Cutting TCO in Half
Innovation Award Podcast Series: Royal London Group
Please install Adobe Flash to listen to the Innovation Award Podcast Series: Royal London Group.
Royal London migration saves $22.9M in Total Cost of Ownership (TCO) and increasing availability
For decades data-intense computing has been done on the same systems. But cost savings opportunities are driving many IT specialists to consider migrating to open systems-based environments and completely refreshing their IT infrastructure. Although hearing “migration” makes IT departments wince, many IT professionals are contemplating and carrying out the task despite the work involved.
UK pensions provider Scottish Life, part of the Royal London Group, relied on systems running ten million lines of code and managing nearly a billion data records. Launched in 1993, the previous system was struggling to maintain service levels. Internal staff & external users experienced a growing number of interruptions as their on-line service requests competed against transaction and background batch processing. Overnight batch jobs frequently ran longer than twelve hours forcing scheduled and ad hoc business processing to be postponed until the weekend.
The previous systems took five years to develop and their software was continually upgraded to respond to changing business needs. Operations tasks, including schedules, printing, performance monitoring and event management, were highly automated. Up to 1800 batch jobs ran nightly with over 64,000 business rules governing the distribution of up to 30,000 print images each day. A performance monitoring and intervention system supported the activities of more than five hundred users, and a remote second-site disaster recovery system was kept in constant lock-step.
Before deciding to migrate to an open system environment, Scottish Life assessed the complexity of moving its massive infrastructure - databases, business rules, operational activities and disaster recovery – and sought the help of MSS International*, a migration specialist, along with Micro Focus*, Oracle* and HP*. In addition to requiring a considerable amount of customization, they needed to replace system software packages with a set of completely new operational support utilities.
Challenge
• Lower IT operating cost. Replace the IT infrastructure to reduce hardware, support and maintenance costs.
• Improve the user’s service level. Make the computing systems more responsive to users and reduce the time required to perform nightly batch jobs.
Solution
• Migrate to an open systems architecture. Implement lower cost operating systems, servers, software and operations tools.
• Deploy scalable server cluster. Run enterprise software on Intel Itanium 2-based HP Integrity servers and PA-RISC-based HP 9000 servers.
Migrating to Open Systems-based Architecture
Scottish Life assembled a diverse team to migrate its Enterprise Application Environment (EAE) consisting of legacy code, data, system software and disaster recovery. MSS International provided translation software to automatically convert programs in proprietary languages and legacy COBOL code into Oracle PL/ SQL, C & Micro Focus COBOL, which can run on any open platform. The pension software application was moved to Cobal85 and data records were converted into an Oracle 9i database. Due to the unique nature, size and complexity of the previous systems, all business processes required testing by both IT testers and business users.
HP and Oracle supplied new system software and operational support utilities, including operating system and database software. Operations staff received comprehensive training, preparing them to setup the new software environment and later to provide ongoing IT support.
Oracle Data Guard configured a disaster recovery clone on the WAN-accessible development system located about two hundred miles from the production system. Other clones were created as part of the application development effort for performance testing, troubleshooting and data mining.
The project included producing new innovative products and modifying software to support new legal requirements, completed by a team of a hundred individuals. This development program ran in parallel with the migration project and both were simultaneously moved to the open systems-based cluster without interruption to the business.
An HP Integrity rx8620 with twelve Intel Itanium 2 processors was the core for the production system. Eight processors actively supported production, while four processors were retained for Instant Capacity on Demand (iCOD), a feature of the HP Virtual Server Environment. It enables extra computing capacity onsite — preconfigured, cabled, powered up and ready to go at a moment’s notice — and there’s no charge before it’s used. This configuration provides fallback for processor failure and additional capacity for special events.
Scottish Life completed its migration in eighteen months, with a remarkable cutover from legacy to new systems occurring over a single weekend and with no additional downtime for users. Large complex code sets were compiled and deployed in a fraction of the time typical for the previous environment. The new open system-based cluster provided immediate performance benefits and demonstrated robustness, which added to the success of the project.
Summary and Metrics
The migration met Scottish Life’s goal of decreasing TCO, cutting it in half over five years compared to their previous computing infrastructure – saving £11.7M ($22.9M US). The savings were a sum of IT support costs and upgrades needed to support a growing business. Hardware upgrades were only 5 percent of the cost for similar changes to their previous environment, relative to performance and capacity. Application and middleware software costs were less, due to open system-based architecture supporting greater selection as compared to a small number of proprietary suppliers. Scottish Life already had skilled staff and contractors and no longer required in-house training of proprietary skills. To grow future computing capacity, Scottish Life plans to scale out the system and take advantage of new effective methods for programming and performance tuning.
Service levels are now higher. The nightly batch jobs run in 8 hours, down from 12 hours, which besides saving operational costs also extends opening hours for both internal & external clients. This speedup in batch processing means additional batch processes can run during the week to satisfy ad hoc requests and handle workload variations. Special processing runs, like the end of month closing, are run within normal batch windows. The processing of some exceptionally large business activities completes in minutes, instead of hours. Between 1500-2800 batch programs run at night compared to 1200-1800 on the prior system.
During the first year of operation, the new system was virtually 100 percent available with no unscheduled downtime due to hardware failures. Users experienced better on-line service operation with response times improving even as load increased - compared to the highly variable response times from the previous systems. The number of long response delays, 10 seconds or longer, decreased from more than 30 events a minute to rarely more than a few per minute. Average response time, measured over a series of months, originally ranged between 0.4 and 1.0 seconds; it later dropped to around 0.6 seconds, although transactions increased by 20 percent. Payment processing time takes five minutes and is done in one pass; previously it took three hours and multiple attempts.
Benefits of Using Itanium® 2-based Solutions
Before deciding on Intel Itanium 2-based servers, the project team investigated various solutions from different suppliers of both proprietary and open solutions. Two separate prototyping exercises were undertaken with different hardware, operating systems and system software. The Itanium® 2-based solution consistently set itself apart as the best technical solution for the following reasons:
• Early test of batch processes were consistently faster than other hardware.
• Code compilation times were faster.
• The Intel Itanium roadmap was impressive.
• The systems proved to be reliable during prototype testing.
• The hardware supported inbox upgrades.
Key Hardware Solutions
• HP Integrity servers with Intel Itanium 2 processors, models rx2470, rx2600, rx8620 running HP-UX 11i
• HP 9000 (PA-RISC) servers, models rp4440, rp3440, rp7410 running HP-UX 11i
• 2 x SANs based on HP StorageWorks Enterprise Virtual Array systems
• 2 x HP ESL9322 tape libraries with three 460 LTO2 tape drives
Key Software Solutions
• Oracle 9i Database
• Oracle Application Server & Oracle Management Packs
• Oracle Data Guard for data replication between test and production systems
• HP OpenView for hardware, UNIX, network and Oracle monitoring
• Micro Focus Studio & Server for legacy applications development
Spotlight: Royal London Group
• The UK’s largest mutual life and pensions company with over three million customers and funds totaling £29.5 billion ($57.7 US) under management.
• The Group’s businesses offer pensions, life assurance, savings and investment products, protection insurance and investment management.
© 2007 Itanium Solutions Alliance. *All other names and brands may be claimed as the property of others.
Note: Information and claims herein are provided by the award recipient and in now way warranted or endorsed by the Itanium Solutions Alliance. The Itanium Solutions Alliance does not control, verify or audit such information or claims and encourages all customers to independently obtain more information about the products.
For decades data-intense computing has been done on the same systems. But cost savings opportunities are driving many IT specialists to consider migrating to open systems-based environments and completely refreshing their IT infrastructure. Although hearing “migration” makes IT departments wince, many IT professionals are contemplating and carrying out the task despite the work involved.
UK pensions provider Scottish Life, part of the Royal London Group, relied on systems running ten million lines of code and managing nearly a billion data records. Launched in 1993, the previous system was struggling to maintain service levels. Internal staff & external users experienced a growing number of interruptions as their on-line service requests competed against transaction and background batch processing. Overnight batch jobs frequently ran longer than twelve hours forcing scheduled and ad hoc business processing to be postponed until the weekend.
The previous systems took five years to develop and their software was continually upgraded to respond to changing business needs. Operations tasks, including schedules, printing, performance monitoring and event management, were highly automated. Up to 1800 batch jobs ran nightly with over 64,000 business rules governing the distribution of up to 30,000 print images each day. A performance monitoring and intervention system supported the activities of more than five hundred users, and a remote second-site disaster recovery system was kept in constant lock-step.
Before deciding to migrate to an open system environment, Scottish Life assessed the complexity of moving its massive infrastructure - databases, business rules, operational activities and disaster recovery – and sought the help of MSS International*, a migration specialist, along with Micro Focus*, Oracle* and HP*. In addition to requiring a considerable amount of customization, they needed to replace system software packages with a set of completely new operational support utilities.
Challenge
• Lower IT operating cost. Replace the IT infrastructure to reduce hardware, support and maintenance costs.
• Improve the user’s service level. Make the computing systems more responsive to users and reduce the time required to perform nightly batch jobs.
Solution
• Migrate to an open systems architecture. Implement lower cost operating systems, servers, software and operations tools.
• Deploy scalable server cluster. Run enterprise software on Intel Itanium 2-based HP Integrity servers and PA-RISC-based HP 9000 servers.
Migrating to Open Systems-based Architecture
Scottish Life assembled a diverse team to migrate its Enterprise Application Environment (EAE) consisting of legacy code, data, system software and disaster recovery. MSS International provided translation software to automatically convert programs in proprietary languages and legacy COBOL code into Oracle PL/ SQL, C & Micro Focus COBOL, which can run on any open platform. The pension software application was moved to Cobal85 and data records were converted into an Oracle 9i database. Due to the unique nature, size and complexity of the previous systems, all business processes required testing by both IT testers and business users.
HP and Oracle supplied new system software and operational support utilities, including operating system and database software. Operations staff received comprehensive training, preparing them to setup the new software environment and later to provide ongoing IT support.
Oracle Data Guard configured a disaster recovery clone on the WAN-accessible development system located about two hundred miles from the production system. Other clones were created as part of the application development effort for performance testing, troubleshooting and data mining.
The project included producing new innovative products and modifying software to support new legal requirements, completed by a team of a hundred individuals. This development program ran in parallel with the migration project and both were simultaneously moved to the open systems-based cluster without interruption to the business.
An HP Integrity rx8620 with twelve Intel Itanium 2 processors was the core for the production system. Eight processors actively supported production, while four processors were retained for Instant Capacity on Demand (iCOD), a feature of the HP Virtual Server Environment. It enables extra computing capacity onsite — preconfigured, cabled, powered up and ready to go at a moment’s notice — and there’s no charge before it’s used. This configuration provides fallback for processor failure and additional capacity for special events.
Scottish Life completed its migration in eighteen months, with a remarkable cutover from legacy to new systems occurring over a single weekend and with no additional downtime for users. Large complex code sets were compiled and deployed in a fraction of the time typical for the previous environment. The new open system-based cluster provided immediate performance benefits and demonstrated robustness, which added to the success of the project.
Summary and Metrics
The migration met Scottish Life’s goal of decreasing TCO, cutting it in half over five years compared to their previous computing infrastructure – saving £11.7M ($22.9M US). The savings were a sum of IT support costs and upgrades needed to support a growing business. Hardware upgrades were only 5 percent of the cost for similar changes to their previous environment, relative to performance and capacity. Application and middleware software costs were less, due to open system-based architecture supporting greater selection as compared to a small number of proprietary suppliers. Scottish Life already had skilled staff and contractors and no longer required in-house training of proprietary skills. To grow future computing capacity, Scottish Life plans to scale out the system and take advantage of new effective methods for programming and performance tuning.
Service levels are now higher. The nightly batch jobs run in 8 hours, down from 12 hours, which besides saving operational costs also extends opening hours for both internal & external clients. This speedup in batch processing means additional batch processes can run during the week to satisfy ad hoc requests and handle workload variations. Special processing runs, like the end of month closing, are run within normal batch windows. The processing of some exceptionally large business activities completes in minutes, instead of hours. Between 1500-2800 batch programs run at night compared to 1200-1800 on the prior system.
During the first year of operation, the new system was virtually 100 percent available with no unscheduled downtime due to hardware failures. Users experienced better on-line service operation with response times improving even as load increased - compared to the highly variable response times from the previous systems. The number of long response delays, 10 seconds or longer, decreased from more than 30 events a minute to rarely more than a few per minute. Average response time, measured over a series of months, originally ranged between 0.4 and 1.0 seconds; it later dropped to around 0.6 seconds, although transactions increased by 20 percent. Payment processing time takes five minutes and is done in one pass; previously it took three hours and multiple attempts.
Benefits of Using Itanium® 2-based Solutions
Before deciding on Intel Itanium 2-based servers, the project team investigated various solutions from different suppliers of both proprietary and open solutions. Two separate prototyping exercises were undertaken with different hardware, operating systems and system software. The Itanium® 2-based solution consistently set itself apart as the best technical solution for the following reasons:
• Early test of batch processes were consistently faster than other hardware.
• Code compilation times were faster.
• The Intel Itanium roadmap was impressive.
• The systems proved to be reliable during prototype testing.
• The hardware supported inbox upgrades.
Key Hardware Solutions
• HP Integrity servers with Intel Itanium 2 processors, models rx2470, rx2600, rx8620 running HP-UX 11i
• HP 9000 (PA-RISC) servers, models rp4440, rp3440, rp7410 running HP-UX 11i
• 2 x SANs based on HP StorageWorks Enterprise Virtual Array systems
• 2 x HP ESL9322 tape libraries with three 460 LTO2 tape drives
Key Software Solutions
• Oracle 9i Database
• Oracle Application Server & Oracle Management Packs
• Oracle Data Guard for data replication between test and production systems
• HP OpenView for hardware, UNIX, network and Oracle monitoring
• Micro Focus Studio & Server for legacy applications development
Spotlight: Royal London Group
• The UK’s largest mutual life and pensions company with over three million customers and funds totaling £29.5 billion ($57.7 US) under management.
• The Group’s businesses offer pensions, life assurance, savings and investment products, protection insurance and investment management.
© 2007 Itanium Solutions Alliance. *All other names and brands may be claimed as the property of others.
Note: Information and claims herein are provided by the award recipient and in now way warranted or endorsed by the Itanium Solutions Alliance. The Itanium Solutions Alliance does not control, verify or audit such information or claims and encourages all customers to independently obtain more information about the products.
