IT KEE

Sanity Check: Shallow Thinking Leads to Deep Regrets

By Bill Carico
Recently, the Microsoft-leaning CTO of a large organization showed me the roadmap he had created for dumping the mainframe in favor of either a Lintel (Linux on Intel) or Wintel (Windows on Intel) server farm. So, I asked him if he was familiar with the difference in Disaster Recovery (DR) capabilities between those two platforms, and how Lintel had a considerably better track record for decreasing downtime (RTO [Recovery Time Objective]) and avoiding data loss (RPO [Recovery Point Objective]). His lengthy response revealed his convictions regarding the nobility of keeping the roadmap pure from technology considerations and the importance of not haggling about petty platform distinctions, because many bright people are continuously working to eliminate shortcomings, wherever they may exist.
Some things worth noting from his reply include:
• Shallow thinkers display blind faith and ultimate trust that someone else, somewhere else, will come up with a way to close capability gaps between platforms.• Ignorance limits one’s ability and desire to engage in technology discussions.• If you let a pseudo-technician serve as your CTO, then you are playing with fire.
DR capabilities are just one area where substantial maturity and functional differences exist between platforms. Naturally, platform suitability depends on what percentage of your applications and transactions must run on your recovery (backup) systems during planned and unplanned downtime, and whether the backup environment needs to handle the same scale and breadth of function (e.g., applications and devices) as provided by production systems. The larger the enterprise, the greater the struggle becomes for those who use Intel platforms to manage explosive information growth and avoid the risk of data loss while running non-stop. Facing regulatory mandates and operational threats ranging from blackouts to natural and manmade disasters to human error to data corruption, a platform’s ability to keep data protected from loss isn’t a trivial matter. If the plan is to have less than full capability during downtime, responsible IT planners should know the difference in robustness and maturity between technology choices and inform their stakeholders about trade-offs and how to mitigate risks, including operational, strategic, and legal risks.
Not surprisingly, shallow thinkers are easily impressed by vendor claims and are too quick to believe vendor promises. Naïveté also is a common trait among pseudo-technicians who live on the reading- edge but lack depth of experience. These are the CTOs and IT architects who childishly trust in the safety of the pack, and in the pack’s collective ability to keep things moving forward.
When it comes to gaps, if you like how virtualizing the server farms has improved DR options for the Wintel and Lintel crowds, then you’ll love the advantages of Linux on System z or running under z/VM. Especially as you move toward cloud computing, consider how the strengths of System z support simplification and consolidation to deliver optimal RPO and RTO. Specific challenges for the large enterprise include the recovery times for different applications, the missing points of data consistency that make data recovery a nightmare, different applications points of recovery in the production cycle, and the need to have all servers included in the replications even though they have different RPO/RTO standards and operate under different Service Level Agreements (SLAs). For both Lintel and Wintel, the greater challenge comes from trying to match hardware requirements with the backed-up system. Intel components vary much more than those found in the mainframe environment, guaranteeing driver and functional difficulties. For mainframes, an OSA card is an OSA card—no issues there. However, on Intel, the NIC can be Linksys (Cisco), Intel, 3Com, etc., thereby creating restore and start-up glitches.
Backup and recovery capabilities wildly vary among platforms—from journals to snapshots to checkpoints to mirrors to encryption to server clusters to logical data recovery with switchover, to name just a few. System z offers by far the most redundancy and the most sophisticated solutions. One example is Geographically Dispersed Parallel Sysplex (GDPS) for both synchronous and asynchronous rapid and complete data recovery. IBM anticipates expanding GDPS to serve as the focal point for additional platforms.
Since large IT shops are mixed-vendor, platform choices will continue to have a substantial impact on complexity, which can be minimized to reduce TCO and eliminate risk. The IT executive’s role is to ensure that the DR and business continuity plans are thorough and reasonable. Shallow thinking leads to deep regrets.

India’s ELCOT: A Next-Generation Mainframe Cloud Services Provider?
By Joe Clabby
Electronics Corporation of Tamil Nadu Limited (ELCOT) is a government-owned provider of Information and Communications Technologies (ICT) services to various government organizations in the Indian state of Tamil Nadu. While ELCOT isn’t exactly a household name, the company would like to establish what could possibly become the first instance of a mainframe “cloud” environment.
ELCOT’s many services include the deployment of systems, storage, and network products and operating environments; custom applications for design and development; technology consulting; and ICT training. ELCOT must support government mandates, including a requirement to promote the use of open source software, as well as help its constituents find ways to reduce the cost of IT. The combination of these two mandates has led ELCOT to the purchase of an IBM System z9 mainframe.
At ELCOT, IBM’s System z9 is positioned as an “open source consolidation server.” As an open source server environment, IBM’s System z9 can run a workload that’s equivalent to 250 Linux/x86 server workloads. And because the z9 also supports Web services, Service-Oriented Architecture (SOA), the Linux operating environment, and Eclipse infrastructure, deploying open source software on System z9 is straightforward and easy.
So far, ELCOT has persuaded several government departments to adopt the open source model. Several eCitizen applications such as the state’s “Family Card” application, which is used to subsidize food purchases, now run on ELCOT’s mainframe. Several of ELCOT’s own Enterprise Resource Planning (ERP) applications are now hosted on Linux on a System z9. But convincing government departments to move to the open source model is a very slow process. And because ELCOT’s customers have been slow to adopt the open source model, as much as 60 percent of the organization’s computing capacity current isn’t being used!
Dr. Santhosh Babu, ELCOT’s managing director and director of e-Governance, wants to fix this problem. He hates wasting IT resources and computing capacity. So he’s aggressively seeking to find a business partner that can manage his System z—and that can sell the unused capacity on his System z9 to other government users or commercial businesses. If successful in implementing this plan, Dr. Babu will essentially build an advanced “cloud computing” environment that runs inside his System z9—quite possibly the first instance of a mainframe “cloud” environment.
Background
Dr. Babu isn’t a technologist; he’s a medical doctor. His background is in operations and project management; his strength is that he knows how to build applications that serve governmental and citizen needs. The strategic decision to adopt mainframe technology was made by Dr. Babu’s predecessor, Mr. C Umasankar, who recognized that a mainframe represents the pinnacle platform for consolidating Linux servers and implementing open source.
Having inherited the charter to drive Linux and open source applications onto a mainframe, Dr. Babu described his two biggest challenges as:
• Trying to convince his clients (a wide variety of government organizations) to more aggressively adopt the use of open source software• Persuading more departmental computing users to embrace centralized, scale-up resources.
Increasing Mainframe Utilization
A s Dr. Babu pursues these initiatives to increase mainframe usage, he’s seeking to:
• Find billing and accounting software to determine computing resource use. With such software (IBM’s Tivoli Usage and Accounting Manager is one example), ELCOT would be able to meter how much computing time and resource an external user used on its mainframe—and could bill for that use accordingly.• Potentially introduce the concept of cloud computing, an evolving computing architectural model that calls for unused computing resources to be returned to a common, “virtualized” resource pool. Resources in this pool can then be made constantly available to users who need computing power.
Dr. Babu is already practicing cloud computing concepts. From his PC desktop, he operates a program that contains links to various applications that run on a variety of servers in the ELCOT infrastructure. He doesn’t know the physical location of those servers, nor does he care. All he cares about is getting access to those application services that are somewhere out there in “the cloud.” To implement a true cloud, servers running Dr. Babu’s applications would need to constantly return unused computing power to a virtualized resource pool where those resources could be automatically provisioned and reused. But ELCOT is part of the way there because the company has a graphical user interface to backend applications—and the users are unaware on which hardware those applications run. If Dr. Babu can host more open source applications on his System z, and if his administrators take some time to automatically provision that environment (provisioning is the build-up or tear-down of system images to make way for new applications to run), then he will have implemented a true mainframe cloud.
Dr. Babu would prefer to use an outside business partner to market and support this potential cloud computing environment on his mainframe. He would like to make his computing resources available to outside users without having to manage those users himself. This concept of providing resources to an outside supplier is again consistent with how the cloud computing model is evolving. Using that model, ICT managers with excess capacity can actually sell that unused capacity to service providers and generate revenue for their organizations or enterprises.
Conclusion
The Indian government intends to broaden the adoption of open source software; the government doesn’t want to spend big money for UNIX or Windows operating systems and related infrastructure and applications. Ultimately, the government would like to see UNIX and Windows applications migrate to the Linux and open source model.
ELCOT needs to be in the forefront of technology trends and clearly understand how technology can be exploited to deliver operational efficiency and benefits to Indian citizens and businesses. One way ELCOT has found to achieve maximum efficiency is to deploy a highly virtualized mainframe environment that can easily run at 100 percent of capacity for lengthy, sustained periods. Further, mainframe technology also helps ELCOT drive down systems acquisition and Total Cost of Ownership (TCO) costs. By consolidating server environments, ELCOT can use its mainframe to drive down management costs and reduce management complexity.
ELCOT’s biggest challenge will be to change government ICT users’ current purchasing patterns. Many government offices and departments still have the “we-want-our-own-server” mentality—a mentality that ensures the proliferation of hundreds or even thousands of stand-alone, towered departmental servers. These configurations are difficult to manage and are massively underutilized. ELCOT and the Indian government know that changing this buying behavior will take time—and a lot of convincing. Buying and deploying a mainframe goes a long way toward demonstrating and proving the kind of efficiencies that can be achieved by moving to a centrally managed scale-up model. The efficiency of using an IBM z9 should help persuade many ICT users to rethink their current buying behavior.
Dr. Babu’s creative thinking regarding how to maximize the use of ELCOT’s mainframe environment is admirable. He eschews waste and is actively seeking ways to maximize his resource utilization while lowering costs. More IT executives should demonstrate similar insight and ingenuity.

India’s ELCOT: A Next-Generation Mainframe Cloud Services Provider?
Published
by
Publisher
on November 19, 2009
in General.
By Joe Clabby
Electronics Corporation of Tamil Nadu Limited (ELCOT) is a government-owned provider of Information and Communications Technologies (ICT) services to various government organizations in the Indian state of Tamil Nadu. While ELCOT isn’t exactly a household name, the company would like to establish what could possibly become the first instance of a mainframe “cloud” environment.
ELCOT’s many services include the deployment of systems, storage, and network products and operating environments; custom applications for design and development; technology consulting; and ICT training. ELCOT must support government mandates, including a requirement to promote the use of open source software, as well as help its constituents find ways to reduce the cost of IT. The combination of these two mandates has led ELCOT to the purchase of an IBM System z9 mainframe.
At ELCOT, IBM’s System z9 is positioned as an “open source consolidation server.” As an open source server environment, IBM’s System z9 can run a workload that’s equivalent to 250 Linux/x86 server workloads. And because the z9 also supports Web services, Service-Oriented Architecture (SOA), the Linux operating environment, and Eclipse infrastructure, deploying open source software on System z9 is straightforward and easy.
So far, ELCOT has persuaded several government departments to adopt the open source model. Several eCitizen applications such as the state’s “Family Card” application, which is used to subsidize food purchases, now run on ELCOT’s mainframe. Several of ELCOT’s own Enterprise Resource Planning (ERP) applications are now hosted on Linux on a System z9. But convincing government departments to move to the open source model is a very slow process. And because ELCOT’s customers have been slow to adopt the open source model, as much as 60 percent of the organization’s computing capacity current isn’t being used!
Dr. Santhosh Babu, ELCOT’s managing director and director of e-Governance, wants to fix this problem. He hates wasting IT resources and computing capacity. So he’s aggressively seeking to find a business partner that can manage his System z—and that can sell the unused capacity on his System z9 to other government users or commercial businesses. If successful in implementing this plan, Dr. Babu will essentially build an advanced “cloud computing” environment that runs inside his System z9—quite possibly the first instance of a mainframe “cloud” environment.
Background
Dr. Babu isn’t a technologist; he’s a medical doctor. His background is in operations and project management; his strength is that he knows how to build applications that serve governmental and citizen needs. The strategic decision to adopt mainframe technology was made by Dr. Babu’s predecessor, Mr. C Umasankar, who recognized that a mainframe represents the pinnacle platform for consolidating Linux servers and implementing open source.
Having inherited the charter to drive Linux and open source applications onto a mainframe, Dr. Babu described his two biggest challenges as:
• Trying to convince his clients (a wide variety of government organizations) to more aggressively adopt the use of open source software• Persuading more departmental computing users to embrace centralized, scale-up resources.
Increasing Mainframe Utilization
A s Dr. Babu pursues these initiatives to increase mainframe usage, he’s seeking to:
• Find billing and accounting software to determine computing resource use. With such software (IBM’s Tivoli Usage and Accounting Manager is one example), ELCOT would be able to meter how much computing time and resource an external user used on its mainframe—and could bill for that use accordingly.• Potentially introduce the concept of cloud computing, an evolving computing architectural model that calls for unused computing resources to be returned to a common, “virtualized” resource pool. Resources in this pool can then be made constantly available to users who need computing power.
Dr. Babu is already practicing cloud computing concepts. From his PC desktop, he operates a program that contains links to various applications that run on a variety of servers in the ELCOT infrastructure. He doesn’t know the physical location of those servers, nor does he care. All he cares about is getting access to those application services that are somewhere out there in “the cloud.” To implement a true cloud, servers running Dr. Babu’s applications would need to constantly return unused computing power to a virtualized resource pool where those resources could be automatically provisioned and reused. But ELCOT is part of the way there because the company has a graphical user interface to backend applications—and the users are unaware on which hardware those applications run. If Dr. Babu can host more open source applications on his System z, and if his administrators take some time to automatically provision that environment (provisioning is the build-up or tear-down of system images to make way for new applications to run), then he will have implemented a true mainframe cloud.
Dr. Babu would prefer to use an outside business partner to market and support this potential cloud computing environment on his mainframe. He would like to make his computing resources available to outside users without having to manage those users himself. This concept of providing resources to an outside supplier is again consistent with how the cloud computing model is evolving. Using that model, ICT managers with excess capacity can actually sell that unused capacity to service providers and generate revenue for their organizations or enterprises.
Conclusion
The Indian government intends to broaden the adoption of open source software; the government doesn’t want to spend big money for UNIX or Windows operating systems and related infrastructure and applications. Ultimately, the government would like to see UNIX and Windows applications migrate to the Linux and open source model.
ELCOT needs to be in the forefront of technology trends and clearly understand how technology can be exploited to deliver operational efficiency and benefits to Indian citizens and businesses. One way ELCOT has found to achieve maximum efficiency is to deploy a highly virtualized mainframe environment that can easily run at 100 percent of capacity for lengthy, sustained periods. Further, mainframe technology also helps ELCOT drive down systems acquisition and Total Cost of Ownership (TCO) costs. By consolidating server environments, ELCOT can use its mainframe to drive down management costs and reduce management complexity.
ELCOT’s biggest challenge will be to change government ICT users’ current purchasing patterns. Many government offices and departments still have the “we-want-our-own-server” mentality—a mentality that ensures the proliferation of hundreds or even thousands of stand-alone, towered departmental servers. These configurations are difficult to manage and are massively underutilized. ELCOT and the Indian government know that changing this buying behavior will take time—and a lot of convincing. Buying and deploying a mainframe goes a long way toward demonstrating and proving the kind of efficiencies that can be achieved by moving to a centrally managed scale-up model. The efficiency of using an IBM z9 should help persuade many ICT users to rethink their current buying behavior.
Dr. Babu’s creative thinking regarding how to maximize the use of ELCOT’s mainframe environment is admirable. He eschews waste and is actively seeking ways to maximize his resource utilization while lowering costs. More IT executives should demonstrate similar insight and ingenuity.