Utility computing's 'dirty little secret' Cloud computing has its very own catch-22. If we can tap into the cloud, grabbing our compute resources on the fly, we can free ourselves from the old school software licensing models. But old school software licensing models may prevent us from tapping into the cloud. "The dirty little secret of the utility computing business …
Yes, lets change all the licensing arrangements, level the playing field, and watch everyone in the IT world go job hunting.
Open source is popular largely due to its reduced costs over traditional software. However traditional software companies make all the money they pay salaries with (and set the standards each and every one of you use to determine fair pay) because of their licensing revenues. If you eliminate the revenue differences then there is no good reason for IT employees to bitch that they are special and want a cut of company earnings - since the company won't be earning much of anything, or will at least be reduced by orders of magnitude.
Cut out the traditional licensing, reduce tech revenues by 1000%'s of percent, and stop whining about low pay in the tech sector (which is hugely above normal rates)
It'll make everyone happier, including me. Yeah for change.
Please stop with the outdated notions. Charge customers for the service not for the software, it's really really simple and if you don't get it that's because you're stupid not because the open source model doesn't work. As a newbie you probably haven't been in the business long enough to know how it works but the price always tends toward zero.
"In other words, all those companies building all those clouds can't rely on the Microsofts of the world for their back-end software"
This implies that anyone CAN rely on the Microsofts of the world ; )
Seriously, though, is this actually news to anyone? The big cloud computing pushers, like Amazon and Google, are quite happy to admit that their systems are open-source-based, not Microsoft systems. Some even consider it a selling point.
Utility Computing's dirty little secret: it doesn't actually have one.
I'm not against open source - I was simply pointing out that if traditional software companies (MS, Oracle, etc...) change their licensing models to accommodate "cloud" computing, then it must have a negative effect on the overall salary in the IT industry.
By default, these companies must sell more licenses for less money, or fewer licenses for more money in order to maintain their current level of revenue. This article points out how this will crush "cloud" computing under its own weight. If these companies make the changes necessary to meet the needs of
"cloud" computing, the resulting drop in revenue will reduce the average salary paid to IT staff because the "big guys" that pay the major salaries (that skew the averages) will no longer be in play. Ultimately this will result in smaller salaries for IT employees because the major players don't have the money to spend on the huge salaries anymore.
I think the posters above misunderstood me - I'm all for changes in the traditional licensing structure. I don't think it will kill the tech sector - I'm positive it will ultimately enhance my bottom line because staff will be cheaper.
It's not optional guys - the average IT salary must decrease if the massive profits from the existing licensing structure are not there.
I've been a senior decision maker in the tech sector for more than a decade. I've seen so many fads come and go it's not funny. My job isn't focused on techie drivel - my job is to make sure the company makes enough money to pay the techies, so that they can drivel. So far I'm doing a pretty good job...
At this point open source is still a fad. Sure there might be some uptake, like with any fad, but it's going to take more than a dozen or so success stories to change decades of industry maturation and business operations.
You know, one the main problems in the IT world is not salary levels, but 'senior decision makers' who not only don't understand the technology, calling it 'techie drivel', but apparently don't understand business either.
Software makers like Microsoft are not the main employers of IT staff - that would be users of IT, not producers of it.
IT licence fees to a company are simply a cost, which, because the applications delivered are packaged, provides no competitive edge. There is no reason to suppose that if that cost was removed, businesses would spend less on IT, thus reducing IT salaries. The opposite is more likely - that those funds would be diverted to gaining competitive advantage using IT more effectively. The salaries of people delivering that competitiveness would be more likely to be increased, not reduced. As long as the 'senior decision makers' knew what they were doing, of course.
Using Open Source in businesses as the basis for tailored systems is likely to increase average salaries, as IT staff are then increasing the market competitiveness of the company, rather than providing a utility service. Currently, software licence fees are a drag on that competitiveness.
The only fad likely to pass away soon is putting people who don't understand the technology in charge of making decisions about it.
though opensource is not going to be the be all and end all here.
If a company creates let's say an email client on the cloud, they don't have to opensource and why should they?
You can write a closed application and run it on a system with a Linux kernel.
You can even use GCC and some of the LGPL libraries and not open source it again (though alter a LGPL library and you have to release that alteration).
All software is about service, the service to entertain, inform, make tasks simple, or even just possible. To maximize your profit, you need to protect your service niche by making it harder than someone just copying your code and jumping on in, then they can offer the same service but not pay for the development costs, so they could probably undercut you.
And have a look at most opensource stuff, most consumers don't want to use it, it is not designed for them, it is designed for developers. Sure, a few developers get lost in the idea of opensource as some type of revolution, but that pretty much stops when the consumers who are paying nothing, start coming in and demanding x and y feature :)
Opensource, works because if a developer wants a feature he/she can roll up their sleeves and include it. Not many consumers are willing to do that.
The GPL is even more confusing to most consumers, but really that is just about a little convention dance amongst developers, companies who don't understand the GPL and use GPL code often get stung, though of course they are often just asked to comply. But failure to comply stops them using the software, so can be quite expensive.
Though, to drive cloud computing you need a solid base, and that will be unix, and quite right you won't want to pay licensing on the base so you will look to make that opensource. And that is the right thing really, consumer applications closed in the main, base infrastructure open, and development tools open. Someone has to pay the ferryman and that is the consumer.
Which Microsoft applications does Mr. Wienman imagine would be useful in a "cload" context.
You are not going to send your word document or spreadsheet out to the cloud. SQLServer database possibly but it's really a "one big box" database and wont spread over '000s of servers like th google offering,
In a competative low margin business like cloud computing ( if its not comptetive whats the point? and competative implies low margins for the competitors) you do not want to throw money for nothing to Microsoft or Oracle so its open source all the way for the "cloud".
The money will be made in "Salesforce.com" type operations that provide a service which businesses need cheaper and better than can be done inhouse.
I have to disagree with you there. If people want to use MS software back ends, or intend to use MS software back ends then MS will simply change the license arrangements accordingly.
The current MS licensing paradigm is generally aimed at specific use licenses, by which I mean you install one license on one PC at the front end and a comparative number of licenses on the back end.
However with dynamic cloudy stuff you cannot just do that. So MS and cloud providers will have to re-negotiate the licensing. If MS are unreasonable then people will go FOSS instead. I will bet MS will be reasonable and get some revenue rather than be stubborn and lose out big time. MS are many things, but you have to agree they are definitely businessly astute.
So you either get effectively cheaper MS licenses or free(ish) FOSS licenses in general use.
Whereas this may reduce profits to MS / Oracle / Sage or whoever it won't affect standard IT personnel *in any additional way to the effects of cloud computing itself*. You will still need the desktop infrastructure, LAN, and WAN and you will still need people to explain things to users.
Now, will a move towards cloud computing itself affect salaries? That is a much more interesting question. Personally I believe any move towards cloud computing or whatever the next big thing is will play out the same way as every change I have seen in the last 15 years has played out - priorities change and people adapt. Salaries generally remain unaffected.
IMHO, obviously
I need cloud computing for all it's benefits... not having to manage a large amount of hardware and network infrastructure would be fantastic. However, what I need to run on all this kit is costly video streaming servers, so I have to buy the licences for all the servers that I _may_ user, in case I ever user them. Each instance with it's own licence key and whatnot.
Now, this is no different from what I'm currently doing, don't get me wrong. However the migration over to cloud computing (or CaaS or whatever) isn't going to be beneficial to me at the moment, as I sit here maintaining existing hardware and software licences.
The first utility computing company that comes along that adds requested software licence costs into the hourly running of an instance is onto a winner.
What are the Slashdot economists doing here?
FYI, SaaS enables GPL code to be embraced and extended in-house without showing customers the source. It's therefore quite ironic that the "free source FTW" comments are coming on a thread about SaaS, because SaaS is the worst thing ever to happen to GPL-based OSS.
Oh yeah, GPL3, that'll save you. (Except there is no useful code yet which is GPL 3-only, but there's a mountain of GPL 2 code ripe for the picking, and GPL 2 cannot be revoked. Oops. How delightful to see Stallman hoist by his own petard.)
(Disclaimer: I'm a GPL hater.)
"At this point open source is still a fad. Sure there might be some uptake, like with any fad, but it's going to take more than a dozen or so success stories to change decades of industry maturation and business operations."
Having been a senior decision maker for over a decade cuts little ice with those who do understand technology. Proprietary licensed software packages may still be dominant on standalone desktop applications to an extent, (does anyone still care about these ?) but this software methodology never went far into the wide area network. As a concept, cloud computing is clearly more closely aligned with the WAN than the standalone desktop.
As to open source being a fad, the reason the Internet protocols were more successful in the nineteen eighties than the then more heavily invested-in competing and proprietary OSI stack was because the IETF specifications and most of the implementations were always freely distributable and anyone interested could participate, i.e. the core Internet is open source. To get copies of the OSI specs without even any software implementing these, someone in my shoes would have had to get a budget sign off from a senior decision maker (who wouldn't have understand why we needed these) for the proprietary OSI specs printed and bound from a single monopoly printer in Switzerland, which were always many months out of date. It was always going to be easier to download the RFCs with a modem and the open source implementations of them which we could contribute towards without having senior-decision makers obstruct the development process.
For someone who was in telecoms in the seventies and eighties and since then became more concerned with the Net, open systems became dominant in telecoms in the eighties to be replaced by open source in the nineties and later, with the same pattern repeating itself shifted a few years later with the development of the Internet/ISP business. I don't hear many of the software engineers who are taking salaries in either the telecom or Internet sector worrying about whether they will be paid from a sale of software packages model that never existed within this industry. There's no shortage of punters willing to pay monthly fees for phone and Internet services provided through open source though. You might as well try to argue that telecom engineers earn less because they are open rather than secretive about standards which determine how equipment from different companies communicate, as if one single company could own the entire world's communications.
So the natural development of cloud computing using open source doesn't threaten the salaries of programmers and technologists in the slightest. You very clearly just don't get the role open source in developing network infrastructure, something which I have rather more than a single decade's experience in. But open source, from this perspective, does make the role of senior decision makers who don't understand how to enable technologists in their employ to be technology leaders somewhat redundant to the process of developing network infrastructure doesn't it ?
"SaaS enables GPL code to be embraced and extended in-house without showing customers the source"
Which is a problem exactly why?
"Oh yeah, GPL3, that'll save you."
The question arises why we need saving, and from what. Probably from a particularly virulent form of sourcofascism.
Black helicopters because the comments in here are probably steganographic messages from the übercloud.
it is about the developers, and even then it is actually about the people who cut the code and integrate the systems.
No one really cares about the desktop, there is hardly any money to be made in that arena, it is has gone web and it ain't going back.
The browsers will really start to improve and very soon, most users will use just use browser based applications.
Tech will always use their own local systems, with local apps, but users, well, they will use what they can get to work and is supported.
All of software is about service it is just automated service, and it is the service to inform, entertain, simplify and make possible.
Only coders appreciate software in and of itself, and they have to, to make good automated services. Show source code to a user and they run a mile. Show it to a coder and they comment on it. I have even seen heads of IT companies refuse to look at source code, bizarre but there you go. And open source is just that, you still have to comply with the license and know how to interface with it. You also need to realise when not to use it, as it will effect certain requirements and potentially increase development life cycle.
You want the languages you use to be open source, the base parts of the operating system and networking open, but perhaps not GPL (which is a spanner in the open source world when you often want the wrench of BSD, MIT licensed code).
The applications most developers use, a user would not have the foggiest about operating those will be opensource and GPL, because the user base are developers who can contribute back.
So, a couple of "experts" saying that the model is not viable because MS can't make money out of it? (I'm not making it up, it's in the article: "Lucas said. "I had a long conversation with Microsoft last week on licensing. I don't think they'll go on record as saying their [current licensing model] doesn't work [in the cloud], but it doesn't work.")
I'm not into this whole cloud thing. Not at all. I work at a quite modest scale, and penny-pinching is less of an issue than keeping control over the data in my field. But the people cited in the article object to "cloudness" for all the wrong reasons. I mean, it's already "standard procedure" in the server world to charge fees on a per-core basis, and most parts of the "cloud" have been applying the same principle for quite a while now. Annual per-user fees, or per-hour charging, or a combination of both. Plus extra fees for special services if applicable. You can even modify the charging grid depending on storage and/or bandwith use. All the tools are readily available, any sysadmin uses them dayly. Why the heck wouldn't it work?
Now I suppose that MS will have a bit of trouble to adapt, but they're known for that. Should we really evaluate a model by looking if MS can make money out of it?
Disclaimer: I just hate the idea of cloud computing, but I believe I hate MS more! ;)
All this licensing ink cloud is really funny, too. Who bothers about license level in the cloud? If you implement it right, it's very difficult for your customers to get to the core code. No more code stealing, reverse engineering made more difficult, it's more open only if you want it to be. A monopolist's dream come true. Only said monopolist has to be able to innovate.
If you don't want to be open-source, well, go and implement your own stuff. I always find it funny that people are happy to use GPL'd -or BSD licensed- code but rant about having to make their own stuff open source. If you're not wanting to share, just build your OS and APIs from scratch. Taking an open-source base and ranting about having to share your code is like using schools and roads and hospitals but not wanting to pay taxes (irony intended).
Let's play a game. You're a head of IT in a medium-sized company, you have an annual budget of, say, $ 100,000. MS licences draw a whopping 20,000 out of that, and, as their stuff is not well documented you have to subscribe to an on-site assistance thing which is likely to be an additional 10,000. You still need some hardware (hardware tends to fail from time to time), so you'd better save $ 20,000 (hey, these things are not cheap).Now you need 2 techies to get the thing running, and to show the average user where the "any" key is. What do you have left to pay them? $25 000 each (fictionnal numbers. Of course). Let's say you switch to open source. The licensing burden disappears, you take the "enterprise" package just to be on the safe side (though as it's well documented, your techies could probably do the job, but let's just be safe), which is $10,000. The hardware is still $20,000, you still have two techies to take care of the machines and show the "any" key to the lusers AGAIN, but all of a sudden you can pay them 35,000 a year each. instead of 25,000. Yeah, sure, open source will reduce IT worker's salaries... or won't it?
Let's assume for the sake of argument that MS do lose out on a significant amount of licence revenue. Where is that money going to go instead? By definition it either goes to Microsoft competitors, or it stays in the coffers of the Microsoft customers. Microsoft themselves don't employ many IT people, they largely employ marketeers and evangelists and channel managers and non-IT B-ark dross like that (sorry if you are one, but your free ride is hopefully terminating soon). Any current MS resellers solely dependent on affected MS product might lose out too, but anyone with a clue knows that "Microsoft business partner" with no non-MS fallback strategy is a disaster waiting to happen.
So remind me again how ordinary decent clued-up non-MS-dependent IT people (?) lose out if the Microsoft marketing machine loses a piece of its pie?
In local Government License costs look like a huge drag on the organisation which prevent us doing much useful with the IT. Support - fine, don't mind paying for that, but the sooner we can ditch the huge license costs and actually get on with using IT to help people do their jobs more efficiently and thus cost the taxpayer less money for a better service then the happier I shall be...
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
