It’s no wonder IT leaders are focusing concern on IT infrastructure and operations costs when it accounts for 60 per cent of total IT spend. Since things are still uncertain despite the economy showing signs of improvement, this area of IT “strikes a resonant cord,” said Jay Pultz, vice-president and distinguished analyst with Gartner Inc. (IT)
“It has such a high interest among CIOs because they’re not going to meet their budget goals if (infrastructure and operations) doesn’t meet its (budget),” said Pultz during a recent Webinar discussing ways enterprises can cut costs in their IT infrastructure and operations (I&O). Gartner defines I&O as everything in IT except business applications.
There are 10 ways to cut costs in I&O that, if done completely, will reduce spend by 10 per cent within 12 months and 25 per cent in three years, said Pultz.
1. Defer those I&O initiatives that don’t meet business needs. “Look at your key initiatives and focus on those that help you meet the business needs, help you reduce costs and help you keep to uptime requirements for crucial systems,” said Pultz. Preferred initiatives include data centre modernization and consolidation, virtualization, improving processes with ITIL, upgrading PCs, unified communications, and a mobile enterprise strategy.
2. Re-examine networking costs. “These tend to be the largest contracts that you have if you’re not doing a significant amount of outsourcing,” said Pultz. By reviewing current telco contracts, enterprises can renegotiate a lower rate, identify billing errors and things they should no longer be paying for. Pultz also suggests revisiting the network architecture and refining uptime requirements.
3. Consolidate I&O. “You must ask yourself if you have consolidated all that you can,” said Pultz. Most data centre managers understand the benefit of replacing distributed and standalone servers with new form factors in the data centre such as rack and blade. Yet less than 10 per cent have consolidated their servers, said Pultz.
4. Virtualize I&O. “Virtualization is an incredibly powerful technology to reduce hardware costs, reduce power costs, improve utilization and so forth,” said Pultz. For instance, reducing server count by 75 per cent will reduce power consumption by a similar amount. Pultz suggests: “Don’t go slow, accelerate as much as possible your plans to virtualize servers” because of the advantages that can be reaped from high levels of virtualization.
5. Reduce power and cooling needs. “We’re just creating more capacity which creates more heat in a similar space,” said Pultz. Don’t think of the data centre holistically. Instead, break it down into pods or modules and design each individually. One module might be designed for heat density with hot and cold racks, while an other designed for higher uptime, said Pultz. Don’t forget energy monitoring tools are useful for knowing your usage and for optimizing that.
6. Contain storage growth. “We’ve got to do more than just throwing terabytes at the problem,” said Pultz. Yet, that’s the popular response among data centre managers. Some techniques include data deduplication to reduce repeats of the same data, thin provisioning to make better use of storage, and compressing files.
7. Push down IT support. “Institute self-service to get end users to do more on their own and call the service desk less often,” said Pultz. Each level of IT support has a different cost, be it self-service, service desk, technical staff, strategic staff. Having end users help themselves is cost efficient because self-service is one-tenth the cost per transaction.
8. Streamline IT operations. “Change management is a very large source of problems that generate incidents that must be solved by the service desk,” said Pultz. Applying an ITIL framework can ensure change is done correctly. Streamlining processes in general, said Pultz, can reduce the operations part of the IT budget. It also means an enterprise can reduce the need for expensive highly skilled IT professionals.
9. Enhance IT asset management. While improving IT asset management won’t result in cost savings, Pultz said having that data source is invaluable for analyzing potential cost savings across I&O. Some areas that the data can highlight are the need to extend server life, combining contracts into volume purchase agreements, and eliminating software licenses for staff who are no longer employed.
10. Optimize multi-sourcing. “I&O over the years has got much more complex with greater demands from the enterprise,” said Pultz. But it’s not a “binary decision” whether to outsource or insource, he added. It deserves a granular look at function, process and platform. Pick one of those areas to outsource. But be careful not to outsource so much that it becomes costly to manage the slew of service providers.
Overall, Pultz has observed enterprises having only completed a third of his 10 recommendations. He suggests comparing I&O costs to that of other organizations to get a sense of where one stands.
Along the same vein, a study last March by Germany-based infrastructure company Software AG found that the majority of IT directors polled considered services-oriented architecture (SOA) to be a good cost-cutting measure in this tough economy.
Specifically, 83 per cent said they will use SOA to cut costs and 40 per cent said they planned to implement SOA governance in the next year.
The reusability of infrastructure components that SOA affords means greater return on investment for an organization, said Tim Holyoake, lead technologist with Software AG. “Tough economic challenges have forced many businesses to break free from traditional application constraints and re-use existing systems to help keep costs down,” said Holyoake.