Does your OS matter in a cloud?
Summary: Putting your computing into a cloud can be like having Blackwater fight your wars. It's outsourcing with no in-house expertise to watch the watchers.
Once applications move into cloud computing do arguments about operating systems, closed source or open source, really matter?
I thought of that while reading about the joint HP-Intel-Yahoo cloud study. (This cloud was caused by a volcano. Man-made clouds are far more sinister.)
My point may sound counter-intuitive. The funding of 10Gen tells Big Money Matt that clouds accelerate the open source trend, not just due to cost but because the only folks dealing with software innards in a cloud are experts.
Tim O'Reilly sees a threat to open source in cloud computing, and I agree, even though it makes all computing, essentially, enterprise computing, and gives both open-and-closed source the same SaaS business model.
Then there's the question of whether all this cloud stuff isn't just hype, another way of selling the failed ASP model from 10 years ago. The recent outage at Amazon, and the need seen by Hyperic to monitor clouds, tells us there is no magic here.
What's really the difference between cloud computing and conventional Web hosting? Scale? Or just hype? Isn't Salesforce.com just a cloud?
If there really is no difference, what we may be seeing is a move by all types of applications into an enterprise space. That would seem to offer open source a level playing field, which is all its advocates have really ever wanted.
But will we really get it? When technology is abstracted from what you're doing, who knows or cares what deals may be made behind the curtain?
Putting your computing into a cloud can be like having Blackwater fight your wars. It's outsourcing with no in-house expertise to watch the watchers.
You may notice I'm drawing no conclusions here. Cloud computing is a lot of hype, a lot of hope, and hard to really get your arms around.[poll id=87]
Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.
Talkback
Some questions
1) Data. Where, exactly, is my data and is it secure. More importantly is it still legally my data. We all put money into the bank knowing that if the bank goes so does our money - does this same model apply???
2) Applications: Is there any reason to assume the "web-based" applications will be more stable, tamper proof and easy to develop as compared to the linux/windows/apple desktop. My experience with main frame computing leads me to believe the opposite - although it was convenient, when finally developing an app, to have the app immediately available.
As to open-source - i think that "cloud-computing" will quickly become commercialized and tiered resulting in more uncontrolled proprietary "standards" which would shut out open source.
Thanks
Is this really a concern?
Web apps have been replacing desktop apps for years now. Most place I go are working to get everything off the desktop and onto the intranet or internet. This is definitely not a new trend.
So in the end I question all of this skepticism. Have we not been paying attention to where our data actually goes? Are we the techies just as buzzword crazy as the business execs? What will probably become all of this is simple...nothing more than whats already going on. Big companies will continue to run their own datacenters and smaller less tech oriented organizations will continue to use hosted services. The only real difference is who is managing the servers in the end.
re: Some questions
Your data probably is secure. It's the potential down time that bothers. It sometimes is that it's more secure than you'd like. GMail for example doesn't delete your email even when you instruct so. Is that the behavior you want? The stuff you put into cloud is actually more than just data but your social pattern which has commercial value in the world of advertising. While you are unware of it, the cloud companies are trying to profit off that.
"2) Applications: Is there any reason to assume the "web-based" applications will be more stable, tamper proof and easy to develop as compared to the linux/windows/apple desktop."
In the current world of AJAX, serious web app development is impossible. Sth else as rich as, or at least close to, desktop framework has to be established b4 web apps take off. The future belongs to the framework utilizing both the connectivity and massive content of web and the advanced, rich hardware installed on a local computer.
It is admin types, not users, who advocate moving everything to the cloud
I'm a User/ Network Specialist in training.
Is this not a common ad theme, 'Power yourself with ... yada, yada computer'. The cloud issue is begining to be an issue of the persons' honesty rather then the persons' privacy? Eventhough, Americans love to watch the crooked get nailed. Don't we? I prefer to send my personal emails via Express, simply because of internet threats from virus infections. Why compromise that security? I think MS continues to offer secure personal computing, it seems to me that it is the Open Sorce end that may be questioned.
RE: Does your OS matter in a cloud?
In "closed clouds" such as Salesforce.com or Google App Engine, it is the cloud provider that as you say will decide what to use behind the scenes and they may or may not choose to use open source based on the usual considerations.
In "open clouds" such as Amazon EC2 or GoGrid (clouds in which the user can choose the application platform), users will decide whether or not to use open source, based on the usual considerations.
So again, I don't see how the cloud affects the open source question.
Or Both
communicate via open and documented protocols so any
one may join and any one may serve and charge. The
cloud and the clients are loosely coupled.
Hard to imagine this ever happening, but perhaps a
company might think it's more profitable to build silos and
control the cloud, client, and protocol. Over the long term,
more expensive and less reliable, but in the short term, via
introductory discounts...
I notice that this morning on slashdot, there are slams on
Mozilla/Firefox from two angles. We're in the news cycle
about the Fortify guy chiding open source projects for
security. Microsoft went to press in its financials with the
assertion that increased Apple purchases may be a threat
to profits and that open source is really purloined
Microsoft
innovation. I put these bullet points together with the
interoperability talk (and some money) and all the tech
press on Midori vaporware and it doesn't spell mother.
I think Redmond would like some time to build their silo. I
thought 300 million was a bit pricy if all one wanted to say
was "Microsoft Vista, now new and improved."
It's the business process - not the application
Maybe when you strip everything down, from the customers view <b>SaaS</b> is simply a business process delivered as a web accessible service instead of a locally hosted and administered software application. The customer is of course always on the prowl for efficient and ever more cost effective productivity gains. Wherever access and bandwidth reliability combine with data security, SaaS solutions knock the crap out of efficiency and cost effective concerns. It's the Web however that makes the productivity part of the equation an explosive leap, with possibilities and opportunities showing up every time a web worker engages the service.<br><br>
Open Source is simply a very effective way for service providers to pool their resources at the standardized "lower" layers of the application stack. Higher levels of the application stack combine with specific services and web accelerators to create distinctive market differentials. Cooperate low, compete high; use open formats, protocols and interfaces and ride the open source - SaaS/SOA wave for all it's worth.<br><br>
For OSS oriented service providers the formula is: Innovate at the top; run out the value of the differential; and feed the innovation back into the OSS machine when you push the stack value higher. The cost of competition is lower, the differential is narrow but focused and specialized, and, your customers enjoy the high value of open standards based substitution.<br><br>
These are good things. Proprietary software vendors will no doubt try to position their applications as SaaS-Cloud based services. To maintain their high profitability though, these vendors will have to seize control of the basic formats, protocols, and interfaces used in their proprietary stacks. To do this, they will most likely use some variation of the infamous Microsoft model of <i>embrace, extend, extinguish</i> applied to open web - open source formats, protocols and interfaces.<br><br>
In other words, as traditional <i>client/server</i> vendors make the transition to Web Stack centric models, they will attempt to lock in transitioned SaaS-Cloud customers at the lowest levels of the application services stack. That way they never have to compete at the higher <i>innovative</i> levels.<br><br>
No surprise there, but it's one more reason why OSS will thrive. IMHO, the future of the Open Web depends on the continued expansion of OSS; with SaaS-Cloud based service providers increasing their support of OSS efforts as a highly competitive means of cracking into the <i>client/server</i> business process stranglehold; a stranglehold exposed and ripe for the taking as established software vendors try to transition their marketshare to new <i>client/ Web Stack /server</i> model.<br><br>
For customers, it's all about critical day-to-day business processes. Unfortunately, for <i>client/server</i> solutions, most of these processes are locked into application bound <i>client/server</i> systems. The data is portable, but the logic behind the process is not.<br><br>
Enter the great transition of these processes to an emerging <i>client/ Web Stack /server</i> model. Whether you look at this as J2EE, SaaS, Cloud Computing, SOA, PaaS or Web 2.0; the potential of this transition is always that of more efficient, more cost effective and infinitely more productive business processes. Traditional systems providers are in a position to <i>promise</i> a more integrated and interoperable <i>transition</i> model. Emerging Web service providers face the challenges of what sometimes looks to the customer as being a disruptive <i>"rip out and replace"</i>.<br><br>
To me, the more interesting aspect of this transition is the reality that Microsoft Office sits in the center of 95% or more of these day to day business processes - no matter what the client/server system. I would even go so far as to say that Microsoft owns the <i>"client"</i> in <i>client/server</i>.<br><br>
So far they have been very successful in protecting the MSOffice platform franchise at the low level of formats, protocols and interfaces. But customers are on to the gambit and are increasingly demanding that Microsoft disclose the proprietary formats, protocols and interfaces - and/or implement open alternatives.<br><br>
Meanwhile, the Microsoft Web Stack, anchored by Exchange, Sharepoint and SQL Server, has turned into what looks to be an unstoppable juggernaut. The fabled Microsoft Developers Network is increasingly engaged with the great transition. Since most of the client/server solutions involved MSDN, the developers network is uniquely positioned to make the transition pitch to existing customers. Whether or not they choose the MS Web Stack-Cloud-.NET platform remains to be seen. But Microsoft has hedged their bets with some excellent, albeit exclusively MS Web directed, transition tools via Visual Studio and the .NET framework.<br><br>
At the end of the day though, the thing to watch is the low level formats, protocols and interfaces. Microsoft owns the information editing space with MSOffice bound business processes dominating without challenge near all business desktops. Alternative Web 2.0 editors are very limited. Useful to consumers and individuals, but near worthless when it comes to workgroup-workflow related integrated business process logic. After five years of efforts to replace MSOffice on those desktops, even the OpenOffice - OpenDocument - EU IDABC crowd admits that the only way forward is to coerce Microsoft into enabling open formats, protocols and interfaces in MSOffice. This seems to have worked. The question remains as to whether or not Microsoft will implement these open formats, protocols and interfaces in ways that will empower competitors to re direct the great transition away from the MS Web Stack - Cloud; and towards Open Web alternatives? I think not.<br><br>
In the December 2007 beta release of the MSOffice SDK there was a nifty little tool for the quick roundtrip conversion of MSOffice XML documents to something called XAML "fixed/flow". The roundtrip aspect is important in that business process specific scripts, macros, OLE, workflow routing and security settings are fully preserved. XAML itself is part of the Windows Presentation Foundation layer; and with other aspects such as Silverlight, XPS, Smart Tags and LINQ replaces Open Web and Semi Open Web initiatives; like HTML5, CSS3, JavaScript, SVG, XHTML2, RDF, SPARQL, XUL, Flash, PDF, and Adobe Flex-AiR.
Clearly Microsoft has a plan to control the low level formats, protocols and interfaces. But that's not to say the challengers come to this fight empty handed. Java, AJAX, PHP, Ruby, Python, Flash, Flex and a host of other methods own the Web 1.0 - 2.0 world. The great transition is still up for grabs, even though Microsoft holds an inside straight.
The MS Web Stack has only become a juggernaut because of the carefully controlled integration with MSOffice-Outlook bound desktops. Some services like Zimbra have cracked the Outlook-Exchange protocol, and are able to connect and compete. Alfresco claims to have similarly cracked the MSOffice-SharePoint protocol. This access to the Microsoft desktop seriously levels the playing field, and Microsoft will have to fight to stay competitive.
Not to discount the importance of Linux and Apache, IMHO the most important OSS initiative today is that of WebKit. This is the advanced browser-device browser <i>layout engine</i> supported by Apple, Adobe, Nokia, Google, RiM, Opera, and KDE. The layout engine implements a WebKit document model that is backwards compatible with the legacy Web, while pushing the envelope with advanced HTML5, CSS3, JavaScript and DOM characteristics.<br><br>
The WebKit layout engine is used by both the Apple iPhone browser and the Adobe RiA AIR runtime. WebKit documents are closely linked to the development of the Mozilla Gecko layout engine through the WHATWG HTML5 and jQuery JavaScript work. Meaning, developers and users can expect sophisticated and complex WebKit documents to display properly across Apple, Nokia, RiM and Google devices; and the Adobe RiA runtime, Opera, and Mozilla browsers. Not to mention the WebKit work going on at Eclipse and QT.<br><br>
Opposite the WebKit-Gecko- Adobe RiA layout and web document model sits Microsoft's Trident and Silverlight model that implements, wait for it now; XAML "fixed/flow".<br><br>
Note that while Microsoft Office Open XML is an open standard (ISO 29500), XAML "fixed/flow" is a proprietary and Microsoft "web ready" document model.<br><br>
Not surprisingly the latest Internet Explorer "Trident" layout engine implements parts of HTML5, CSS2.1 and JavaScript. It looks to me though that they drew the line right at the cusp of complexity that business process documents demand. So my guess is that we won't be seeing anytime soon a MSOffice converter capable of producing Open Web ready documents; especially sophisticated "flow" documents at the level of the WebKit document model (- HTML5, CSS3, JavaScript). What we're sure to get from Microsoft and the Microsoft Developers Network is a high quality conversion of MSOffice business process documents to XAML "fixed/flow". That's a done deal.<br><br>
Imagine all those billions of MSOffice documents "web ready", MS Web-Stack ready, yet still capable of full participation in existing business processes. In this sense, Web values such as collaboration, data and services integration are <i>in-process</i> value added attributes. Want to excite a business customer? Tell them you can add extraordinary productivity value without disrupting existing business processes.<br><br>
From this perspective it's hard to see how SaaS hurts open source. If anything, the pressure is on as never before. It took some time, but Microsoft has put their stake in the ground and are ramping up the great transition. The pieces seem to be in place. So i see OSS as just now moving into the thick of the fight, with the future of the Open Web at stake.<br><br>
As for the guys who will get hurt by this great transition? That's easy to see. It's the standards consortia and organizations charged with advancing the reach and usefulness of formats, protocols and interfaces whose influence is increasingly diminished. Maybe it's just that we are at a time where innovation rules over interoperability? Maybe once the dust settles the world will once again return to reliance on slow moving and vendor dominated standards consortia? Or maybe not. Till that time though, the future of the Open Web is in the hands of open source communities. <br><br>
As a final thought; IMHO, just as XAML-Silverlight will dominate the Microsoft transition process, WebKit will come to dominate Open Web alternatives. Including SaaS, SOA, Java and Web 3.0.<br><br>
~ge~