|
vaporware.... 
_________________________
Asu no koto o ieba, tenjo de nezumi ga warau.
Talk about things of tomorrow and the mice in the ceiling laugh. (Japanese Proverb)
John Andrew Holmes "It is well to remember that the entire universe, with one trifling exception, is composed of others."
|
|
|
|
|
Hmm, vaporware that you can use today?
|
|
|
|
|
upload virus to cloud and ...wait ;}~ 
peace & serenity
|
|
|
|
|
it`s not so easy to upload a virus on it !!!
but sure it is possible 
|
|
|
|
|
The cloud isn't a take it or leave it proposition for deploying for applications. I will argue that 90% of hosted applications can benefit from incorporating parts onto the cloud.
Common arguements:
-What if XYZ service will go belly up
Build the app to support 2 or 3 different services. If one goes down, simply switch your "cloud" provider. Unless WW3 breaks out (in which case who cares anyway), all three services won't simultaneously cease to exist at the same time. Storage is DIRT cheap on most services, so keeping redundant info on a couple for seemless failover costs almost nothing.
-How can I trust XYZ to host my database? THEY ARE GOING TO STEAL MY BITS!
a) Yeah, right...
b) Don't store your sensative data on the cloud if that is an issue. Nobody said you need to put your entire application onto the cloud. I, for instance, edge cache every image on the sites I host. The applications and databases still run on my servers. This reduced the bandwidth requirements on the server by 90% and it frees up the disks to perform database lookups instead of reading static images. I can serve way more applications from each server this way. Pages are served much faster because the images come from the closest cache after the first hit. My transfer costs are substantially reduced.
-OK...lets assume that all the cloud providers simulaneously explode...then what will you do with your application???
Then I change a single CNAME entry that points to my image repository on my cloud infrastructure provider to point to my own server. I go pay thousands of dollars a month more to setup all the mirrored servers I will need to serve the content, and I'm still way ahead of the game.
I tell my clients up front that the price of my service depends on the availability of X, Y, and Z services. They can either pay an order of magnitude more right away, or they can take an essentially zero risk chance and start pay way less right now. Setting up the more complicated cloud infrastructure produces an ROI of 3 months generally. I have yet to have a client say "No, I'd rather pay 10x more per month for hosting so that I can save a few dollars on the initial development cost. God forbid cloud hosting will cease to exist in 3 months...I would lose money!"
-I like to pay for my software once and own it
Good for you. The question was about you developing something for the cloud, not purchasing something from the cloud. But, let me argue this regardless - I don't know how many of you actually sell software, but almost every piece of software has either annual support charges or upgrade fees for new major releases. How many software programs do you have installed on your PC right now that you will never have to pay an upgrade license for? Guess what, just about every peice of software is a recurring cost.
Regardless of how YOU want to purchase your software, I'd be very surprised if your legit software company (no, that part time hobby with 5 partially completed projects doesn't count) isn't hosting something that can benefit from some offloading to a cloud environment.
---
I'm not even sure why I'm arguing this, trying to enlighten you folk...I'm just reducing my competitive edge against you :P. Everything that's a new fad isn't always a bad thing...I can understand the skepticism, but that doesn't mean you should irrationally run away from something that has the potential to help you greatly. There are plenty of very good use cases for cloud environments. To just completely write it off without considering how it can help you is foolish IMO.
modified on Tuesday, November 4, 2008 3:46 PM
|
|
|
|
|
Some good points, and plenty to think about. Thanks.
I voted no because of the types of apps I write. I write software that monitors and controls our machinery. There is a very predictable amount of data generated each day and stored in a SQL server database, I have no need for elastic scaling to meet demand. Reports are generated on the client PC, and take only a few seconds, so there is no need for big servers. I see no benefit to moving any part of our current app suite to the cloud.
I think the cloud has it's uses for some types of software, but I think the amount of developers just writing client based business apps that have no benefit to gain from using the cloud is underestimated. That is where I think the large portion of no votes come from.
Simon
|
|
|
|
|
So, basically you're saying that we should triple amount of code (in case one provider pulls the plug) so we could use 'cloud' to serve static content and we still have to develop old fashioned server-side and and client-side application to handle sensitive (almost all?) information. Easy and clean...
And for the subscription, your arguments just don't stand. If you bought, let say, VS2005 you don't have to buy release if you don't need/want/like it, or you can skip few releases, save some money and then by new release that you want, simple you continue to develop and deploy your application using older version of tools. But on the cloud you HAVE to pay month after month, year after year to your provider...
|
|
|
|
|
Yes, you should triple the amount of code that handles your cloud storage implementation, which if you properly abstract your classes, is very little.
And yes, thats true...but you WILL eventually upgrade. My point is simply that all software is a recurring cost. A particular cloud application may cost more or may cost less in the long run depending on the particular situation, but thats true of any application. Subscription software isn't something that is new to the cloud - there are many software applications that you have to re-license every year.
Is it possible X desktop software will be cheaper than Y subscription software over the long run? Yes, of course, just like it is possible that X desktop software will be cheaper than Y desktop software...companies can set whatever pricing they want. My point is simply that flat out refusing to use subscription software isn't always smart, since you may end up paying more for "perpetual" software licenses + upgrades in many cases. It's a case-by-case situation. It was simply a response to "I like to own my software instead of paying for it continually, so I will NEVER use subscription software".
|
|
|
|
|
And...depending on the situation...you don't have to write ANY additional code. If you structure the site properly, you can configure edge caching to dynamically cache often requested files. Switching edge cache providers can be as simple as switching a CNAME in your DNS server.
You have to design a client/server application to handle your sensative data anyway...I'm not changing anything there. I'm simply serving static content from a cloud. You make it sound so difficult...it really isn't. The longest I've spent reconfiguring a site so far to work with an edge cache provider is 1 hour. It's a matter of changing image URL's to come from a prefix like images.site.com instead of site.com/images. Nothing else needs to change.
|
|
|
|
|
Mike Marynowski wrote: Build the app to support 2 or 3 different services.
<sarcasm>How come did I not think of that?! Such a simple easy solution! No extra work either!
Mike Marynowski wrote: I, for instance, edge cache every image on the sites I host. The applications and databases still run on my servers.
<sarcasm>Uau!! Is that what the Cloud is?! Amazing! I have been using the Cloud for many years and did not know it! Imagine that!
Mike Marynowski wrote: How many software programs do you have installed on your PC right now that you will never have to pay an upgrade license for?
Guess what, just about every piece of software I use is free as in freedom and cost.
The "Cloud" (who comes up with these buzzwords?) is just another tool, not a miracle solution.
Regards.
|
|
|
|
|
PedroMC wrote: How come did I not think of that?! Such a simple easy solution! No extra work either!
Actually, it really is quite simple...you just abstract the specific provider implementation into an interface, and then implement the various providers. It's a one-time task that a) doesn't take a lot of time and b) is reusable in all your future projects and c) saves you money very quickly.
PedroMC wrote: Uau!! Is that what the Cloud is?! Amazing! I have been using the Cloud for many years and did not know it! Imagine that!
I don't get it ... what's your point?
PedroMC wrote: Guess what, just about every piece of software I use is free as in freedom and cost.
Yes, a few of you "I'll only use free stuff" developers exist, but most of the people on this forum understand what "ROI" is and thus they have at least a few commercial software apps on their computers.
PedroMC wrote: The "Cloud" (who comes up with these buzzwords?) is just another tool, not a miracle solution.
That's my point. You don't have to base your entire application off of the cloud if it doesn't make sense (and it usually doesn't). But used properly, this "tool" can in fact have "miraculous" results on your monthly bill.
|
|
|
|
|
Mike Marynowski wrote: and then implement the various providers
And it will work ... until the provider interfaces change. A significant possibility since some of those services are very recent.
Mike Marynowski wrote: I don't get it ... what's your point?
Get an "old" tool, rename it with a magic buzzword and you get a shiny magical new tool ... *NOT*.
Mike Marynowski wrote: Yes, a few of you "I'll only use free stuff" developers exist
We "free stuff" developers like to have our feet on firm ground and not over a rug that can be pulled under our feet. 
Now seriously, I have used many non free commercial software, most I only give a trial run but a few have been paid and used for some time. In the end I settled on an almost completely free environment, not because of a philosophical view but because it worked best for me. Note that my most frequent target environment is also a free environment (e.g. Linux or *BSD servers and desktops, LAMP) so it was a natural fit.
Mike Marynowski wrote: most of the people on this forum understand what "ROI" is
Have you been talking to my accountant again?
Mike Marynowski wrote: But used properly, this "tool" can in fact have "miraculous" results on your monthly bill.
Now, that is a argument I completely agree and support. Look at the tool and what it can (and can't) do for you and your bottom line and ignore the buzzwords and hype.
Regards.
|
|
|
|
|
PedroMC wrote: And it will work ... until the provider interfaces change. A significant possibility since some of those services are very recent.
Sure, its possible...but because you have a nice set of base libraries you reuse, its easy to update all your software/websites with a single edit. Most providers are well aware of the implications of changing an interface and continue to support the old interface for a long time (i.e. Amazon S3 introduced a new interface a few months back but continues to support the old one indefinitely for existing apps). I have yet to see this happen. And say it does happen - I still rather save $50g on my annual hosting bill and do an 8 hour update to my interfaces once in a while.
PedroMC wrote: Now, that is a argument I completely agree and support. Look at the tool and what it can (and can't) do for you and your bottom line and ignore the buzzwords and hype.
Well, then I'm not even sure what we are arguing...wait...are we arguing? Wasn't that my original point? "Cloud bashing" is just as much a hype right now as "Cloud services." Ignore the hype (and anti-hype hype), and be willing to critically look at what the service can do for you. My point was that most people aren't willing to do this, or they just "assume" anything that falls under the category of "cloud service" is useless. The fact that you have been "using the cloud without knowing it" just further reinforces this point. Hell, I don't even know what you are arguing if you are in fact using cloud services.
|
|
|
|
|
Mike Marynowski wrote: Most providers are well aware of the implications of changing an interface and continue to support the old interface for a long time
It is a very reasonable expectation (one that I did not see!). My concern about API changes was not for the systems I administer (those I can handle, assuming a reasonably timed warning) but for the systems I could potentially provide to my clients that I have no control over and could stop functionning because of a API change (I have seen some very slow moving admins/clients that take months to apply patches I provide).
Mike Marynowski wrote: Well, then I'm not even sure what we are arguing...wait...are we arguing?
Apparently not! We (may) have a different view of this tool's potential (probably because of different needs) but that is it.
Regards.
|
|
|
|
|
If the cloud lives up to it's potential, then I would definitely consider using cloud-based technology for many applications, not all, but many. The problem for established companies is trust. Why should I trust MS to host my applications, or Google or Amazon, or anyone else?
For startups, in many segments of the market, the cloud may provide an option to compete with "the big boys" in that market at a fraction of the cost direct investment would require. There are many services that, if I were a startup, I'd love to outsource to a cloud-based vendor. Payment processing, backups, load balancing, etc. Once that's done, it's not a far stretch to simply host my app on the cloud. This has the potential to allow developers to focus on the development of business-value components and pay less attention to infrastructure.
I would say that much of what the "cloud" claims to offer already existed prior to the "cloud". The difference is that now, not just specific targeted services can take advantage of mass scale-out, now even small apps can take advantage of it.
The current structure, organization and trustworthiness of the "cloud" is highly speculative and suspicous, but it could definitely be a real contendor for future software.
|
|
|
|
|
I think it is hard to answer this right now. Everything (*theories*) sounds very well right now but who knows what we will end up with?!! I would probably just wait and see...
Anyone else agree on that?
- Stop thinking in terms of limitations and start thinking in terms of possibilities -
|
|
|
|
|
Amazon EC2 is not a theory.
|
|
|
|
|
Crikey. Over 60% "won't" or "probably won't."
I'd like to know what percentage of those voters are desktop app. developers and what are web-developers.
|
|
|
|
|
Paul Watson wrote: I'd like to know what percentage of those voters are desktop app. developers and what are web-developers.
I'd like to know what percentage of those voters actually have a clue what a cloud is. I bet most don't, and most people around here won't care much about it until MS promotes it. So, maybe in a year or so that percentage may change on CP.
|
|
|
|
|
Looking inside the cloud we see it's just a term that describes connected, service based applications. I know everyone thinks the world somehow needs to be connected 24-7, but it doesn't and it isn't and isn't likely to be any time soon.
So unless your application expressly requires 24-7 connectivity and your users trust their data to those that store it for them, it is a ridiculous concept that in fact is nothing new. Somehow we live in a time where taking old technology and spinning it off with a new buzz word excites people to the point of their heads asplode. I find it all boring. 
led mike
|
|
|
|
|
led mike wrote: So unless your application expressly requires 24-7 connectivity and your users trust their data to those that store it for them, it is a ridiculous concept that in fact is nothing new
On the contrary, it provides an excellent way for your application to scale *without modification* to your app for process intensive items. You take a traditional clustered server/client scenario and add on more processing power and then what. If the apps don't take full advantage of it then it's effectively useless. The cloud gives you a way to leverage the awesome power of 100 machines (or whatever) *very* easily. It's not just all about connectivity.
Besides, if it was just connectivity there would be no point to it as client/server has been around for a while already. But, just think of how CP could've benefited from a cloud for the slow downs it gets every so often before they decide to upgrade. Not sure how MS will pull off their incarnation of it, but you can bet it'll be much easier to keep everything in sync and clustered OS-wise than to leave it to each piece of software individually (the old way).
Sure, Amazon's "web" services is just a first iteration of the concept of it, but even still it's also about processing power too and scalability too (in such a way people never done before). Anyway, like I said already, once more people jump on the bandwagon, you'll find this concept being adopted on Intranets, etc. It's the new of life in the parallel/quantum way we're heading to.
If anything, a cloud is more like a cluster farm that doesn't show itself. The process doesn't know any better or even hast to like the old way, it just suddenly gets faster with another cpu unit added to it. Amazon just happens to use it for web hosting, but it's the start of the new way of looking at computing and parallel processing a fun sneak peak into the future of software/hardware.
[edit] Of course I just noticed how the marketing hoopla for Azure mentioned MS data centers up front, so perhaps they'll be taking a bogus (IMO) direction with it. Not sure yet. But, I think *having* to use their data centers would take a nice concept and tarnish it. [/edit]
modified on Tuesday, November 4, 2008 6:32 PM
|
|
|
|
|
Jeremy Falcon wrote: On the contrary, it provides an excellent way for your application to scale
What? Scaling is an issue with connected apps only, my post is specifically stating that NOT all apps need to be connected. 
Jeremy Falcon wrote: The cloud gives you a way to leverage the awesome power of 100 machines (or whatever) *very* easily.
No, not easily, you have to be CONNECTED, and that doesn't even account for bandwidth.
Jeremy Falcon wrote: It's not just all about connectivity.
Now that is contrary to your preceding sentence.
Jeremy Falcon wrote: but it's the start of the new way of looking at computing and parallel processing
There's nothing new about it other than the commercial products that you can buy to achieve it rather than developing it yourself.
Jeremy Falcon wrote: a fun sneak peak into the future of software/hardware.
Dude, you might want to cut back on the coolaid.
led mike
|
|
|
|
|
led mike wrote:
What? Scaling is an issue with connected apps only, my post is specifically stating that NOT all apps need to be connected.
I never said it's an issue with connected apps only. I said it provides a nice way to do it. Apples and oranges here.
led mike wrote: No, not easily, you have to be CONNECTED, and that doesn't even account for bandwidth.
Yeah, but MORE MACHINES DON'T MEAN YOU NEED MORE BANDWITH OVER THE INTERNET - JUST THE INTERNAL NETWORK THE CPU CYCLES ARE CONSUMED ON. See I can use caps too.
led mike wrote: There's nothing new about it other than the commercial products that you can buy to achieve it rather than developing it yourself.
Then you don't know anything about how it does its thing. I mentioned it in other posts, so I won't repeat myself.
led mike wrote: Dude, you might want to cut back on the coolaid.
Dude, why must you confuse being connected to the network with scalability? And dude, it's Kool-Aid.
Now, please pull the stick out and try again.
|
|
|
|
|
Jeremy Falcon wrote: why must you confuse being connected to the network with scalability?
In a discussion about cloud computing? Isn't that obvious?
Jeremy Falcon wrote: Then you don't know anything about how it does its thing.
Anyone that thinks there is something new about how the internet works is a fool. I've been developing TCP/IP applications since 1994. I have a decent handle on how it works. Next your going to try to convince me that Push exists. 
Sorry about the misspelling of Koolaid, it's been years since I've seen the word, my hands typed faster than my brain can operate. That frequently happens to me. However don't mistake my unfamiliarity with spelling Koolaid with my familiarity about how the internet works. 
led mike
|
|
|
|
|
I build service oriented applications, but will not be building cloud applications any time soon.
Privacy concerns and concerns about performance, availability and all that stuff are keeping customers from going there.
Also I think the technology involved is not mature enough.
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
