3. cloud definition: NIST
the ‚IT Light Switch‛
Use Computing power
...
− On demand
− Only for the time
you need it
− Give it back when
you are finished
− Pay for metered
usage
4. cloud definition
attributes of a typical cloud
Re-
Self Purposed
Service Pool of Scalable
Enabled Resource
Charge
Back /
Show-Back
Metered Web
Elastic Usage Enabled
5. private cloud preferred by enterprise
accounts
IDC survey of enterprise accounts
‚Among the
enterprise execs
IDC surveyed, the
so-called
"private cloud"
was preferred
over public cloud
services
by four to one.‛
„ IDC
6. cloud recipe
a sample
All baked in the oven
of governance and security
1. End user catalogue
2. Automated infrastructure
2. Automated infrastructure
provisioning
provisioning mechanism
3. Charge back
7. cloud recipe critical success factor
automated infrastructure provisioning
Includes: Virtualisati
− comput on plays a
e big role
− networ (makes the
k job easier)
Involves 2 types of devices
Automation: Involves 2 types of
− storage
Automation:
−Toolset Automation − Toolset Automation
− People Automation
−People Automation
8. two types of automation
toolset
Three types of
Automation Tools
Involves 2 types of
Automation: Network
Server Storage
−Toolset Automation
Build device allocation
(Physical configuratio (Discs)
−People Automation
and Virtual) n
(Switches,
routers etc)
9. two types of automation
people automation
„ People are the main reason IT projects fail
„ You cannot run a successful cloud service if
there are too many people involved
10. people automation
defined
An automation solution that
does the job that people
would normally do is
sometimes called:
Process Process Run Book
Orchestra Automation Automation
tion
11. process orchestration
defined
Reboot Upgrade Restore
Web Server SQL Server Backup
Unix Network
Clever Linux Boffin Securit
chap Smart y
Bloke Guru
Patch Backup Open
Oracle Exchange Ports
Operation
s
Manager
13. process orchestration
an analogy Process Orchestration
Toolset Automation
Server
Lets’ build a House Carpenter
Management Site Architect
Building site
DataCenter
Network
Plumber
Management
Storage
Management
Brickla
yer
15. the cloud is a journey
ca can support that journey
Yesterda Today Tomorro The
y w Future
BROKER
Internal Virtual Cloud
Private Data Cloud Ecosyste
Cloud centre Brokera m
ge
Today The Coming Just Around
Here and
Now
Big Thing the Corner
?
16. cloud journey
today here and now: internal private cloud
Process
Orchestrat
ion
User
ITIL Securi
Business
Catalog
Toolset ty
Services Automatio
n
$
Availabi Perform Netwo Compute Stora
lity ance rk (P&V) ge
Service
Green
Level
Capacit
17. FY12 sales kick off
grow “ delight “ lead “ unite
Use Case Process and Demo:
Testing and Development of Business
Services
Using an Internal Private Cloud
17
18. Product Boss of Behind the scenes
Developer
Developer
Process Automation
Request Management
Toolset Automation
Receive System
request ITIL Request
Go to Task 1 ”
catalogu Process Email
e and Request Approver
request
Receive Task 2 ”
a Test
s Email Receive
Compute
request approval
r System
for Task 3 -
approv initiate
al Provisioning
Replies
to
Provision Close
email
Server request Server
giving
Receives Automatio
approv
email Load n
al
giving Software Toolset
details
of the
Email User
availabl
e system
Close
Ticket
19. FY12 sales kick off
grow “ delight “ lead “ unite
Use Case Process and Demo:
Elasticity of a Business Service
Using an Internal Private Cloud
19
20. Employee Head of IT Operations Line of Business Manage
Catalog
Business Service Business
Request Operations Service Level
Web Trending
Access
Sub 2
Over
Second
2 Second
Respon
Response Automated
se Provisioning
Company
Web Site
Server Farm
21. FY12 sales kick off
grow “ delight “ lead “ unite
Internal Private Cloud
Pre Sales Support
21
22. moving to a private cloud
the ca cloud workshop: helping customers
estimate ROI
22
23. the ca cloud workshop
not why? but how?
Facilitat Solution TSO
or
Service Definition & Catalog Strategist
Capture
Demo
Discussion Service Definition & Catalog
ROI Tool
Orchestration & Fulfillment Capture
Demo
Discussion Orchestration & Fulfillment
ROI Tool
Monitor ” Measure -
Capture
Manage Demo
Discussion Monitor ” Measure -
ROI Tool Manage
Elasticity Capture
Demo
Discussion Elasticity
ROI Tool
Project
Project Plan
Documen ROI
t
27. ca internal private cloud program
ca automation suite for clouds + services
single services to sponsored industry air
solution implement program cover and
SKU private cloud ensures leadership
in 45 days customer
success
27
28. it’s real, it’s here, it’s now and will continue
launch activities and timeline
Puma/FY12 Release I .0
Planning Launch SKO Release I.0 LAUNCH
Declare GA/
Lighthouse Sales tools, Start Lighthouse Public
Puma Puma Program Design
Demo, PROGRAM Announcement
P0 P1 Enablement
Q1 FY12 Q3 FY 12
Q4 FY11
January Feb March April May June July August Sept Oct Nov
Puma Offsite Launch
Campaign
Messaging/Positioning SALES KICKOFF CA World
GTM Planning Media Tour
PUMA LAUNCH TIMELINE
28
29. continuum of cloud use cases
for ca’s automation suite for clouds
Integrated
Flexible
Unified
Adding Application
Architecture
Release Mgt.
(via Process Automation or
Enabling Grid ARM)
Option
Cloud management
optimization/
vBlock enablement
Service
Orchestration
Solution
29
30. internal private cloud for test and
development
example
“ Virtual
Provisioned
“ 20 Servers for 6 “ Physical
User weeks Seek “ Change Ticket Provisio Provisioned
“ Network
Requests “ Web server layer Approva “ Email
n the Provisioned
“ Seibel system
Service l integration Systems “ Storage
“ Database system allocated
“ Integration
established
“ Notify the
“ Raise Incidents Deliver user
Test the “ Notify the user Charge “ Start the
the clock
systems “ Update Change
“ Email Back
Ticket systems integration
“ Update
“ Notify User before
time Change Ticket Hours
“ Charge
Decommissi “ Offer reschedule End User
on “ Extend or Closure
customer
“ Update
Not
decommission
“ Stop the Clock
Resource
pool
Weeks
31. proof point: ca @ ca
ca technologies own private cloud
‚CA saves $16
million and more
than
25 years of
developers’ time
by automating
provisioning for
Labs on Demand
service.‛
32. proof point: ca @ ca
significant benefits
‚Deploying just one
‘green’ blade server
can enable us to
create 700 or more
virtual machines,
and we typically
have more than
4,000 virtual
servers active at
any time.‛
„ George Watt
Vice President, Software
Engineering
33. proof point: ca @ ca
results by moving to the cloud
„ Lab consolidation from 44 to 4
− Average 700 concurrent, active sessions
− Today ~6500 physical machines and ~ 4000 VMs
„ Cost Savings
− Last year, saved more than 20 years of
developer time and more than $4 million
− Server utilization up to 90% plus
− Server to admin ratio was at 60:1, now at 500:1
„ Real estate savings
− 15,904 sq ft reduction
− $2.4 million projected 5 year savings
− 2,330 machines relocated, ~ 50% virtualized in Q4
„ Automatic power-off between reservations
− 450 metric tons of carbon reduction
Editor's Notes
VocabularySimpleMore at KickoffThe purpose of this pre learning is simply to equip you with the minimum amount of vocabulary necessary to be able to say at least something sensible when faced with a cloud discussion. Almost all of you will know all of this already, so please treat the exercise as a validation of your understanding.For others, I hope you will see just how easy the ideas and concepts are.This is just a starting point and there will be more of the same during the Kick Off, all of it very simple and very easy to absorb.
Because the cloud does not lend itself to a concise definition – an analogy serves us much better.The old country house in the 19th century might well have had an electric oil-fired generator of electricity. It was old, dirty, smelly, prone to break down, very costly, and difficult to maintain. Edison came up with the idea of centralising electricity generation and providing it to multiple homes via an electric switch. So electricity came at the flip of a switch and you only paid for what you used.Most people now believe that this is a parallel for the IT industry and just as the locally-owned electric generator is not common, so the future of the data centre is uncertain.
NIST definitionWhen you ask the business to explain what they expect from IT with regard to their cloud they say .On demandOnly for the time I need itGive it back when I am finishedPay for metered usage
When the business asks for IT to “Deliver new services faster and cheaper” what are the challenges to IT because of the “Attributes” of this business cloud?Self Service Enabled Users drive what they want from a limited set of options in a catalogue. For example, they can have large, medium, or small with an associated service of Gold, Silver, or Bronze. The example here is Salesforce. com vs. Seibel. SFDC won out because it was just good enough but available dynamically.Re-Purposed Pool of ResourceNo longer do we have dedicated systems to an individual business service.This needs careful capacity management like an Airline Booking System that overbooks an aircraft on the assumption that not everyone will turn up.ScalableMust grow quickly as the business growsElasticEach individual Business Service in the cloud must be supported by IT infrastructure that can grow and shrink automatically, or on a schedule, depending on the work load of that Business Service.Charge Back/Show-Back – metered usageThe business will only pay for their metered use of the systems.If there is no internal cross-charging mechanism in place they will at least want to have “Show -back” to decide which departments are consuming the most IT resource – this is part of capacity management and resource planning.
The low hanging fruit for today is Internal Private Cloud for Test and Development.This is where the money is and has been for some time.
The core to a cloud is:End-user focusRemoves people Pay per clickTo put it another way:CatalogueAutomationCharge back
Automation is key.Virtualisation is almost essential – BUT ONLY PART OF THE STORY.We need to understand the difference between Toolset Automation and People Automation.
IT infrastructure consists of three types of generic objects:Server build (Physical and Virtual)Network devicesStorage devicesConsequently we need tools to automate the configuration and build of all three types.
One of the most critical concepts that the cloud depends on is based around how we can remove people from the building and on-going management of cloud solutions:People are too slowPeople do not perform repeatable tasks wellHand-offs between people are problematicPeople don’t audit their activity wellPeople don’t document processes wellEven if people have an automation toolset for configuring the network, servers, and storage, it is still almost impossible to make them use the tools in a coordinated, consistent, repeatable and auditable fashion.
Fortunately the IT industry has known about the requirement for “People Automation” for some time.It goes under various titles, some of which are highlighted here:The oldest name for the discipline is “Run Book Automation”The CA Technologies preferred name, and a more common modern term is “Process Orchestration”
In a data centre operations department there is usually a group of individual domain experts allocated to a particular discipline that they understand to a great depth.When a request comes to the operations group to perform some seemingly straight forward task such as patch the Oracle system, it will in reality, involve a significant amount of coordinated effort between all the domain experts. They will need to perform their individual tasks in specific order with an expected outcome, and then pass responsibility to the next domain expert, and so on.In order to minimise risk, the processes for achieving all the standard tasks expected of the operations group is usually documented somewhere, quite often as a collection of ring binder folders, called a “Run Book Library.”Capturing the Run Book Library as a series of work flows in a Software tool, and replacing the people with direct interfaces from the work flow into the systems being effected, was the origin of Process Orchestration.
This is what a process flow looks like in CA’s Process Orchestration solution from request to provisioning including authorization.
The cloud is a journey:Major enterprises are looking for: Thought leadershipPartnershipThe timeline very much depends on the maturity of the organisation, what some implemented yesterday, others won’t implement until tomorrow.For Ca’s sweet spot – the fortune 500 – what comes first is the Internal Private Cloud usually for Test and DevelopmentWhat is becoming more and more important is the concept of the Virtual Data centreCloud Brokerage is where the MSP’s and Systems Integrators are rushing to But the final end game is when most clouds are external and by their very nature they interoperate
This session is to focus on Today here and now – the Internal Private CloudStep 1: Choice restriction – SLA - CostAll things start with the End User and his or her list of choicesThe user chooses the service they want from the service catalogueThis will be a restricted choice but the user does not care because it will be good enough and delivered rapidlyThere will be an agreed SLA There will be an agreed priceStep 2:The business service is delivered to the user (We will come back later to discuss how that occurs)Step 3: ChargebackWe start the clock so we can measure how long the business service is usedThis is for Chargeback or Show backStep 4: MonitorThe business service can be monitored for 4 thingsStep 5: Availability monitoringIs it up or downStep 6: Performance monitoringFor example is the end user response satisfactory?Step 7: SLA monitoringIs the service about to breach it’s SLAAuto monitoring of SLA’s is a major differentiator for CaStep 8: Green monitoringIs the business service within its approved carbon footprint.This is another differentiator for CaStep 9: ElasticityThe monitoring tells us that the Business service is about to fail it’s SLA and needs more resourceStep 10: PeopleSome of the people that could be required to bring the service back to health include:Change ManagementApproval emailsProvisioning teamsSecurity teamApplication teamDB teamStep 11: No peopleRemove people with Process OrchestrationPeople are the main reason projects failNot repeatableNot auditableHand offs means delaysKnowledge in people’s headsBut cannot automate Fork Lift TruckBut cannot automate waiting for approvalsStep 12: Auto Configuration ToolsProcess orchestration will need to configure the infrastructure equipment and because we do not want to use people we will need configuration automation tools for process orchestration to supervise.Step 12: Network configuration toolCa has very limited functionality here and this is a weakness. Objection handle as follows:Ca does not want to reinvent a wheel because the market is already dominated by proprietary tools such as Cisco WorksWe have a business relationship with Cisco and we use Process Orchestration to make Cisco works behave as if it was a Ca product. The same would apply to Juniper etc.Step 13: Server Configuration ToolThe tool must be able to build and configure Physical and Virtual Servers from bare tin all the way to the full application stack.The tool must be able to handle all operating systems, hardware manufacturers, and hypervisors used by the organisation.Ca built a very powerful tool from the ground up for the modern world dominated by Virtualisation and the Web.Step 14: Storage configuration toolCa does not have a storage configuration tool and this is a weakness. Objection handle as follows:Ca does not want to reinvent a wheel because the market is already dominated by proprietary tools such as Net App.We have a business relationship with Net App and we use Process Orchestration to make Net App behave as if it was a Ca product. The same would apply to Emc, HDS etc.Step 15: Left ArrowThis is a cyclical processStep 16: ITILAutomation can be a dangerous thing unless your processes are managed well and governed.If you automate bad processes you will do bad things very fast and very often.One way to manage the processes is to make sure the Cloud is underpinned and governed by ITIL.Step 17: SecuritySecurity is the biggest inhibitor for organisations moving to the cloudAutomating Identity and access rights to the newly provisioned services is important otherwise you introduce unnecessary delays in delivering the systemsCa has a market leading solution for managing the balance between “Trust and Risk” when delivering cloud solutionsStep 18: Capacity – Airline bookingThis elastic cloud borrows and gives back resource to a resource pool based on the monitoring.The management of the resource pool is a sophisticated problem – usually called “Capacity Management”A resource pool should be managed like an Airline Booking system wher you don’t expect everyone to turn up at the same time.Modelling and predicting the resource pool requirement is a strong play for Ca TechnologiesFINAL STEP – How was the Business Service delivered in the begining?Business service is defined in the catalogue with the associated SLA and cost per dayThe request is managed by process orchestration which will handle the ITIL management of request for approvals and send emails to the approver and requesterProcess Orchestration will then use the Automation Configuration tools to provision the Business Service from the resource poolProcess orchestration will integrate with the security system to add the access rights for the user Process Orchestration will switch on the MonitoringThe system will be delivered to the userThe resource pool capacity will be analysed and reports delivered by Process orchestration to the Resource Pool Capacity Management team.
The next scenario concerns elasticity
In EMEA we have taken the base Xaas system and produced two simple use cases tailored for ease of understanding in the early stages of a sales cycleThe first case is Infrastructure as a service for test and developmentVery quickly click through the process flowThe following are a series of screen shots showing the look and feel of the demoVery quickly click through the screen shot slides
The next scenario concerns elasticity
An employee requests access to a web site that has previously been provisioned, with an associated SLA that comits it to provide sub 2 second response.The Head of IT operations requires a dynamic real time view of the health of the business serviceThe line of business manager requires a business trend and cost impact view over time of the business serviceThe scenario automatically provisions additional resource when the SLA is breachd due to response time becoming in excess of2 secondsclick through the process flow
There is also an value tool that has been builtIt is an interactive tool to be used during conversations with a customer to reveal an ROI for implementing an Internal Private Cloud
In April we are also going to be launching a presales resource currently called the cloud workshop A year ago most organisations need educating as to what the cloud is and why it is importantNow they want to know how to build a cloud and where to startThe work shop is between a half and a full day in durationIt is envisioned initially to be for multiple individuals from the same companyA solution strategist will introduce the idea to the prospect and qualify the opportunityThere is a lead facilitator who runs the workshop and also uses the interactive Value driver toolThe solution strategist will capture the information during the workshopThere is a TSO requirement to deliver live demo’s during the workshopThe demos are based on the 2 use cases of “Provisioning test and development systems” and “Elasticity” that we have just been looking at.
This slide represents the workflow for the workshopclick through the process flow
For test and dev provisioning we can adjust the amount of Toolset automation and Process Orchestration and have a immediate feed back of projected savings (as shown in the next slide)
Blue bar costs have reducedVU meter savings have gone upVU Provisioning time has gone down
And of course we can produce automated documentation of the findings
Private Cloud “0-60 in 45 Days”Focused, Pre-Packaged Solutions for Private Cloud/IaaS Use CasesBest in Class – Integrated Technologies from CA and PartnersRapid Implementation Methodology – 4-8 weeks Reference Architecture of Certified Platform (*VCE & x86)Complete Services Package (Soup to Nuts – Delivery)Quantified/Fast Time to Value/ROI Integrated into IT Operations eco-systemFoundational Solutions for advanced customizationsOn ramp to integrated hybrid clouds
The official launch is due his month
They are building a road map of enhancements as the market place matures
This is based on a real use case where a major Pharmaceutical needed to provision a Siebel system for test and developmentThey were receiving 2 requests per week for Test systems to enhance Seibel.The process involved 5 departments in 3 countries and many levels of approvalThey were able to provision a single system every 8 weeksAt Ca we provision systems in minutes
We eat our own dog foodAround 4,000 systems being used, built or dropped at any one timeAround 30 minutes from request to delivery of a systemCa’s business is Software Development and we have no fall back option to our own Internal Private CloudWe have bet the company on our own Internal Private Cloud
Pick out the underlined and commit to memory for any elevator pitch you may need