iseries Recovery Options Pro s & Cons Presented by: David Tedford Lakeview Technology, Inc. tedfordd@lakeviewtech.com 978-671-9211 office 617-901-7909 - cell
Presenter background 1987, BS in CIS, Quinnipiac University 18+ years of IT industry experience 10+ years of software development/applications programming experience 2002, MBA, University of Massachusetts 8+ years of high availability experience Solution management Project Management Channels Management Sales Management Certified iseries Sales Specialist Certified High Availability Sales Specialist
Agenda: High Availability Market drivers Define RPO, RTO, ROI Calculating Cost of downtime Available DR Solutions ROI for your DR solution Some examples Q&A
Facts: 87% of businesses experience unplanned downtime each year. 50% of business experience up to 5 unplanned outages each year The average number of days per outage is 3 90% of businesses go bankrupt within 2 years of a significant outage - Financial and Functional Impact of Computer Outages on Businesses, University of Texas
Availability Trends Business initiatives drive higher levels of availability ebusiness On Demand Enterprise networks Globalization of enterprise operations Business competition Increased hours of operation / customer service Government regulations HIPPA Remote computing Support of mobile and home-based employees
Traditional iseries Availability Market High Clustering Characteristics of the solutions is based on: RPO/RTO Cost of Ownership Skill Level Environment Complexity High Availability Disaster Recovery Tape/Restore Low Level of Information Protection High
Environment Review. Mission Critical Application Planned downtime Unplanned downtime Cost of downtime
Two Types of Downtime Availability Unplanned Downtime Disaster Recovery Reliability Issues Planned Downtime System Backups Hardware Upgrades Application Upgrades O/S Upgrades File ReOrgs
HA Requirements Define Business & Technical requirements Define Sponsors, Influencers, decision makers, contact info, etc. Define systems, locations, applications Current & desired RTO & RPO Existing DR solution Procurement and implementation timeframes Fiscal year, budget $, approval HA and H/W competitors Impending events
What Are the Components of Availability? Recovery Time Objectives (RTO) How long can you tolerate your applications being offline? Recovery Point Objectives (RPO) How much data loss can you tolerate? RPO RTO
What Then Is Optimum Availability? Optimum (op te-mem) noun, adj: The best or most advantageous condition, degree or amount Your Recovery Point Objectives (RPO) Your Recovery Time Objectives (RTO) PLUS your Optimum Return on Investment (ROI) Startup Costs Recurring Costs RPO Skill & Personnel Investment Operations, Administration and Infrastructure ROI RTO
What s it costing the business when their systems are unavailable?
Impact on Cost of Downtime Tangible Costs Intangible Costs Lost Revenue Lost Wages Lost Inventory HIPAA Violations Legal Fees Lost Opportunity Employee Retention Goodwill Brand Damage Patient Respect
Why Availability Average Hourly Cost of Downtime Telephone Ticket Sales $ 69,000 Manufacturing $ 80,000 Airline Reservations $ 89,500 Retail Catalog $ 90,000 Infomercials / Promotion $ 199,500 Retail Banking $ 1,000,000 Retail Brokerage $ 6,500,000 Contingency Planning Research
Industry vs Needs
50 Million People in the Dark Blackout of 2003
Picking the Wrong HA Solution Could Give You Nightmares
The IBM iseries Is the Most Reliable Server in the Industry! 99.94% Platform Availability < 5.2 hours per year True, but 95+% of all downtime is planned!
Planned Downtime Events other than disasters disrupt availability much more frequently. Hardware failures, localized power failures and so on are still rare but more frequent than disasters. Maintenance may halt applications daily: Database saves. Database reorganizations. Software upgrades. Hardware upgrades. General system maintenance.
Minimal Solution: Off-Site Protects data availability, not application availability. Requires long recovery times. Tape Backup tape sent off-site nightly. Advantage: low hardware and software costs. Disadvantages: Leaves Orphan Data unprotected. Requires disruptions during saves. Provides value only during a disaster.
Better Solution: Data Vaulting Send backup tape off-site nightly. Continuously replicate updates, but don t apply them. Discard yesterday s updates when new tape arrives. When a disaster strikes: - Load the newest backup tape. - Apply the updates.
Better Solution: Data Vaulting Advantages: Can share facilities in 3rd party Disaster Recovery site. Avoids orphan data problem. Disadvantages: Still takes a long time to recover. Only provides value in the event of a disaster.
Best Solution: Remote redundancy (1) Replicate databases off-site in near real-time. Maintain a fully redundant copy of data and objects on a hot-standby remote system. Perform database saves at the remote site. If disaster strikes, switch users to the remote site.
Advantages: Best Solution: Remote Near instant recovery. Recover right to the point of failure. Provides robust Continuous Operations and High Availability. redundancy (1) Disadvantage: Must maintain two fully operational sites.
Best Solution: Remote redundancy (2) Same as remote redundancy (1) except: Split each physical system into 2 logical systems using LPAR. Production. Backup. Split the application load between two systems. Use the backup partition on one system to back up the production partition on the other.
Best Solution: Remote redundancy (2) Advantages: All of the advantages of remote redundancy (1). Provides higher performance by splitting the application load between two systems. Caveat: You must either accept less performance when one system is offline, or size both machines to handle the full load. LPAR
Replication: A Closer Look at Asynchronous RJ Remote Journaling Replication Process: Asynchronous
Replication: A Closer Look at Object Replication
HA Option Summary: Solution Option: Pros: Cons: Backups Hotsite High Availability software Home Grown Application Cheap Solution Part of Normal Operations Provides DR Provides DR & HA ROI for planned outages Allows regular testing Proven technology Environment Specific Own Intellectual property Slow RPO Data Loss Tape Failures Uses backups Site Location unknown FIFO Lack of testing Perceived Cost System requirements Comm requirements Has HA con s Ongoing dev/maint New technology improvements Cost
Availability Solution Options Backup to tape 3 rd Party Recovery Sites Storage Area Network (SAN) High Availability Software (MIMIX)
Requires staff to operate and manage backups May require downtime system backups Difficult to test recovery requires downtime or backup hardware Re-entry of lost data may be required Backup to Tape
3 rd Party Recovery Site Based on tape backup practices Travel Expenses and time to test recovery process Focused on system recovery not business and application
Storage Area Networks Requires downtime for failover testing High initial cost Can only be used for recovery
High Availability Software Daily operations and monitoring can be completely automated & integrated into IT operations Integrated into all business systems / applications. Easy to audit and test requires very little downtime
Case Study Simple MRP Has only one iseries system Performs nightly application backups and weekly system backups Application downtime to perform weekly backups Has no means to test backup quality and integrity Tape
Case Study Simple MRP Partition the server or add 2 nd iseries Eliminate backup downtime & staff costs Higher Quality Backups / Quicker Recovery Perform Application maintenance without downtime MIMIX Partitioned iseries Server Tape
Case Study Simple MRP Test backup recovery on 2 nd partition Faster Recovery at 3 rd Party Recovery Sites No off site testing expenses test on partition MIMIX Partitioned iseries Server 3rd Party Recovery Site Tape
Case Study Complex ERP Single iseries Server & Application Server Environment Application backups performed nightly and weekly SAN used for faster recovery of production system 3 rd Party Backup Site for Disaster Recovery User User User SAN Tape
Case Study Complex ERP Application downtime due to backups Backup s are resource intensive Application recovery at 3 rd party site takes too long SAN recovery requires IPL SAN has low cost of ownership User User User SAN Tape
Case Study Complex ERP Higher Quality Backups performed during daytime hours 3 rd Party Recovery Site no longer needed No off site testing expenses test on backup server or partition Solution based on application, not server scalability Backup iseries Server MIMIX User SAN Tape User User
Case Study Complex ERP Backup iseries Server Ability to test recovery on your own servers Application load balancing reduces upgrade expenses MIMIX Tape User SAN User User
Benefit From HA & DR software Without a Disaster! Eliminate virtually all downtime from: Planned maintenance. Hardware failures. Disasters. Thereby: Increase revenue. Improve productivity. Enhance customer satisfaction and service. Reduce recovery costs. Reduce other expenses.
MIMIX Makes it Possible! High-performance, continuous replication over any TCP/IP network. Enables data vaulting and redundantremote solutions. Monitors primary system availability. Automatically switches users (failover) to the remote site when the primary system becomes unavailable due to a hardware failure or disaster. Manually switches users to the remote site during maintenance activities.
Optimum Availability Solutions: MIMIX DB2 Data Replicator Object Replicator ClusterServer Promoter Monitor Outage RPO RTO App Planned & Unplanned Seconds to Minutes Seconds to Minutes Native OS/400 DB2 Replicator Object Replicator Monitor Outage RPO RTO App Planned & Unplanned Seconds to Minutes Seconds to Minutes Native OS/400 Outage RPO RTO App Planned & Unplanned Seconds to Minutes Seconds to Minutes Windows DR1 Outage RPO RTO App Unplanned 2 4 Hours 12+ Hours Native OS/400
Lakeview Technology Experience & Choice 14 years experience Not just iseries HA Mimix iseries, OS/400 Echostream IBM, HP, Apple, Unix, Linux.. HA Cluster IBM, HP, Apple, Unix, Linux. Mimix for Windows xseries, W2K & Advanced Server HACMP pseries, AIX HA Omni Cross-platform Data Replication solutions Network1 Data Optimization Global support & operations
Sleep at Night by Using MUST HAVE Criteria with which to Evaluate HA Solutions & Providers 1. Assessing HA Solution Functional Attributes 2. Assessing HA Solution Provider Abilities & Track Record
Summary A MIMIX Disaster Recovery solution ensures: Faster recovery. Less lost data. In addition, it simultaneously provides robust High Availability and Continuous Operations. Thus, MIMIX provides a significant Return on your Disaster Recovery Investment, even if you never experience a disaster!
Thank You! Questions and Discussion