Cloud Crossroads: Which Way PaaS? - InformationWeek
Cloud // Platform as a Service
09:48 AM
Connect Directly

Cloud Crossroads: Which Way PaaS?

As allies and competitors line up behind the new Cloud Foundry Foundation, where does that leave Red Hat and its support of OpenStack and Project Solum?

Top 10 Cloud Fiascos
Top 10 Cloud Fiascos
(Click image for larger view and slideshow.)

There's a fight brewing over which form of open-source platform-as-a-service should command developer loyalties and serve as the basis for future cloud applications.

Should it be native to the main open-source cloud project, OpenStack, and work closely with OpenStack components?

Or is it good enough that it runs on top of OpenStack but has a broad mix of powerful backers to sponsor its development?

Red Hat was counting on the former and working to combine its OpenShift PaaS and a new OpenStack project, Solum, to form a powerful new PaaS platform. Red Hat is a leading contributor to OpenStack, and a set of Red Hat engineers jumped into Project Solum within hours of its announcement by Rackspace. The combination of OpenShift and Solum was an exciting prospect, wrote Red Hat's OpenShift guru Matt Hicks back on Oct. 24.

"While OpenStack is a fast moving space, we have a lot of experience with it and believe that there is tremendous potential to align our PaaS approach with this project," he wrote in that blog post. Red Hat would supply development tools and Linux container expertise; Solum would provide application deployment and lifecycle management. It seemed a perfect marriage. But that's not how it's working out.

Rackspace's principal architect, Adrian Otto, launched Project Solum and also posted a complimentary blog Oct. 24 on the Rackspace website. He said, "We feel very fortunate to be working with eBay, RedHat, Ubuntu/Canonical" and others in forming Project Solum.

That mutual admiration makes it difficult to explain why, after just four months of collaboration, Rackspace appears to have changed its mind and is now a sponsor of Cloud Foundry, the competing open-source PaaS. It's not just contributing code but ponying up big bucks to launch a Cloud Foundry Foundation and sit on its board of directors as a platinum sponsor. Red Hat's other erstwhile ally, Canonical, is sitting on the community advisory board of the new foundation, due to appear this summer, according to an announcement Feb 24.

[Are you addressing the top threats to cloud security? See 9 Worst Cloud Security Threats.]

Cloud Foundry is open-source PaaS, started by VMware and spun out with other VMware-owned open-source code into EMC and VMware's Pivotal Software subsidiary, headed by Paul Maritz. Maritz sensed Pivotal had a successful open-source project on its hands that already enjoyed broad backing and the activity of 750 developers. It would be more successful, he and others concluded, if it could be freed from what many viewed as VMware and Pivotal's dominance of the project.

"Cloud Foundry ... has evolved into the basis for a true open platform ecosystem," Maritz wrote in a Feb. 24 blog on the Pivotal site. PaaS is one of the primary forms of cloud computing, probably second to infrastructure-as-a-service (characterized by Amazon Web Services). With PaaS, developers get a broad set of tools and development assistance on the basic plumbing of an application, allowing them to produce end results faster. PaaS also looks to assist the time-consuming task of preparing a new application for deployment.

Both Cloud Foundry and OpenShift look to help developers produce applications and then give them several cloud-deployment options. Maritz's comments mirror an ongoing debate within the cloud community over how effective PaaS can be by itself. Isn't it really just a feature of good IaaS? His answer was no, it's important on its own, and Cloud Foundry will prove it.

(Image credit: Wikipedia Commons.)
(Image credit: Wikipedia Commons.)

To a still-unknown extent, that's newly hatched competition and potential bad news to Red Hat, which was counting on the "native to OpenStack" Solum, combined with Red Hat Linux and Red Hat OpenShift.

Red Hat is a dominant player in its own right when it comes to the Linux operating system. Workloads that enterprises send into the cloud, when they run under Linux, are overwhelmingly running under Red Hat Enterprise Linux, thanks to Red Hat's extensive testing of the system. By linking OpenShift to OpenStack, Red Hat had a shot at capturing more developers and tying them more closely to its growing catalogue of Red Hat middleware and virtualization software.

Key industry players don't want to go along with it on that path. IBM and HP are backers of OpenStack, but they're also platinum sponsors of the Cloud Foundry Foundation, along with SAP, EMC, VMware, and Rackspace. Platinum sponsors commit to support the foundation with paid developers and financing. (In the case of OpenStack, platinum sponsors pay $500,000 a year and get a guaranteed board seat.) CenturyLink and ActiveState will also join the new foundation as gold sponsors. They will contribute a lesser amount and nominate a candidate for the board, with a limited number of seats drawn from all gold sponsors.

Red Hat could only have watched in dismay as Pivotal's moves shifted the terms of discussion over open-source PaaS. Not only were some of the early and savvy backers of the Apache Web Server and Linux, such as IBM, lining

Next Page

Charles Babcock is an editor-at-large for InformationWeek and author of Management Strategies for the Cloud Revolution, a McGraw-Hill book. He is the former editor-in-chief of Digital News, former software editor of Computerworld and former technology editor of Interactive ... View Full Bio

We welcome your comments on this topic on our social media channels, or [contact us directly] with questions about the site.
1 of 2
Comment  | 
Print  | 
More Insights
Newest First  |  Oldest First  |  Threaded View
Page 1 / 2   >   >>
User Rank: Apprentice
7/30/2014 | 11:55:40 AM
Re: How will this PaaS tragedy play out...
WSO2 has a cross-IaaS PaaS that in Apache called Apache Stratos.  Apache Stratos is being used by companies such as Cisco for Telecom applications.  Cisco is a major contributor as well as half a dozen other major companies.  It implements cross-platform capabilities utilizing Apache jclouds and can support many containerization technologies and is fully polyglot as well.   This is similar to Cloud Foundry's capabilities.  Apache Stratos can support multiple IaaS simultaneously including a combination of local IaaS and public.  

I personally distinguish a PaaS from IaaS by looking at PaaS as the application layer of IaaS.  I think this distinction is critical from an architectural and market point of view.  IaaS provides the services for managing hardware infrastructure.  Services that are built on top of that are SaaS services or PaaS services that help you manage applications layer.  I think that this is the way the market will have to divide.  This doesn't mean that an IaaS player can't also provide SaaS and PaaS services as many of them do, but they also support other SaaS and PaaS offerings by their customers as they must.  

For an enterprise there has always been the issue of whether to tie yourself to a specific vendor.  I believe Enterprises for their own good and survival in some cases will require that IaaS vendors keep their SaaS/PaaS offerings separate and not force them to use their offerings.  I will agree that so far the IaaS vendors are crossing the line and trying to couple IaaS/PaaS services and some Enterprises may go all the way with them but a large number of customers will choose to use their own PaaS for many reasons.  

In many cases an enterprise has to use multiple IaaS vendors because of cost issues, specific compliance issues in certain countries, features in other IaaS vendors or SaaS companies they want to leverage or optimize.  Larger Enterprises will find it hard to standardize on one IaaS vendor across the whole company leading eventually to cross platform requirements.
User Rank: Apprentice
3/10/2014 | 10:50:04 PM
Welcome To The Playoffs, Who's in the Superbowl?
Mark Thiele's post touched on the point that the two projects discussed here represent a subset of the overall PaaS market, but neither is focused on migrating all types of enterprise workloads. Indeed, both are focused on deploying next generation cloud-design applications. Ultimately, whatever the outcome of these two projects battling it out, it's merely the playoffs. They still have to square off with Azure, AWS, Apcera in the public arena and the likes of established vendors that Gartner groups into aPaaS in the private cloud space to see what share of the overall PaaS market they will garner. 

Personally, I wouldn't count RedHat out. PaaS adoption is a very small subset of cloud initiatives today. As the market for these platforms grow, they will be looking for vendors that have success in the application development space. JBoss owns a lot of that attention today. If RedHat can convert their JBoss clientele into OpenShift users, this opening salvo by Pivotal et. al. will merely be the lucky kickoff return for a touchdown followed by no points on the board for the rest of the game.

Moreover, consider what IBM did to Sun with Java with regard to fostering it's most successful use and becoming it's number one benefactor. Now, look at what's occurring with Pivotal and look what they're doing with BlueMix. The strategies are very similar. They rode Sun's large-scale investment and ownership of the platform until Sun was gone and now they have the opportunity to repeat with CloudFoundry.

In the endgame, AWS claims they're not a PaaS, but they're not fooling anyone. The next round of PaaS brackets break down as follows:

AWS vs. IBM <-> Microsoft vs. RedHat
Charlie Babcock
Charlie Babcock,
User Rank: Author
3/10/2014 | 8:15:58 PM
Is PaaS a feature of a cloud or gateway to a cloud?
Is PaaS just a feature of infrastructure as a service? Reuven Cohen made that argument convincingly here:  

Or is it the other way around. The best way to access cloud infrastructure is through a PaaS. For the enterprise, I think it's this prospect that has long term appeal, though I don't know that all enterprise IT shops think this way. It takes a DevOps mindset and that's still a suspect proposition in some cases. Despite disadvantages at the moment, Red Hat is in a position to deliver on OpenShift as a way to build RHEL applications in the cloud and deliver RHEL applications to the cloud. Will it still strive to do so through an OpenStack project, with its former compatriots defecting to Cloud Foundry? I don't know. I'm thinking Krishnan Subramanian may offer part of the answer in another contributed IW column soon.
User Rank: Strategist
3/10/2014 | 7:02:36 PM
Re: Meh....
Ben, you are entitled to have your own opinion. But your claim that we haven't spoken in a long time is not true. Just last week we exchanged email regarding the past business relationship involving money. I am not debating with you on other topics you raised as it is your own opinion but I just wanted to point out that we have coimmunicated as recently as last week.
User Rank: Apprentice
3/8/2014 | 12:33:22 PM
Does there have to be an either/or?
I am ready to take on Josh for $10, heck, I can even up the ante and go for $25 bucks! Josh, are you on?

Having said that, I am not sure why we have to choose between Solum (or whatever it will end up being) and CloudFountry. The two, in my opinion, have two distincly different value propositions. 

CloudFoundry is a cross-platform PaaS that is designed to work on top of multiple IaaS's and provide the application developers and owners cross platform support. 

Solum is a native PaaS for OpenStack that will guarantee the best native integration with the core of OpenStack, thus develiering better performance, seamless integration and better economics. 

A rough analogy from the yesteryear would be ODBC vs. PL*SQL on top of the Oracle database. ODBC was commonly used by people wanting to support Sybase, Informix, DB2 and MS SQL, while PL*SQL was also commonly used by the folks who wanted to optimize their Oracle database performance only. 

Both approaches are viable and are likely to have many customers. 
User Rank: Ninja
3/8/2014 | 9:19:55 AM
Traction is the key
I think Mark Thiele has nailed the issue here: PaaS is still niche. It reminds me of Linux in the mid-to-late-1990s: quite fragmented, and while it offers benefits, the learning/implementation curve combined with legitimate concerns that you can't actually implement what you want without breaking the benefits with hacks and workarounds means much more limited adoption than OpenStack/CloudStack/other private-cloud-enabling platforms. Add to this the fact that the PaaS choice for many marquee organization is a completely different one--CloudBees--that focuses on a more narrow goal (smaller learning curve, more guaranteed implementation). (Also in this category: Azure, Apprenda). I remain skeptical that non-VM-based deployments (e.g., based on any Linux scripting language) can fit properly into the PaaS mold. Those deployments are overwhelmingly choosing configuration management (usually Chef or CloudFormation) over PaaS. So I see this debate as a lot of soon-to-be-legacy companies fighting over a small space that's being won by different solutions. Citing 750 developers is pretty pathetic--check GitHub for cloudformation-based code, and you'll see why.
User Rank: Apprentice
3/7/2014 | 10:58:12 PM
With the usual series of disclsoures (for details, just look at my disclsoure page but briefly, I'm an adviser to ActiveState by way of their acquisition of Appsecute. Previosuly I've worked with Krishnan Subramanian who, despite saying his tweets and blog comments are completely his alone, also works for OpenShift/RedHat. Note that while Krish and I used to work together, we haven't spoken for a long time and the only time I hear from him is indirectly by way of tweets questioning my sanity/morality/ethics/etc. I've spent time, shared beers and dinner and generally hung out with a bunch of folks in the PaaS world), here's my take.

10 months? Sorry Josh, if only for ego and appearances sake, RedHat won't pull the plus on OpenShift in that time. That's not to say that it won't happen, it'll just take a little longer. The very interesting thing would be to see how RedHat spun it and what it meant for the vehement "OpenShift forver" crew within RH.

Clearly Rackspace screwed over RedHat in launching Solum and then jumping into the Cloud Foundry world. But what else could they do? Solum is a nothing, an announcement with little substance. OpenShift may have customers but it's unfortunately either created or been thrust into an apparent zero sum battle with Cloud Foundry and it's patently obvious who is winning that - CF has almost every big vendor on board, it's winning the attention battle and unfortunately OpenShift responds not with good stories of customer wins (which I'm sure exist) but with a continual barrage of snark - regardless whether it comes form individual's accounts and is appended with a "personal views only" statement, that stuff sticks and it feels like a very defensive RedHat coming through.

It's also worrisome that apart from a small band of folks, we don't see a lot of love given to OpenShift from the RedHat execs, maybe the RedHat summit will change that but, for obvious reasons, I'm about as likely to hear about that as James Watters is to get an invite to the OpenShift drinks.

Underlying all of this is the fact that Docker specifically, and containers generally, are the flavor de jour, and much of the PaaS value prop is delivered by containers (at least for Linux apps). In this regard Cloud Foundry has the advantage since (for Pivotal at least) it's part of a greater whole and Pivotal can quickly spin a story about all the other goodness that was spun into it from EMC and VMare.

PaaS isn't dead (despite comments from pundit-critics who suggest I've said it is), but PaaS is going to change as it works out where it sits between IaaS vendors rapidly moving up the stack, container offerings doing much fo the protability stuff, management tools doing scaling, and application vendors moving down into the declarative platform space. Lots of thinking to do huh?

Right - that's my 2 cents, now we just wait for the barrage of comments suggesting that I know nothing about the space. Coming in 3, 2, 1...
User Rank: Apprentice
3/7/2014 | 6:58:30 PM
Re: RedHat joining Cloud Foundry foundation is wrong for PaaS
Renat, I agree with your point:

"If one thing that RedHat could do to make the world a better place - is to collaborate with Cloud Foundry community, to standardize around run-time and services, which are now incompatible."

That's why Stackato decided to use buildpacks (only used by Heroku at the time, IIRC) almost two years ago. Of course, Cloud Foundry is now on buildpacks as well. Everybody standardizing on buildpacks would be a step in the right direction. Services themselves are pretty standard across PaaS platforms but the "connectors" to them are not yet.
User Rank: Apprentice
3/7/2014 | 6:25:02 PM
Re: How will this PaaS tragedy play out...
Hi Charles, just want to clarify that I work for SUPERNAP who runs the SUPERNAPs.  :)
User Rank: Apprentice
3/7/2014 | 6:24:36 PM
Re: Does Joshua only pay north of the border?
Charles, I only mentioned Canada in response to you wondering whether you'd have to pay tax on your winnings; I have no idea how the IRS works, but here in Canada you'd be fine. I didn't even know Josh was a fellow Canadian... And, sadly, the loonie is only worth about 90 cents US currently so you're better off collecting down there.
Page 1 / 2   >   >>
Register for InformationWeek Newsletters
White Papers
Current Issue
Cybersecurity Strategies for the Digital Era
At its core, digital business relies on strong security practices. In addition, leveraging security intelligence and integrating security with operations and developer teams can help organizations push the boundaries of innovation.
Twitter Feed
Sponsored Live Streaming Video
Everything You've Been Told About Mobility Is Wrong
Attend this video symposium with Sean Wisdom, Global Director of Mobility Solutions, and learn about how you can harness powerful new products to mobilize your business potential.
Flash Poll