disturbing that you feel the need to specify the language here. I get the feeling that you have chosen the language first, and the problem second. This is not a good way to do things; I appreciate that you may well be jazzed up about Java, and fairplay to you but you really need to stop and consider what the problem is that you are trying to sort out first.
The first thing that you need to do is to gather your requirements. Work out what you are trying to do; what problem you expect to solve. Talk to people who are doing the task that you are trying to do and find out how they do it. Where possible, you should always try and model your software after the process it is imitating.
Finally, good luck. I look forward to you passing your course, and entering the world of professional development.
I have CDO, it's OCD with the letters in the right order; just as they ruddy well should be
Cloud computing is where the program that an end user wants to run is kept on a company's server. Steam is cloud enabled. For example COD6 had the ability to dl it from steam if you bought it. You can then play the game just as if you had it on a disk. I suppose its kinda like that. Maybe you could define it as a web page too tho...
Well I was drinking with a guy who works for MS here in Singapore, job is to sell cloud/azure services. As I work for a bank he didn't bother pushing it, banks are just not potential customers.
His line went something like, we set up a very secure flexible platform supplying all the dev and office tools you need on a subscription basis. You (as a SME, his target market) cannot afford to implement and support the same level of security MS supplies. It was a good line and he believed it so I didn't have the heart to try and pick it apart (besides I was well into my cups at that stage).
Like most tools it has it's application, it is not the panacea the evangelists claim but has it's uses I guess.
My problem, beside the connectivity nightmare, is that everything is via a browser. I know it's getting better but it is still a pretty crappy interface (this from the guy dragging himself kicking and screaming into Silverlight).
Never underestimate the power of human stupidity
No it isn't. Cloud computing is about having access to resources (compute power, memory, perstistent storage) on an ad-hoc basis, with no practical limits to those resources, at a lower cost than you could supply yourself. It has almost nothing to do with web browsers.
Mycroft mentioned some of the tools built on the platform, but from a development standpoint the way I look at it, cloud computing is a suite of development tools built around making scaling an application out the primary optimization.
That's why many of the cloud platforms have things like NOSQL databases and are web based. Relational databases with strong transactional support are notoriously difficult to distribute so there are many different new models of data storage each of which optimizes a different aspect of the whole scale out problem. There are also different constraints on code management and optimization when it is designed to run as efficiently on one system or one thousand. There are also tools that run horribly on anything under 10 systems but awesome on large clusters.
To further muddy things, Amazon, google and microsoft are selling cloud services based on their particular platforms. But if you have a bunch of systems lying around you can set up your own cloud stack. Some of the systems management for the large installations are mind boggling, with setups of cargo containers full of hardware. They leave the container plugged in and wait for a percentage of the systems to fail in the container and just send the whole thing back to the factory to be rebuilt.
It's a server/client relationship where you are the client, and:
1. You don't own the server or anything installed on it.
2. You have no control over changes to applications installed on the server.
3. The server might be a zillion miles away, and could be routed through a zillion miles of inferior wire or severe bottlenecks at any given time, giving you all the performance of a 386.
4. You can be disconnected from the server and/or services at a moment's notice, for any reason (network problems, billing problems, caffeinated keyboard at the server site problems, the-guy-you-complained-to-on-the-phone-doesn't-like-you problems, etc.) (And don't forget the ubiquitous "computer error" problems.)
5. Your work is accessible to a bunch of people who probably couldn't get jobs as developers, or who barely know what a computer is -- and their secretaries.
6. Anyone with a packet sniffer knows exactly where to put it to access your work, as well as the work of everyone else who uses the same server.
7. Most of your processing power is spent on maintaining contact with the server, up and downloading all manner of unnecessary files, background-installing whatever the server owner decides he wants installed on your local machine, and running background processes for licensing agents, update agents, and secret agents (And that's if you only use a word-processor. If you want spreadsheets as well...)
8. You pay and pay and pay for the above privileges, because the server owners obviously deserve money for nothing.
I wanna be a eunuchs developer! Pass me a bread knife!
Hey, I do enjoy reading your description of cloud computing!
Does anybody remmember the grand idea of "Network Computer"? At that time it was said
that a PC should have all its OS and SW on the network, and it needs neither OS no HD
as network will prvide everything. Pleanty of noise and talks were made about it, with
greatest expections from many. And the result? Total bullshit + nonsense. After a couple
of years, we are all still using PC WITH a OS, and with one or multiple HD.
The present claims about Clouding seems to be repeating that rediculous story. Where
is the common sense? Fortunately no everybody has been carried away by the nonsense!
I have found that the only people excited about Cloud computing is the people trying to sell it (MS, Amazon, etc) and executives / high level managers that are (A) too cheap to buy the necessary hardware themselves, and (B) too dull to understand cloud's limitations. Great examples of the lack of understanding at the executive level: My former CTO drew a network diagram on my desk that had our WAN Routers connecting to the "Cloud" when he truly meant internet / IP connections. This was at a mid-size company making financial software. Now my new CIO is too cheap to buy a single server and is dooming my project to failure by placing it on a Cloud. I say dooming because the network speed, security, and placement of my requested server is extremely important and the Cloud won't provide this.
On a side note, has anyone noticed the Azure commercials from Microsoft that are marketing to the general public? A housewife using the "Cloud" to edit her family photo's. How is the "Cloud" service relevant in this situation, isn't her PC running Windows 7 enough?
Last Visit: 31-Dec-99 18:00 Last Update: 6-Jul-22 1:17