Click here to proceed.

1
The Changing Face of Business Continuity in the
21st Century
James R. Hackett Lakeview Technology, Inc Area
Manager North Central
2
Agenda

• What is HA DR
• Defining RPO RTO
• Business Drivers DR Trends
• Approaches and Methodologies
• HA and ROI, mutually exclusive?
• Solutions from IBM and ISVs

3
Business Continuity (BC)
24x7x365

• Capability to withstand outages
• Planned or Unplanned
• Uninterrupted Mission Critical Service
• Pre Defined Service Level Agreements
• Primary Solution Sets
• DISASTER RECOVERY(DR)
• Resources, plans, services, and procedures used
  to recover mission critical applications and to
  resume normal operations for these applications
  at a remote site
• Stated disaster recovery goals
• Certain levels of degradation are acceptable
• HIGH AVAILABILITY (HA)
• Provide continuous processing for mission
  critical applications during planned unplanned
  outage events
• Permanent regular switching of roles between
  backup and primary resources and remaining so for
  an extended period of time (weeks verses hours)
• More demanding RTO and RPO (as compared to a DR
  solution)
• Capable of fully automated failover

4
High Availability DR Definitions

• High Availability (HA) Solution
• Capable of providing application data
  resiliency with minimal impact to the application
  user (assumed that the underlying infrastructure
  for data, power, communications etc are
  resilient)
• Can provide a recovery point objective (RPO) of
  the last application commitment boundary
• Can provide a recovery time objective (RTO) of
  less than 10 of the time for a single system RTO
  (IBM) when best practices are used
• Planned and unplanned outages can be a fully
  automated and transparent to the production
  environment
• A solution environment where at IT operations
  discretion, primary and backup resources exchange
  rolls regularly and for an extended period of time

5
High Availability DR Definitions

• Disaster Recovery (DR) Solution
• A solution primarily focused on data and
  application recovery where recovery may involve
  many hours verses minutes.
• A solution where the RPO may be hours (i.e.. tape
  recovery)
• A solution that typically requires remote
  location recovery
• A solution that is normally not used to mask
  disruptive maintenance procedures from the end
  application user
• A solution topology where the roles of the
  primary and back up servers are static (regular
  role switching is not a requirement)

6
RPO and RTO - Definitions

• RPO is the targeted tolerance threshold for lost
  business transactions. It defines how much data
  and objects can be lost or how far you can go
  back in time to resume application functionality.
  Do you need to be as near to now as possible?
• RTO is the targeted amount of time required to
  resume and recover application functionality.
  Focus should be on how fast the application can
  be restarted. and brought online.

7
IT Priorities

• 79 of IT managers say Business Continuity the
  Top Priority

Q3a
of 4/5 responses (important /extremely
important)
Source IBM it Trends 1H 04
8
Business Profile
Spending for HA is growing faster than spending
on the server market as a whole (Source Sept
2003 IDC 3008)

• Drivers
• Internet Competition is just a click away, but
  so is the hacker with a malicious virus or worm
• Just in time and real time operations, SLAs.
• Asset Utilization 24x7 operations
• Consolidation (data centers, servers,
  operations)
• Global operations or just across time zones
• Corporate Governance - i.e. SOX, HIPAA

9
Business Profile
Five Top Consequences of a Disaster (Source
Database Trends, May 2005)

• Decreased employee productivity (62)
• Data loss (43)
• Reduction in profits (40)
• Damage to customer relationships (38)
• Reduction in revenue (27)

10
Business Profile
51 Percent of companies surveyed actually had to
execute their Disaster Recovery Plan in recent
years. Another 45 Percent admit they do not know
how long it would take to get back up and
running. (Source Database Trends, May 2005)

• Most Common Incidents
• Computer system failure (37)
• External computer threats (35)
• Natural disasters and fires (14)
• Accidental and malicious employee behavior (13)
• Average Projected Recovery Time
• 1. Over three days (3.23)

11
Disaster Recovery A Grudge Buy or Providing
ROI?
Emerging Business Continuity (BC) Value Based

• The fact that most organizations are unlikely to
  ever use the full extent of the services they
  have paid for has in the past made disaster
  something of a grudge buy and not something
  that most companies are eager to spend money
  on.
• From ITWEB
• 25 September 2001

12
Disaster Recovery Bottom Line or Bottomless
Pit?
Emerging BC Value Based

• Recovery services dont add anything to the
  bottom line, but the consequences of not having a
  plan in place can be disastrous.
• Dave Linacre
• Managing Director
• IBM Business Continuity and Recovery Services
• from ITWEB
• 25 September 2001

13
Is Your BC Solution an Expense?

• Can You Predict Your Next UNPLANNEDSystem
  Outage?
• Can You Predict the Durationand Amount of
  Losses?
• Can you place an ROI on Your BC Solution?
• Does Your BC Solution Provide Benefits or
  Alternate ROI Without A Disaster?

14
Or, Is Your BC Solution Also an Investment?

• Can You Predict Your Next PLANNEDSystem Outage?
• Can You Predict the Durationand Amount of
  Losses?
• Can you place an ROI on Your BC Solution?
• Does Your BC Solution Provide and Benefits or ROI
  Without A Disaster? Yes, by eliminating the
  PLANNED Downtime

15
Emerging BC Value Based

• Backup Elimination
• Perform backups without downtime.
• Single or Dual System Solutions
• Single or Dual Dataset Solutions
• Parallel Processing / Parallel Dataset
• Simultaneously run the application on multiple
  systems / locations
• The Data is maintained on all systems
  simultaneously
• Storage and Storage Virtualization
• Combined w/ Capacity On Demand maximizes IT
  Infrastructure Utilization and Lowers Costs

16
Traditional iSeries Availability Market
CLUSTER
Recovery in Hours to Minutes
Enterprise and SMB Markets
HIGH AVAILABILITY
SMB Space Large Gap Costly jump to HA
RPO
IBM BCRS / QUICK SHIP
Recovery in Days to Hours
TAPE BACKUPS
RTO
Recovery in Hours to Minutes
Recovery in Days to Hours
17
Current iSeries Availability Market
0
CLUSTERING
Enterprise Market
HIGH AVAILABILITY
minutes
SMB Market
DISASTER RECOVERY / PRACTICAL AVAILABILITY
SOLUTIONS
RPO (data loss)
1-4 hours
12-24 hours
BCRS / SUNGUARD QUICK SHIP
12-24 hours
TAPE BACKUPS
1-2 weeks
2-3 days
4-10 hours
minutes
0
RTO (application downtime)
18
Prices Making High Availability Ever More
Affordable!
i870 Config Before 04-20-04 System
Specifications 8/16 way (8 active) 11,500 CPW 24
GB Memory 8.4TB Disk 3481-L23 Tape (9,086) HW
List Price 2,543,934 Price Per CPW 221 SW
Tier P40 SWMA 44,403 List MMMC 7,230
i870 Config After 04-20-04 System
Specifications 8/16 way (8 active) 11,500 CPW 24
GB Memory 8.4TB Disk 3481-L23 Tape (9,086) HW
List Price 1,920,774 Price Per CPW 167 SW
Tier P40 SWMA 44,403 List MMMC 7,230
i5 570 Config on 06-11-04 System
Specifications 3/4 way (4 active) 11,700 CPW 32
GB Memory 8.4TB Disk 3581-F28 Tape (11,420) HW
List Price 1,477,168 Price Per CPW 126
(-43 from April 20th) SW Tier P30 SWMA
27,372 List MMMC 4,007
19
How HA Solves Business Needs

• Get High Availability with Disaster Recovery
• Replicate locally between two servers for
  Continuous Operations during
• Planned maintenance
• OS hardware upgrades
• Disk failure
• Replicate to a remote location for added Disaster
  Recovery protection against unplanned downtime
• Site loss

Production iSeries server
Backup iSeries server
Tape Backup, Maintenance, BI/Analytics,
Development Without Downtime
20

• Planned Outages
• Estimated to be 95 of all i5, iSeries and
  AS/400e outages
• Types of Planned Downtime
• Daily/Weekly/Monthly Saves
• Software installation/upgrade (OS, Application,
  Middleware, etc)
• PTF installs
• Operating Systems Upgrades
• Hardware Upgrades

• Unplanned Outages
• Estimated to be lt5 of all iSeries outages
• Unscheduled/Unplanned Downtime due to
• Power outage
• Human error or program failure
• Unprotected DASD or multiple DASD failure
• Other hardware failure
• Disasters
• Estimated to be lt1 of all iSeries outages
• Site wide problem, weather, virus/worm, terrorist
  related

21
LPAR Replication/Intra-Server

• Eliminate Most Planned Outages (most of 95)
• Daily/Weekly/Monthly Saves
• Software installation/upgrade (OS, Application,
  Middleware, etc)
• PTF installs
• Operating Systems Upgrades

Production iSeries with Logical Partitions
(LPARs)
Intra-server replication eliminates planned
downtime with no additional hardware investment
3
2
1
22
Local Replication Between Two iSeries

• Eliminate ALL Planned Outages (95)
• Daily/Weekly/Monthly Saves
• Software installation/upgrade
• PTF installs
• Operating Systems Upgrades
• Hardware Upgrades
• Eliminate Most Unplanned Outages (Most of 4)
• Power Outage (if different power source)
• Human error or program failure
• Unprotected DASD or multiple DASD failure
• Other hardware failure

Production Apps Keep Running
Maintenance work performed on the Backup Server,
then roll swap with the Production server
Maintenance Without Downtime
23
Local Replication and Remote Replication

• Eliminate ALL Planned/Unplanned Outages AND
  Disaster Recovery
• 100 of Downtime!

Datacenter 1
Production

• Daily/Weekly/Monthly Saves
• Software installation/upgrade
• PTF installs
• Operating Systems Upgrades
• Hardware Upgrades
• Power outage
• Human error or program failure
• DASD failure
• Other hardware failure
• Site wide problem

Backup
Datacenter 2
Remote Backup
24
(No Transcript)
25
Solution Topology For DR The More Common
Deployment
Requirements Feasibility of Approach Dictate
Deployment Approach
High Availability or CBU Server
Primary Server
Disaster Recovery Center
Data Center
Remote recovery outage support topology for
relatively static environments Temporary
roll-swapping done periodically, Save window
reduction rarely done, Work load distribution
rarely done.
26
Solution Topology for HA The Best Practice
Capacity BackUp Server
HSL Loop
High Availability Server
Primary Server
Disaster Recovery Center
Data Center
Data Center Regular permanent roll swaps, save
window reduction, workload distribution Disaster
Center Relatively static and dedicated for
disaster recovery scenarios
27
Implementing HA is A Collaborative Process within
an Organization
28
Business Resiliency Requirements

• BUSINESS REQUIRMENTS
• Business Impact Analysis
• Establish Existing Planned Outage Profile
• Asset utilization (planned outage unavailability)
• Customer Impacts
• Develop Worst Case Outage Profile's
• Financial and Customer Impacts
• What-If Scenarios
• Business Benefit Analysis
• Establish a Financial cost/benefit assessment
• HA can pay for itself !
• Risk Assessment
• Probability Versus Risk Aversion

• IT REQUIREMENTS
• DR Solution, HA Solution, Combination
• Planned Outage Transparency
• Backups Maintenance Procedures
• Role Swap Objectives (for an HA Solution)
• Workload Balancing Objectives
• RPO, RTO
• Recovery Point Objective (last transaction or
  greater)
• Recovery Time Objective (minutes, hours..)
• Data Center, Disaster Center, Combination
• Solution topology will be driven by requirements
  and cost
• Consider Offering i5 Check
• IT Resiliency Profile

29
Company Drivers As Business Objectives Expand,
IT Spending Increases.
Relation to DR HA Technologies

• Business Objectives
• 24X7 Operations
• Best of Breed
• Acquisitions/Expansion
• Cost Expense Cutting
• Status Quo

Enterprise HA/DR
SMB HA Solutions
Tape Backup

• IT Capabilities
• Stage 1 Website
• Legacy Systems

• Stage 2 or 3 Website (Transaction processing)
• Vendor Data Exchanges
• BI Analytics

30
Customer Sample Preliminary Availability
Justification

• Questions
• Company Annual Revenue
• Company Annual Profit
• Planned Annual Downtime
• Prior Year Unplanned Downtime
• Average concurrent users supported?
• Average User Annual Compensation (25/hr)
• Estimated employee DT productivity ()?
• Calculated profit loss during DT (per hour)
• Hours of Operation per day?
• Days of Operation per year?
• Hours of work per employee per week?
• Weeks of work per employee per year?
• Prior Year UPDT Additional Costs
• - Overtime (15K/hr) , Penalties, External
  Support, Emergency Plan Activation

Downtime (DT) Factors Costs Annual Revenue Per
Hour Annual Profit Per Hour Planned Downtime Cost
(Profit) Prior Year Unplanned Downtime Cost
(Profit) Outage Costs Based On Revenue Unplanned
DT Salary Costs (400 x 12.5 x 76) Agreed Upon
HA Costs - Unplanned (150K 433K
380K) Additional Justification (Business
Initiatives) - Business Intelligence Project
(400K) - Developer Server Partition
(200K) TOTAL JUSTIFICATION
750M 50M
86K 5,700
32 hrs
182,000
76 hrs 400
433,000
9.3M
50,000 50
380,000
5700 24
963,000
365 40
600,000
50 150,000
1.6M
How Far Can You Stretch Your Service Levels
Without Planned Availability!
31
The Solution Choices

• The Availability Building Blocks
• Basics Raid5 and Mirroring
• Static DR Implementations
• Dynamic HA Implementations
• The Solution Choices For Data Resiliency

32
The Availability Building Blocks
Continuous Availability
Application Resiliency
Data Resiliency IASPs, HA Replication
Disk Replication, IASPs FlashCopy PPRC
Transaction logging Journaling
Disk, HW, Data Protection Mirroring
Disk Protection Raid 5
Backup Recovery and Tape Automation
33
Disk Mirroring Solution Package for Integrated
DisksProvide Protection against Single Points of
Failure
34
iSeries Multi System Data Resiliency Solutions
Remote Journaling
Logical Replication
Switched Disk IASPs
Tape
ESS with IASPs
XSM (Cross Site Mirroring)
35
iSeries Business Continuity Offerings

• IBM HA Express Portfolio on iSeries
• Models 520, 810 825
• Support mission critical 24x7 environments
• Minimize planned and unplanned downtime
• Minimize save window
• Lower complexity and cost

• iSeries for High Availability Offering
• Models 570, 870 890
• Support mission critical 24X7 environments
• Role swapping and workload distribution
• Minimize planned unplanned downtime
• Support Heterogeneous Environments

• iSeries for Capacity BackUp Offering
• Designed for disaster recovery scenarios
• Not intended for 24x7 HA solutions
• Not intended for workload distribution
• No option to permanently activate standby
  processors

36
Logical Replication Solutions Vendors For HA
Logical Replication
37
Logical Replication Solution Vendors For DR
Logical Replication
38
Heterogeneous OS Mixed Technology Solutions
Replication, Switchable Resources Multiple OS
Environments
39
ISV Solution Coverage
IASPs
Windows
I5/OS
Linux
AIX
DS
Data Mirror
iTera
Lakeview Technology
NoMax
Traders
Vision Solutions
40
Data Center
Hybrid Cluster Data Center SW Disk with
Replication Off-Site for DR
Production Server
Provides Overall Replication
Management and Control
Replicate Backup eServer i5
DR Site
Backup Server
41
Summary

• HA is a Cooperative Implementation
• Business Units Standards Drive IT Capabilities
• BC must add value to achieve investment in the
  short term
• BC must integrate with the business units goals
• IBM solutions portfolio feature a rich set of
  business continuity and high availability
  solutions
• ISV Solutions, IBM Hardware Solutions
• HA CBU Servers

42
Additional Resources

• iSeries HA External Website
• Several HA articles available such as HA 101
• www.ibm.com/eserver/iseries/ha
• iSeries HA Sales Kit
• http//w3-1.ibm.com/sales/systems/portal/_s.155/2
  54?navIDf220s240geoIDAllprodIDiSeriesdocIDi
  hask.skitdocTypeSalesKitskCatDocumentType
• iSeries for HA/CBU
• http//www-1.ibm.com/servers/eserver/iseries/hardw
  are/is4ha
• http//www-1.ibm.com/servers/eserver/iseries/hardw
  are/is4cbu
• IBM Redbooks
• i5/OS High Availability Clusters Data
  Resilience Solutions
• http//www.redbooks.ibm.com

43
Questions? Thank you!
44

Back Up Charts
45
Cross Site Mirror (XSM) for remote HA/D/R
iSeries Cluster Device Domain
Open TCP/IP Network
WAN
TCP/IP Network
TCP/IP Network
Sync to memory
Optional Sync to disk
Concurrent I5/OS writes
iASP-A
iASP-A
46
Data and Disaster Center Availability with
ESS-PPRC (with PPRC-Toolkit)
SPCN
Prod. iSeries-A or LPAR-A iASP-A varied on
Switchable Tower(s) A
ProdESS
iASP-A
HSL
Fiber(s)
HA iSeries-B or LPAR-B iASP-A varied off
SysBas
TCP/IP Network
FC / ESCON
Backup iSeries or LPAR
iASP-A
Tape Backup
47
Protect Your AIX and Linux Applications
Switched Disk Solution Using ESS
Backup AIX/Linux Partition(s)
Production AIX/Linux Partition(s)
48
AIX Data and Application Protection with HACMP
and HACMP/XD
Lakeview offers the skill service capability
To deploy HACMP and HACMP/XD
Backup AIX Partition
Production AIX Partition
HACMP for Control/Switching
HACMP/XD for Replication
Direct I/O
Direct I/O
49
Protect Your AIX and Linux Applications and
Data and
Replication-based Solution Using Virtual I/O
EchoStream FS
Backup AIX/Linux Partition(s)
Production AIX/Linux Partition(s)
50
Product Offerings
51
iSeries Solutions for Business ContinuitySmall
and Medium Business
IBM HA Express Portfolio on iSeries Models 520,
810, 825
Support mission critical 24X7 environments Role
swapping and workload distribution Minimize
planned downtime Minimize unplanned
downtime Save window reduction
52
iSeries Solutions for Business ContinuityLarge
and Medium Business
iSeries For High Availability Models 570, 870
890
Support mission critical 24X7 environments Heterog
eneous OS Environments Role swapping and workload
distribution Minimize planned downtime Minimize
unplanned downtime Save Window Reduction
53
iSeries Solutions for Disaster RecoveryLarge
and Medium Business
iSeries For Capacity BackUp
Models 570, 825, 870, 890
Support For Disaster Recovery Environments
54
IT Resiliency Profile Analysis

• Achievability of current SLA
• Availability Coverage
• Single Points of failure
• Interdependencies
• Ownership and Process
• Skills and Staffing

• Develop a Resiliency Baseline
• Resiliency Gap Analysis
• Plan of Action

CTC (HA consulting) http//www-1.ibm.com/servers/
eserver/services/havail.html
55
1H05 Events, Offerings and Promotions (AG)

• Rochester HA CIO Fly-ins
• Quarterly Business Continuity CIO Fly-Outs
• Regional Cities
• Quarterly ISV Webcasts
• Availability Assessments On Demand
• Buy One Get an HA Server Free

56
iSeries Technical Sales Support, Americas
57
(No Transcript)