automating infrastructure as a service deployments and monitoring – tec213

24
Chris Kernaghan 14 th August 2012 Automating Infrastructure as a Service Deployments and monitoring – TEC213

Upload: chris-kernaghan

Post on 29-Jun-2015

140 views

Category:

Technology


1 download

DESCRIPTION

Automating Infrastructure as a Service Deployments and monitoring

TRANSCRIPT

Page 1: Automating Infrastructure as a Service Deployments and monitoring – TEC213

Chris Kernaghan14th August 2012

Automating Infrastructure as a Service Deployments and monitoring – TEC213

Page 2: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

SAP and IaaS

Insert "Title, Author, Date"

2© 2012 Capgemini. All rights reserved.

IaaS• IaaS is a Public or Private environment which is billed on a utility

“what you use” model• Defining characteristic of Cloud is self-service• Lots of competing technologies eg.

KVM, LPARs, VMWare, Solaris Containers, EC2, Eucalyptus• IaaS environments are essentially, for individual customers,

infinitely scalableSAP/AWS• SAP have been working in AWS since 2008• Working on Project Titanium for 3 years• Large supporter of Virtualisation technologies• Use in Training, Project and Development systems• Helping partners to build and develop services supporting

Cloud

Page 3: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Supported SAP Cloud Partners

Insert "Title, Author, Date"

3© 2012 Capgemini. All rights reserved.

Total of 63 partners who are certified to support SAP’s supported cloud services.

15 of these partners are Globally certified

Many run their own hosting/Cloud services in addition

Page 4: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

SAP AWS Deployments

Insert "Title, Author, Date"

4© 2012 Capgemini. All rights reserved.

SAP usage across 53 user groups groups

Use of Project Titanium to control instances

Page 5: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Capgemini UK SAP AWS Deployments

Insert "Title, Author, Date"

5© 2012 Capgemini. All rights reserved.

Dec-0

9

Feb-1

0

Apr-1

0

Jun-

10

Aug-1

0

Oct-1

0

Dec-1

0

Feb-1

1

Apr-1

1

Jun-

11

Aug-1

1

Oct-1

1

Dec-1

1

Feb-1

2

Apr-1

2

Jun-

12

Aug-1

20

10

20

30

40

50

60

Template landscape

Customer Project 3

Customer Project 2

Customer Project 1

Demo landscape

Internal Pilot

AWS SAP deployments used by Capgemini UK for Internal and Project use.

Managed by project technical teams or practice members.

Page 6: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Cloud and Shadow IT

Insert "Title, Author, Date"

6© 2012 Capgemini. All rights reserved.

Consultant Cloud Admin

Facilitator

• Help IT administrate Infrastructure

and applications

• Design processes which meet

business compliance standards

• Build/Design flexible

infrastructures which are adaptable

but controlled

• Migrate customers from outdated

platforms to next generation

platforms

• Help business develop and deploy

leading edge applications

• Curate and administrate flexible

landscape

• Additional to project/chargable

work, so primarily done outside

hours

• Team members need to be highly

adaptable, flexible, autonomous,

bought into the SoPs

Page 7: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Automation – Why do it

Insert "Title, Author, Date"

7© 2012 Capgemini. All rights reserved.

S – ScheduledC – ConsistentA – AuditableR – RepeatableS – Sharable

• Systems administration is full of boring, repetitive tasks• Evolve from thick manual run guides to smaller guides on

automated processes

• Create better results driven compliance processes and records

• Quicker serial/parallel executions• The execution of the tasks is more controlled

• There are lots of methods of automation, depending on the layer you are working within• OS – System schedulers• Database – triggers or stored procedures• Application level – scheduled batch or background jobs • Captured input session tools

Ability to execute consistent process to serve multiple compliance frameworks with different documentation requirements – eg, ITIL, SOX, FDA, etc..

Page 8: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

DevOps

Insert "Title, Author, Date"

8© 2012 Capgemini. All rights reserved.

• DevOps is a methodology and philosophy which is used to create and maintain flexible infrastructure and teams

• When used properly it can provide• Consistent information,

• Auditable Processes,

• Automatic Measurement

C – CollaborationA – AutomationM – MeasurementS – Sharing

We are uncovering better ways of running systems by doing it and helping others do it.Through this work we have come to value:• Individuals and interactions over

processes over tools• Working systems over

comprehensive documentation• Customer and developer

collaboration over contract negotiation

• Responding to change over following a planThat is, while there is value in the items on the right, we value the items on the left more.

Cloud depends upon close collaboration due to rapid and shortened development cycles• Flexibility and durations are important as the

costs are directly charged and very visible • Projects tend use smaller teams• Administrators often still have a day job

managing the Production infrastructure

Page 9: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Types of Automation

Insert "Title, Author, Date"

9© 2012 Capgemini. All rights reserved.

• Build automation• Configuration management automation• Change management automation• Scheduled tasks• Exception management automation• Instance management• Security Management• Cost management and reporting

Page 10: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Build Automation

Insert "Title, Author, Date"

10© 2012 Capgemini. All rights reserved.

Activity Manual duration Automation technology

Software Download Hours/days (dependant upon connection)

RSS Feeds and automated download manager

Hardware configuration (Disks, NICs)

Mins/Hours Scripted tasks, Puppet/Chef

O/S configuration (DNS, Users, patches)

Hours Scripted tasks, Puppet/Chef

3rd party software install Hours Scripted tasks, Puppet/Chef

SAPInst (Initial answers) Mins/Hours Answer Files, prebuilt XML files

Build Hours Exception based alerting scripts

Post build steps Hours Scripted tasks executed over RFC

Page 11: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Demo

Insert "Title, Author, Date"

11© 2012 Capgemini. All rights reserved.

Page 12: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Configuration Management

Insert "Title, Author, Date"

12© 2012 Capgemini. All rights reserved.

• Configuration implementation automation can be achieved in a number of ways• Copy from other systems

• System copy• Client copy

• Scripting using SDKs and RFCs• Java• Ruby (Show User creation code)

Runmyjob.comCPS/LVM/UC4

Configuration item Automation technology

O/S and DB configuration Scripts, Puppet/Chef

SAP & Window Users Scripts, LDAP, CUA

SAP RFCs Scripts, Direct SQL Scripts

SAP Number Ranges Scripts, Transports

SAP Transport Landscape Manual

SAP Logical system name Script, manual

SAP Background job scheduling Scripts, 3rd party applications

Page 13: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Change management automation

Insert "Title, Author, Date"

13© 2012 Capgemini. All rights reserved.

Change management applies inside and outside of SAP

Purpose is to accurately record changes for a number of reasons

Infrastructure change management

• Information discovery tools• Belarc• SolarWinds• Squirrel• Configuration Management Databases• Chef/Puppet

SAP change management

• Configuration discovery tools• LiveCompare• Solution Manager• ABAP tools

• Transport management• CTS+, Transport Express, RevTrac

Each automated system for applying change management usually also has a work flow based approval system.

Page 14: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Demo

Insert "Title, Author, Date"

14© 2012 Capgemini. All rights reserved.

Page 15: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Exception Monitoring

Insert "Title, Author, Date"

15© 2012 Capgemini. All rights reserved.

Exception monitoring is valuable because• It identifies and alerts against things which are issues• Can be narrow focus, for example only looking for critical issues• Compliance is often based on exceptions/deviations, therefore exception

based reporting works well with compliance

Using scheduled checks on specific components and an alerting framework• Solution Manager

• SAP Diagnostic agents• O/S scripting

• O/S, DB, SAP checks• CMDB checks

• Configuration management checks

Cloud systems are usually systems created in addition to normal BAU systems

Page 16: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Demo

Insert "Title, Author, Date"

16© 2012 Capgemini. All rights reserved.

Page 17: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Cloud Instance Management

Insert "Title, Author, Date"

17© 2012 Capgemini. All rights reserved.

• Instances are the number 1 cost contributor to the normal AWS EC2 bill• AMI’s are the biggest work load contributor to the EC 2 workload

• AMI ageing and patching• Working with Dynamic resources

• Few AWS resources are static – this leads to interesting design/architectural choices• Use of Elastic IP addresses vs DHCP&DynDNS• Disk subsystem structuring – SAP vs World

Page 18: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Cloud Instance Management

Chris Kernaghan

18© 2012 Capgemini. All rights reserved.

• Instance control evolution has 3 main stages• Always up

• Primarily each instance is up 24*7• Start up and shutdown is manual

• Dumb Schedule• Instances are brought up and down using EC2 API• 3rd Party services

• Default down• Chris Paine’s work with Google Apps provides a website with ability to start

instances

Instance type Typical Runtime Cost

Always up 720 hours/month (24hrs*30) $662/month

Dumb Schedule 450 hours/month (15hrs*30) $414/month

Default Down 120 hours/month (4hrs * 30) $110/month

Page 19: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Demo

Insert "Title, Author, Date"

19© 2012 Capgemini. All rights reserved.

Page 20: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Cloud Security Management

Insert "Title, Author, Date"

20© 2012 Capgemini. All rights reserved.

Security Management is very important for 3 reasons

Flexibility is the mortal enemy of security

Cheaper = Less security because good security costs money

Faster = leaves no time to do security analysis and implementation

Data Protection is paramount, internally and externally

Putting Production quality data in the Cloud is inherently risky

Cloud resources and projects have high degree of flux of Human resources and reuse of technical resources

Cost, utility billing means a direct cost can be placed on each instance

Security can provide details of who is spending what money

Page 21: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Cloud Management

Insert "Title, Author, Date"

21© 2012 Capgemini. All rights reserved.

Cost Management and reporting

• Cost management is vitally important due to perception

• Move the perception to Value and ROI

• Define the KPIs early

• Use Cost calculators to show how much each instance costs

• Use automated reporting

Page 22: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

Evaluation

Insert "Title, Author, Date"

22© 2012 Capgemini. All rights reserved.

FeedbackPlease complete a session evaluation for this session!

Page 23: Automating Infrastructure as a Service Deployments and monitoring – TEC213

| Sector, Alliance, Offering

More information

Insert "Title, Author, Date"

23© 2012 Capgemini. All rights reserved.

Please contact:

• Chris Kernaghan

[email protected]

Page 24: Automating Infrastructure as a Service Deployments and monitoring – TEC213

www.capgemini.com

The information contained in this presentation is proprietary. ©2012 Capgemini. All rights reserved