Maintenance and Upgradation

Maintaining and upgrading IT systems is a vital link in the IT management life cycle. Loopholes here can lead to significant losses and services going down, writes Varun Aggarwal

Choosing the right mix of software and hardware is in itself a daunting task. CIOs often take months to deploy a new technology. This task involves selecting the right technology, the right vendor, convincing the management, getting approvals from the finance heads et al. However, the proverb ‘all is well that ends well’ does not hold true here because this is not the end but rather just the beginning. The real task, which is often more complex than the procurement of systems itself, is that of maintaining and upgrading your software and hardware assets. Maintenance and upgradation form the final and most important link of any IT project’s life cycle.

There seems to be little in the way of any formal methodology to categorize the labor requirement and skills that you need to install, operate and maintain hardware or software once purchased. Typically, software developers will put substantial efforts into quality assurance to minimize bugs and performance optimization prior to releasing the code. In the case of a customized software application, which the IT department develops for its own use, it establishes a formal training plan for those who have to install, maintain and secure it and for the user/operator community as well. There is currently no way to estimate or judge if that training is sufficient and if the designed staff for installing it and operating it is sufficient either.

The rising cost of administration

Manikkam VS, Head IT, Henkel CAC opined, “Today [this decision] is basically an educated guess. As a software system become more complex, the personnel and skill required for the information technology (IT) team becomes extremely demanding and costly to maintain. The people, dedication of the staff to properly manage such IT systems, and costs associated with those operations is generally glossed over.”

R P Singh, President and Head, Global Deliveries, Nucleus Software, said, “As time goes by, the dependence of business on deployed software rises. This makes continued maintenance a critical component of the investment plan. Maintenance covers production support as well as upgrades to keep in line with changing business needs. Typically, the provider of the software asset provides maintenance. The cost of transitioning it to alternate hands is too high and deemed feasible only in case a provider is too expensive or stops supporting the asset. Organizations must secure maintenance and support of software for the long term when acquiring the same.”

At DSIIDC, the system manager is responsible for the maintaining and upgrading the system after the deployment. However, it is a different story at Ranbaxy. David Briskman, CIO, Ranbaxy said, “We follow a model that shares implementation and support with the same team. We find that this balance works best with internal teams as one can balance the excitement of projects or new implementations with the stability (work-life balance) associated with support. In certain areas we augment support with external resources for day-to-day maintenance.”

Complexities in maintenance

Changes in the business environment can put great pressure on running systems. Most businesses undergo significant change within three years of IT deployment, which include new regulations, new product offerings, new operational processes, new channels and more. Adaptation to new needs, scalability to expanded business operations are critical demands in addition to correction of defects and this drives both the cost and maintenance of IT systems and software.

Companies have to ensure that the new solutions deployed are able to provide necessary support to operations and business departments. They have to ensure that the end-users of a solution are able to derive value from the solution. Scalability of the solution is a key factor which the stakeholders and managers are always cautious about. Singh opined, “Another critical success factor for the managers is cost of change because cost of change to a solution in production-mode is always higher than the cost of changing a solution under development or design. Actually, rather than focusing on cost of maintenance alone, companies will be better off looking at cost with reference to the value delivered to a business.”

Maintaining and upgrading legacy systems is one of the most difficult challenges that CIOs face today. Constant technological change often undermines the business value of legacy systems developed over the years involving huge investments. CIOs struggle with the problem of modernizing these systems while keeping their functionality intact. Despite their obsolescence, legacy systems continue to provide a competitive advantage through supporting unique business processes and containing invaluable knowledge and historical data.

“Legacy architectures are antitheses to modern distributed and layered architectures. Such architecture executes hardwired business policies and decisions by rigid, predefined process flows, making integration with customer relationship management (CRM) software and Internet-based business applications torturous and sometimes impossible. Addition, IT departments find it increasingly difficult to hire developers qualified to work on applications written in languages no longer found in modern technologies,” explained Manikkam. According to him, several options exist for modernizing legacy systems, defined as any monolithic information system that is too difficult and expensive to modify to meet new and constantly changing business requirements. Techniques range from quick fixes such as screen scraping and legacy wrapping to permanent, but more complex, solutions such as automated migration or replacing the system with a packaged product.

Many complexities go with not just legacy but also new systems. Briskman said, “As a global MNC with operations running 24x7 around the globe, having the right skills available all the time can be a challenge. Having a global help desk assists with this. The other area of complexity tends to be with specialized / niche (pharmaceutical-specific) applications, where the skills in these specific applications are hard to find. We work on building our in-house expertise as much as possible to avoid expensive outside expertise.”

To cater to such complexities, huge sum of capital investments go into maintenance and it is across verticals. Talking about the costs involved in the process, Arun O Gupta, Customer Care Associate & Group Chief Technology Officer, Shoppers Stop Limited, said, “Our operational IT budget is about 0.8% of revenue and that includes the maintenance spend. The maintenance focuses on ensuring that a system works as designed, provides the functionality consistently and incremental changes that may be required based on new requirements or dynamic business need. There are no significant issues with our current technology architecture and to that extent we have been able to maintain a stable operation.”

The situation is more complex in the government sector. Surinder Kapur, CIO, DSIIDC, explained, “The lack of knowledge in the government sector and the stringent rules that exist do not allow us to spend money. You have to explain the ‘what and why’ of each activity. To do this, you have to grab the knowledge either from publications or by interacting with other CIOs. It is difficult to explain the economics because you cannot buy each time so one has to adopt ‘request, steal or borrow’. I have been using all of these techniques at the time of need when I am supposed to find a solution to a problem.”

Gupta added, “The cost of maintenance is in-built into the project charter for a period of 3-5 years, which is the typical useful life of any application. When we take any project for approval, it includes the next 3-5 years of maintenance based on certain estimates and licensing models from vendors. Exceptions to this are managed as they come up, which could be major enhancements or changes, add-on modules or scaling up.”

In-house vs. outsourced

Whether to go for in-house maintenance or an outsourced model—is a question that often comes to the minds of CIOs and the answer is not as simple as it seems to be. Organizations with skilled IT staff usually do not face this challenge and often go with in-house maintenance. However, organizations especially non-IT organizations take either the outsourcing route or a mix of both.

With specialization available, it is often beneficial to outsource maintenance in the long run. Economies of scale help in bringing down cost. It is easier to demand performance and cost improvements with good vendors. Of course, it is not easy to find and manage the right partner.

Gupta said that his company uses a mix of in-house and outsourced resources to manage the maintenance of the systems. Ranbaxy also follows a similar policy. “I believe that a mix of in-house and outsourced expertise is the best model. While a system requires many people, you can utilize outsourcing firms to assist with day-to-day operations, but there is no substitute for internal expertise,” added Briskman.

Briskman also said that outsourcing takes place for strategic reasons rather than economic ones. “Given the type of career that we offer, we find it easy to attract talent. However, day-to-day maintenance jobs tend to have higher attrition rates. Having partnerships with firms with broader roles for people, tends to help reduce the impact of attrition at in these roles and allows the internal team to focus on improving the business,” he added.

One of the biggest advantages of outsourcing is that it normally brings in the potential of scaling up maintenance capacity for short and medium term in a cost-effective manner. Singh said, “Organizations normally engage on fixed price with fixed scope of work but build the flexibility of scaling up when necessary. There are also performance-based incentives and penalties introduced to keep everyone focused on business uptime. Do be aware that such plans could change the focus to earning incentives or evading penalties rather than the real goal.”

Preparing for an upgrade

Many organizations are reviewing business practices and upgrading information technology (IT/ICT). The need to achieve a smooth system implementation, free of omissions and delays, is obvious for successful business management.

While you cannot guarantee a smooth and error-free implementation, particularly where new technology is involved, there are processes and actions that can maximize the positives and minimize problems.

Manikkam explained that factors that merit consideration include the extent and complexity of the change/upgrade; the availability of staff for the extra work and time involved; staff skills and experience in implementing change, particularly technological change. Changes, particularly those involving technology, intensively consume resources. As resources are scarce in most organizations, special attention is required. Effective control in day-to-day administration is essential. Consequently, control is critical when complementing changes or upgrades,” he added.

Software upgrades of proprietary systems can sometimes require an upgrade of the operating system or network software to the latest or later version. This may even require upgrades or replacement of hardware and network hardware. These combinations increase the complexity dramatically, as interfaces between the various components may cause operational problems. Extensive testing of a new system is essential in this type of environment.

The cost of upgradation

Every organization faces time and budget constraints. Budgets may have considered the capital costs involved with the implementation of a new system.

According to Briskman, “In most cases, business requirements will drive the need for upgrading applications and new functionality supported in later releases. Performance and reliability are two of the factors that drive hardware upgrades. We do not like to be ‘out of support’, however, we believe in leveraging investments for as long as makes business sense to get RoI. Strategic reasons (e.g., data center consolidation) drive other upgrades. However, there is always a business case for any upgrade or investment that is approved / agreed / sponsored by the business.”

At Shopper’s Stop, Gupta believes a methodical process, which involves answering certain questions before taking a decision on any upgrade. The questions are:

• Does the upgrade offer any material improvement over the existing version?
• What is the investment required to make the change?
• How long will the vendor continue to support my current release?
• What is the cost of sustaining the current infrastructure for the application in comparison to the new one?
• Are there any other benefits (tangible or otherwise) by moving to the new platform?

According to Kapur, “Upgrading of technology is done when the hardware loses its life, because explaining technology to CEOs, who are not from the IT side, is impossible. Under such circumstances, justifying the technology is a Herculean task. I had changed the hardware after every five years because as per IT policy any hardware which is five years old can be replaced for which no justification is needed.”

Incremental or Big Bang

The biggest challenge of any technology deployment is change management. Although the incremental approach poses the additional challenge of managing a slew of changes, it can be quite effective as you introduce small changes over a longer period. It is clearly more expensive as you end up repeating testing, training and deployment. However, the big bang approach faces strong change barriers and may result in a failed project at times.

Briskman said, “It will depend on the system and the situation. I am personally a ‘big bang’ approach fan. In order to achieve improvement or change step-by-step, you need to make big leaps. If you do everything incrementally, you will eventually by a laggard. That said, there is only so much risk or change and organization can take. Understanding the organizations tolerance for both would be crucial in deciding whether to go for a ‘big-bang’ or incremental upgrades. The more big-bang you can achieve, the faster you will move as an organization, the faster the return on investment.”

Briskman added that his team’s methodology for most upgrades was to lead the effort and collaborate with key solution providers to deliver the upgrades. This approach also helps the internal team get ready to support an upgraded platform. “End-of-life upgrades do happen, but we find that most application providers will arrange to provide support even for these platforms. That said, this does usually add risk and heartache as problems occur that are either ‘fixed in the next version’ or unique. The most important thing to do in an upgrade is get on with it and make it happen. Too much focus on getting everything to work or the perfect upgrade can turn a six month project into a twelve or eighteen month project—thus pushing costs further up or lose business opportunity. Plan it, work the plan, execute on time, ensure accountability, and deliver should be mantra for handling upgrades,” he added.

“In a virtualized environment this is not a major challenge. We have moved almost all critical applications to virtualized hardware, which helps us manage the increase in the mid-term without additional investments, “said Gupta.

If the upgrade is purely technology related, simple testing to ensure (a) operations (b) performance comparison is enough. However, if the upgrade is functional, companies should perform elaborate functional testing in addition to performance assurance testing. In case of technology upgrades, if there is no support for a particular version of a product, there is little choice but to go for an upgrade. However, in case of functional upgrades, the decision is mutual. Organizations restrict functional upgrades to one or at the most two in a year and make small defect correction updates more often.

varun.aggarwal@expressindia.com