Dissecting The PaaS LandscapeRishidot Research Webinar
Bookkeeping and Disclaimers
Speaker’s Twitter handle: @krishnan Webinar hashtag: #paasmkt This is not deep dive research rather a 30000 feet
overview on the market. Not all players covered. Deploycon 2013 100% off discount code sent to
webinar attendees Research report shared after Deploycon 2013 Some of the vendors listed in this webinar are either
Rishidot Research clients or Deploycon sponsors
Plan For The Talk
Defining PaaS PaaS Ecosystem and Spectrum PaaS Differentiation PaaS and Big Data Conclusion
SPI Model
PaaS Definition
PaaS is defined as an elastic on-demand platform for applications that completely abstracts away the underlying infrastructure with the application scaling seamlessly with the platform.
P, ugh, aaS?
The idea of PaaS has evolved Enterprise reluctance and diverse needs
has changed the aaS usage in PaaS PaaS Platform Services where services
can be hosted or private #ongoingdiscussion
Let us stop thinking about the debate and start talking about the usage
What is PaaS and What is Not
PaaS definition has broadened but certain characteristics hasn’t changedApplication scales with the platform that offers “infinite” scalabilityNo human interventionNo hardware in the discussion
What is PaaS and What is Not
Application Scales With Platform
Legacy Applications on Elastic Infrastructure
Is PaaS Middleware for Cloud?
Why PaaS?
Developers: Faster Development and Continuous
Deployment (Agile) Fewer Bugs (Similar environments in Dev,
Test and Production) Richer applications due to add-on services Easy seamless collaboration
Why PaaS?
Organizations:Platform for services worldReduced and efficient operationsCost effective ITAgility
PaaS Evolution
From Hosted to PrivateFrom Proprietary to Open
SourceFrom Startups to Enterprise
PaaS Ecosystem
PaaS Ecosystem
PaaS Spectrum
Abstraction Control
InfrastructureServices
Hosted PaaS
PrivatePaaS
DevOpsPaaS
DevOpsTools
PaaS: Axes Of Differentiation
Hosted Vs Private PaaS Single Language (Best of Breed) Vs
Polyglot Proprietary Vs Open Source DevOps Vs NoOps Vertical PaaS
Hosted PaaS Vs Private PaaS
Hosted:• On Demand• Pay per use• Economic benefits and agility• Higher Levels of Abstraction• Lose some control and “lock-in risks”• GAE, Heroku, Engine Yard, Windows Azure,
Appfog, Tier 3, Dot Cloud, Force.com, Orangescape, etc..
Hosted PaaS Vs Private PaaS
Private PaaS:• Less agility• More control and less lock-in risks• Varying levels of abstraction• Struck in CAPEX model• Apprenda, ActiveState, CloudFoundry, WSO2,
Cloudsoft AMP, Cloudify, Cumulogic, etc..Then there is Hybrid like OpenShift, CloudBees, Oracle Java Service, etc..
Best Of Breed Vs Polyglot
Best of Breed: Single Language Platforms Best of breed evolution More depth than Polyglot platforms (Today) Support for legacy applications Enterprise target Apprenda, Engine Yard, WSO2, CloudBees,
Cumulogic, Oracle, etc..
Best Of Breed Vs Polyglot
Polyglot:• Single platform supporting multiple languages
and frameworks• Suitable for modern apps and orgs with multi-
language developer teams• More adoption in startups but enterprises are
slowly embracing polyglot• Heroku, CloudFoundry, OpenShift, Tier3,
ActiveState, AppFog, Google App Engine, etc..
Proprietary Vs Open Source
Proprietary Platforms:• Non availability of source code. Less
flexibility in Platform customization• Hosted or Private• Higher Lock-in risks with hosted platforms• Heroku, Engine Yard, Google App Engine,
Apprenda, HP, CloudBees, Cumulogic, etc..
Proprietary Vs Open Source
Open Source:• Usual moral reasons• Source code available for easy customization• Hosted or Private• Lesser lock-in risks with hosted platforms
under certain conditions• CloudFoundry, IronFoundry, OpenShift,
Brooklyn Project, Cloudify, etc..
DevOps Vs NoOps
Convenience Vs Flexibility Question NoOps Platforms -> More Constraints Certain Applications like Marketing Apps
fits well with NoOps Platforms DevOps Platforms -> More Control Certain Big Data Applications need more
control
Vertical PaaS
Focused on specific verticals like Financial, Health Care, Media, Gaming, etc..
Media PaaS: Azure, AWS, Google, Federated Cloud Platforms
Vertical PaaS for regulated industry Gaming PaaS on federated clouds could offer
high performance gaming based on real time data
Other PaaS
Visual PaaS -> Force.com, Orangescape, WorkXpress, etc..
ALM Services -> CloudMunch, Electric Cloud, etc..
IDE Services -> Cloud9, Codeenvy, Neutron Drive, etc..
Beyond PaaS
PaaS Market Platform Services Market
Mobile Backend as a Service Platform Services
Component Services like Identity, Social, Real time streaming, etc.. Platform Services
PaaS Platform Services
Platform Services
PaaS and Big Data
Current Generation of PaaS is built for scaling users
PaaS v2.0 PaaS for scaling data a.k.a PaaS for Big Data
The evolution has started already but will accelerate in the coming years
PaaS v1.0
Internet of things
Intelligent Applications
Intelligent Platforms (PaaS v2.0)
Conclusion
Enterprise PaaS is real Platform Services are still evolving We need platforms for the internet of
things Platforms for big data Next 3-5 years is going to see emergence
of intelligent platforms
Connect with me
Work: Principal Analyst, Rishidot Research and Editor, CloudAve.comEmail: [email protected]: @krishnanWebsite: www.rishidot.comBlog: www.cloudave.com/author/krishnan/Slides: www.slideshare.net/rishidot