Managing Software from Development to Deployment in the Cloud

Harpreet Singh, Vivek Pandey | CloudBees Managing Software from Development to Deployment in the Cloud

Why PaaS?
PaaS eliminates these deployment friction points and more…
Entire dev-to-deployment to the PaaS
Get to the market faster

Agenda
What’s the cloud and the PaaS?
What should be the criteria choosing a PaaS?
Demonstration of taking applications to the PaaS.

What is the PAAS?
Where does it fit in the whole cloud…

DC versus AC: Replay?

Traditional Software Stack

What’s the cloud?
” Cloud computing is a model for enabling convenient, on-demand network access to a shared pool of configurable computing resources that can be rapidly provisioned and released with minimal management effort or service provider interaction”*
Is it about the hardware, servers, storage, infrastructure, middleware, application software?
What’s your view?
* Definition from NIST

Isn't this hosting?
You own resources
Pay for what you use and not use
No sharing
No elasticity
Renting resources
Pay for what you use
Sharing
Multi-tenant
Elasticity
Hosting = Cloud =

Cloud Computing: Outsourcing headaches Cloud Provider

5 Key Cloud Characteristics* * As specified by NIST On Demand Self Service Measured Service Rapid Elasticity Broad Network Access Resource Pooling

4 Cloud Deployment Models Public Cloud Community Cloud Hybrid Cloud Private Cloud

3 Cloud Services Model

Cloud Service Provider View IaaS Build Your Own Stack Manage stack Build app Manage app User Concern Provider Concern PaaS Pre-built Platform Managed stack SaaS Pre-built app Use App

App Developer (Cloud Service Consumer) View IaaS SaaS PaaS DevOps Me? How do I standout? Custom apps!

App Consumer View

Middleware in the cloud?
Did we just take the App Server to cloud?
And threw in the database too?
“ I get my own instance of App Server, DB in the cloud and I have a PaaS. Sounds great!”
But is that what it is…

Multi-tenancy
Service Provider View
Resources shared amongst tenants
Economies of scale
Spectrum of multi-tenancy
Nothing multi-tenant to
Each tier of the stack is multi-tenanted

PaaS Reference Architecture Core Platform Services Middleware Services Runtime Services ALM Services Dev Services Provisioning Billing Monitoring SLA Policy IaaS Virtualization/OS Metering App Server Mail Logging DBMS/NoSQL Messaging Auto Scaling CI HA SCM SDK APIs UI IDE

Choosing a solution
Determining criteria…

What is your use case?
Mimic production environment
all about deployment
IaaS view – companies moving upstack
Shorten lifecycle
No more just deployment
SaaS view
Companies moving downstack

Agile or Expansive? Or Both?
Turnkey solution
Rapid development
Rapid deployment
5 Key cloud characteristics
Configuration over Convention
ALM capabilities
Management
Monitoring
Multi-language
Multi-cloud
4 deployment models
Agile Expansive

Criteria for evaluating a PaaS
What friction points does it eliminate?
Am I architecting solutions in a new way?
Am I maintaining tiers of my stack?
Is it IaaS-like or SaaS-like?
How am I paying for the service?
Any add-on services available?
How mature is the service?

Criteria for evaluating a PaaS
Am I locked in to the vendor?
What standards do I care about?
Am I leveraging my existing dev skill set?
Did I compress my dev-deploy cycle?
Am I going faster to the market with better tested and better scalable solutions?

Lay of the Land

Amazon BeanStalk
IaaS view; in beta
Java based. Tomcat; No API restrictions
No vendor lock-in for the application
Auto-scaling, loadbalancing, monitoring
Versioning, management, logging
SDK, Eclipse plugin
Need new service – bake a new server ami and use it

BeanStalk Architecture Provisioning Billing Monitoring SLA Policy Core Platform Services IaaS Metering App Server Mail Logging DBMS/NoSQL Middleware Services Auto Scaling HA Runtime Services ALM Services SDK APIs UI IDE Dev Services Virtualization/OS Messaging CI SCM

Google App Engine
SaaS-like, In preview, final this year
Multi-language (Python, Java, Go)
Restricted Sandbox
Some Java apis supported (JDO, JPA, no JDBC)
Goodbye MySQL, hello DataStore
App URLs served by multiple instances
No streaming data

Google App Engine
Vendor Lockin
Requires to code in a new way
E.g: 30 seconds to respond
Other features
Versioning of apps
Administration console
Integrated with Google Apps
Logging
Eclipse plugin and SDK

GAE Reference Architecture Provisioning Billing Monitoring SLA Policy Core Platform Services IaaS Metering App Server Mail Logging DBMS/NoSQL Middleware Services Auto Scaling HA Runtime Services ALM Services SDK APIs UI IDE Dev Services Virtualization/OS Messaging CI SCM

CloudFoundry
Public, Private and Micro
Multi-language
Java (Spring), Ruby (Rails, Sinatra), Node.js, Scala and Grails
Multi-database
MongoDB, MySQL, Redis
Open Source

CloudFoundry
Multi-cloud:
VMWare vSphere. Option to do AWS through Rightscale & others
Private Cloud: sometime in the future
CLI, SDK, UI (through STS)
No Vendor Lock-in

CloudFoundry Reference Architecture Provisioning Billing Monitoring SLA Policy Core Platform Services IaaS Virtualization/OS Metering App Server Mail Logging DBMS/NoSQL Middleware Services Messaging Auto Scaling HA Runtime Services ALM Services SDK APIs UI IDE Dev Services CI SCM

Redhat OpenShift
Which OpenShift? Express, Flex, Power
Multi-language: which OpenShift?
Ruby, PHP on Express; Java EE & PHP on Flex and C, Linux for Power
Multi-AS: Zend, Tomcat, JBoss – Flex
CLI, SDK or UI: depends on Flex, Express

Redhat OpenShift
Multi-cloud
In Flex: EC2 more later
Private Cloud: No official announcements
Flex Services
Monitoring, Metering
Logging
Ability to go to a service level and configure parameters

Flex Reference Architecture Provisioning Billing Monitoring SLA Policy Core Platform Services IaaS Virtualization/OS Metering App Server Mail Logging DBMS/NoSQL Middleware Services Messaging Auto Scaling HA Runtime Services ALM Services SDK APIs UI IDE Dev Services CI SCM

Key Takeaways
Why is the development stage not in the cloud?
What happens when I take complicated apps to the cloud?

Vendors - Positioning Flexibility Compatibility SaaS-like IaaS-like OS Flex is server-centric Basic orchestration of AWS blocks .Net-on-a-Server, with only … more constraints Spring-centric offering, focused on vCloud JVM-based offering (soon EE) fully server-agnostic Oldest SaaS-like offering but with lots of env. constraints

CloudBees
What we bring to the table?

About CloudBees Our Mission Strategy Become the leading Platform as a Service (PaaS) for Java™ Why We’re Different CloudBees services the complete lifecycle of Cloud application development and deployment. No Servers. No Virtual Machines. No IT.
[email_address] – Cloud Services for Developers
RUN@cloud – Frictionless runtime PaaS for Java apps

Why CloudBees is Better? Manage environment ( recurring ) Metering/Scaling ( recurring )
Remove constant friction between dev and IT
No need to provision masters, slaves
Environment constantly monitored, managed and maintained
Metering/scaling updated in real-time
Business Continuity built-in
Deploy App to traditional Java platform Continuous Integration on CloudBees Build Manage CI Environment: Setup CI Environment Analyse sec. bulletins Analyse QA bulletins Analyse AS/OS logs maintain OS maintain JVM maintain AS maintain DB maintain FW Validate/QA stack Measure App performance Acquire Hardware Provision node Update Cluster Update LB Update Security Install Plugins Manage Plugins Apply Patches  Setup Master Setup Slaves

Why CloudBees is Better? Provision node: Provision cluster: Provision DB: Deploy App: (not a transparent update process) Manage environment ( recurring ) Metering/Scaling ( recurring ) (transparent, multi-version, HA, scalable)
Remove constant friction between dev and IT
No need to provision nodes, clusters, load-balancers, databases, etc.
Environment constantly monitored, managed and maintained
Metering/scaling updated in real-time
Business Continuity built-in
Deploy App to traditional Java platform Deploy App to CloudBees DEPLOY Configure OS Secure OS Setup JVM Install AS on OS Configure AS Secure AS  provision nodes Install load-balancer (LB) Configure LB Secure LB Setup BC for cluster QA cluster Configure DNS  Provision node Install DB Configure DB Secure DB Setup BC for DB  Provision cluster Provision DB Deploy schema Deploy App to each node Analyse sec. bulletins Analyse QA bulletins Analyse AS/OS logs maintain OS maintain JVM maintain AS maintain DB maintain FW Validate/QA stack Measure App performance Acquire Hardware Provision node Update Cluster Update LB Update Security

Why CloudBees is Better? Provision node: Provision cluster: Provision DB: Deploy App: (not a transparent update process) Manage environment ( recurring ) Metering/Scaling ( recurring ) Manage CI Environment: Setup CI Environment DEPLOY Build Continuous Deployment Configure OS Secure OS Setup JVM Install AS on OS Configure AS Secure AS  provision nodes Install load-balancer (LB) Configure LB Secure LB Setup BC for cluster QA cluster Configure DNS  Provision node Install DB Configure DB Secure DB Setup BC for DB  Provision cluster Provision DB Deploy schema Deploy App to each node Analyse sec. bulletins Analyse QA bulletins Analyse AS/OS logs maintain OS maintain JVM maintain AS maintain DB maintain FW Validate/QA stack Measure App performance Acquire Hardware Provision node Update Cluster Update LB Update Security Install Plugins Manage Plugins Apply Patches  Setup Master Setup Slaves

CloudBees: PaaS CloudBees Core Platform Services [email_address] [email_address] SaaS for developers: Repositories, build, test, etc. Nectar Inside Run your applications: Auto-scale, maintain, etc. Core runtime services & back-end services 1 2 3 A B EC2, Rackspace, … Verizon, Terremark, … Enterprises C Public Edition IaaS Providers APIs Private Edition On-Premise Hybrid Edition Mixed - new

CloudBees Platform as a Service Continuous Deployment Production Build Test Provisioning Maintenance JVM – Java EE Java, Grails, Scala, … CloudBees Core Platform Services RUN DEV Repositories Data Services ,NoSQL Metering Billing Monitoring Management Messaging Identity Logging/ Auditing Back-end services Runtime services GIT Maven … SVN Cloud anagement / Virtualization / OS Jenkins APIs Databases Other SaaS On Premise IaaS Virtualization/OS

CloudBees PaaS Provisioning Billing Monitoring SLA Policy Core Platform Services IaaS Virtualization/OS Metering App Server Logging DBMS/ NoSQL Middleware Services Auto Scaling CI HA Runtime Services SCM ALM Services SDK UI Dev Services APIs IDE Mail Messaging

CI in the Cloud
No disruption when code changes break things
“ Working” build always available for test, demos, etc.
Provides critical metrics about development – quality, productivity, functional
Build machines always available when you need them, lowers CAPEX
Scale DUO – Pay for what you use, when you need it
Integration between CI modules reduces dev complexity, lowers OPEX
Seamless integration to runtime PaaS
Continuous Deployment Capability
Why Continuous Integration? Why in the Cloud?

[email_address] Deploy
Leverage PaaS in testing
CLI, HTTP API… PaaS
(MySQL)

Stable
RUN@cloud Features Java/JVM Based Monitoring HA UI 3 rd party integrations Free Offering CloudAnt New Relic Jenkins L&F Mgmt UI CLI, HTTP API, SDK Eclipse. DB Tomcat 4.5k + Apps Starter For FOSS Java EE 6 Web Profile (soon) Grails JRuby Play Coldfusion Scalable MySQL Non JVM Based Node.js POC Scale DUO APIs, SDK Session Persistence

Focus on development-to-deployment cycle (all in the cloud)
DEV@cloud + RUN@cloud
A la carte: allows for 3 rd party integration
Public cloud offering available today
Takeaways

Demo
Taking and managing applications in the cloud

Demo 1
Simple Application to Google App Engine and CloudBees
Bringing in Continuous Integration in the Cloud
Bringing Continuous Deployment to the application
Managing Application on CloudBees

Managing Software from Development to Deployment in the Cloud

by CloudBees

Accessibility

Categories

Upload Details

Usage Rights

1 Embed 1

Statistics