Cloud // Platform as a Service
Commentary
1/15/2014
02:40 PM
Connect Directly
RSS
E-Mail
50%
50%

PaaS Is Dead. Long Live PaaS

Platform-as-a-service is not dead -- Red Hat exec Krishnan Subramanian says it has evolved to a higher state.

Recently a new meme has started among the cloud chatterati: the death of PaaS, or PaaS morphing into functionality for IaaS or PaaS. This discussion peaked when an analyst from 451 Research wrote a report titled "Is PaaS becoming just a feature of IaaS?" (subscription required). An example of how others have picked up the theme appeared Jan. 14 in Network World.

In today's fast-evolving technology world, it's natural for pundits to write premature obituaries, but it's also important to look deeper and understand the reason for their confusion. One explanation is that PaaS has gone beyond hype and reached a new level of maturity. Last year, many enterprises adopted PaaS for their production workloads, but 2013 was also the year when the demarcation between two flavors of PaaS became clear.

Another reason is the strategy adopted by cloud vendors like Microsoft and Google. Both started off with PaaS but eventually started offering IaaS to compete with Amazon Web Services. Microsoft in particular has blended its offering with both PaaS- and IaaS-like features.

Two flavors of PaaS
In the beginning, PaaS was constructed as the middle layer in the cloud stack between IaaS and SaaS to describe the evolution of application deployment platforms in the cloud. PaaS was dominated by services like Google App Engine, Heroku, and Engine Yard in those early years. As the space matured, we saw the evolution of PaaS meeting the needs of modern enterprises.

[For another perspective on the state of IaaS and PaaS, see Is The Cloud Platform Battle Over?]

Innovations by Red Hat in OpenShift, VMware (now Pivotal) in Cloud Foundry, Docker in its dotCloud platform, and others led to a new generation of PaaS offerings that are architecturally different from the first generation. In this article, I will discuss the two flavors of PaaS.

PaaS by service orchestration
This flavor of PaaS, offered by early providers like Google App Engine, built the platform by composing different services needed for application deployment. It started with a compute fabric, on which services for data storage, monitoring, logging, etc. were added. In this flavor, PaaS is nothing but a composition of services needed for the applications to run. These were purely hosted offerings, and building and managing a platform using service orchestration was easy for providers.

PaaS by container orchestration
As PaaS matured and enterprises started to demand a private version, another flavor developed. This was mainly due to the maturity of Linux Container technology and the feasibility of building a platform based on containers that could also be implemented and managed on the premises.

Docker is a perfect example of fast, lightweight Linux containers that make it easy for users to port their applications across different cloud providers. Unlike virtual machines that abstract only raw compute, containers can encapsulate entire applications and application environments. Vendors like Red Hat, dotCloud (now Docker), VMware (now Pivotal), and others saw value in the container-based approach: An orchestration and management layer on top of containers can offer the same seamless developer experience as the first-generation PaaS. For a growing number of vendors, container orchestration is fast becoming the norm. This is further proven by the growing interest in OpenStack's standalone Project Solum to build an abstraction layer for application deployment using container technology.

When you step back and consider the two flavors of PaaS, it's easy to understand why some industry watchers are confused. The service orchestration-based approach of PaaS is similar to SaaS architecturally, and the container orchestration approach of PaaS is similar to the VM-based approach of IaaS. Though SaaS and IaaS can be extended to resemble one of these two flavors, there is a clear value proposition for PaaS as it is defined as a separate layer in the cloud stack. This is especially true with the container orchestration flavor.

PaaS removes the complexity and costs associated with IaaS, whereas companies like Netflix use additional tools to build a platform atop Amazon. Though Netflix approach might fit its needs at web scale, PaaS can greatly simplify enterprise IT departments' efforts as they transition to a more modern system.

It's natural for similarities between platforms to cause confusion, but a premature obituary for PaaS will end up hurting an industry that could otherwise derive great value from ongoing development.

Krishnan Subramanian is director of OpenShift strategy at Red Hat and a frequent cloud commentator. He was formerly an analyst with Rishidot Research and a physicist.

Can the trendy tech strategy of DevOps really bring peace between developers and IT operations -- and deliver faster, more reliable app creation and delivery? Also in the DevOps Challenge issue of InformationWeek: Execs charting digital business strategies can't afford to take Internet connectivity for granted.

Comment  | 
Print  | 
More Insights
Comments
Newest First  |  Oldest First  |  Threaded View
Ruslan Synytsky
100%
0%
Ruslan Synytsky,
User Rank: Strategist
1/25/2014 | 9:58:25 PM
Next gen of container based PaaSes
Krish, thanks for the article.

Absolutely agree that PaaS history has different ages of evolution. I believe it makes sense to add the link to our 3 years old vision about PaaS 2.0.

Also it's worth to mention that Jelastic is the pioner platform that uses containers from the first days of the architectural design. Now it's "must have" architectural element of a mature PaaS.
cbabcock
50%
50%
cbabcock,
User Rank: Strategist
1/21/2014 | 7:17:37 PM
The lines will blur
Re: Stratustician, "PaaS will really just evolve..." I can agree with that statement. The lines are likely to blur considerably, with most IaaS offering PaaS-like deverloper services and PaaS taking on many of the optoins of IaaS. The definitions are not hard and fast. They are more NIST's descriptions in 2009-2010 of what existed as we first started talking about cloud computing.
Stratustician
50%
50%
Stratustician,
User Rank: Ninja
1/20/2014 | 1:35:36 PM
How will they use it?
I think that PaaS will really just evolve depending on how we see the market start to really take advantage of cloud.  If we see more SaaS style adoption, PaaS services might evolve to be a simple sandbox to run applications, without really building out too much more of the service.  If it is used for a testing ground for developers, we might see it morphed into a fully-stocked IaaS solution that offers some of the current basics of PaaS.
chipchilders
100%
0%
chipchilders,
User Rank: Apprentice
1/16/2014 | 8:29:40 PM
It's not the Death of PaaS, it's just getting put in the right place
I'm certain that we aren't seeing the "death" of PaaS, but are seeing a reality develop in the market around it's relative import and the value of different approaches to supporting application owners.  "Pure" PaaS solutions are certainly valuable under the right conditions, but come with all the same implementation and learning curve challenges inherent in any new and complex technology.  More importantly, they often require rethinking the applications themselves, which is something that I find hard to swallow as an engineer. When you step back and look at the evolution of IT in a more general way, one stark reality stands out...  the long tail.  Enterprise developers need to find ways to take advantage of the enormous value possible with orchestration and abstraction, but in ways that bridge the gap between the future and the current realities.

In my opinion, Amazon, Google and Microsoft have it right...  they offer the power of modular services (think RDS, DynamoDB, etc...), coupled with composition *options* (call this the "PaaS" if you want to) for the application owners is the path forward.
philwhln
50%
50%
philwhln,
User Rank: Apprentice
1/16/2014 | 6:47:42 PM
Laurianne
50%
50%
Laurianne,
User Rank: Author
1/16/2014 | 4:24:00 PM
Re: PaaS is evolving to a higher state
Krishnan, thanks for weighing in, and thanks for the link to the explanation on Docker.
iamkrishnan
50%
50%
iamkrishnan,
User Rank: Apprentice
1/16/2014 | 2:21:10 PM
Re: PaaS is evolving to a higher state
I would be ok with argument that says IaaS will take advantage of PaaS layer to offer a more integrated offering because PaaS needs some infrastructure to run atop (IaaS or traditional servers). I am pushing hard because they seem to dismiss the very usefulness of PaaS itself. This is a good intro to Docker http://www.slideshare.net/dotCloud/why-docker2bisv4
iamkrishnan
50%
50%
iamkrishnan,
User Rank: Apprentice
1/16/2014 | 2:18:42 PM
Re: We are simplifying too much
Johan,

I read your post. Good post. It will align with what I am talking about if you combine Layer 2 with layer 3 into one stack. In the current version, it aligns with my orchestration of containers flavor but not sure how you will fit Google App Engine kinda approach to PaaS.
cbabcock
50%
50%
cbabcock,
User Rank: Strategist
1/16/2014 | 12:49:38 PM
PaaS is evolving to a higher state
It's not generally understood what containerization (example, Docker) is doing for cloud workload deployments. I don't think I"ve done a very good job of explaining it myself. (But I will keep trying). In the long run, Reuven Cohen hsa a point; PaaS will be built into most cloud platforms. Today, PaaS is showing the most rapid evolution of the different forms of cloud computing and it's a bad call  to say it's about to disappear. Anyone disagree? How about Jay Lyman, at Forrester Research?
JohanDenHaan
100%
0%
JohanDenHaan,
User Rank: Apprentice
1/16/2014 | 2:54:02 AM
We are simplifying too much
> It's natural for similarities between platforms to cause confusion, but a premature obituary for PaaS will end up hurting an industry that could otherwise derive great value from ongoing development.


I agree with your conclusion. We see the confusion all around: everything is called a PaaS nowadays, not in the least because vendors want to be part of the "hype". I think the problem starts with the fact that we, as an industry, are still using the three flavors of cloud (IaaS, PaaS, SaaS) as a way to describe the landscape and categorize offerings. This is too much of a simplification of reality nowadays.

It will be helpfull in my opinion to clarify the different flavours of cloud platforms a bit more. I did a first attempt in making a layered model that distinguishes between different flavours of IaaS, PaaS, and SaaS, including their target audience.

Do you think your distinction between PaaS by service orchestration and PaaS by container orchestration aligns with this model, or did I miss something?
Google in the Enterprise Survey
Google in the Enterprise Survey
There's no doubt Google has made headway into businesses: Just 28 percent discourage or ban use of its productivity ­products, and 69 percent cite Google Apps' good or excellent ­mobility. But progress could still stall: 59 percent of nonusers ­distrust the security of Google's cloud. Its data privacy is an open question, and 37 percent worry about integration.
Register for InformationWeek Newsletters
White Papers
Current Issue
InformationWeek - September 2, 2014
Avoiding audits and vendor fines isn't enough. Take control of licensing to exact deeper software discounts and match purchasing to actual employee needs.
Flash Poll
Video
Slideshows
Twitter Feed
InformationWeek Radio
Listen Now This InformationWeek
In in-depth look at InformationWeek's top stories for the preceding week.
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.