Q&A: Getting a Handle on IT Hardware
Getting a handle on your IT servers can be difficult. We explain how one company tackled the problem.
Getting a handle on IT resources can be a daunting task. As The Carphone Warehouse grew and completed an acquisition, the firm knew it needed to get a handle on the assets in its data center. To learn how the company approached its data center transformation initiative, what the project required, what solutions were used, what lessons the company learned from its new project, and how IT intends to expand the new technology, we spoke to David Byrne, architecture and design director at The Carphone Warehouse.
Enterprise Strategies: What is The Carphone Warehouse (CPW)?
David Byrne: The Carphone Warehouse is the world's largest independent mobile phone retailer, as well as one of the UK's leading telephone and Internet service providers. The Carphone Warehouse has more than 2,450 stores in nine EU countries and 1,000 Best Buy Mobile locations in the U.S.
What was the problem CPW faced? How was CPW handling the problem when you went searching for a better solution?
Knowing what's "under the hood" of our IT organization has been a particular challenge for us, especially with our growth and acquisition of AOL UK in the Telco space. We needed to gain greater control and cost savings within IT and "ERP for IT" seemed to be a good solution to increase our efficiency. Specifically, post-acquisition, we had no measure of our more than 1,400-server estate and interdependencies, we had no reliable, centralized database to support ITIL initiatives, and our CMDB integrations were slow, therefore expensive. Additionally, our Best Buy joint venture required better clarity between Carphone Warehouse Retail Systems and Carphone Warehouse Telco Systems. This would enable the two businesses to be separated at a technology level in preparation for the de-merger of the two businesses.
The business objectives were to:
- Accurately measure the server estate to remove redundant servers
- Identify server-to-server dependencies and network configurations to manage change
- Populate and maintain CMDB data to support ITIL processes
- Enable migration of multiple virtual servers to ensure continuity for key business services
- Support priority computing on demand (COD) initiatives to improve utilization and efficiency
Before searching for a better solution, we were attempting to handle these issues using another commercial discovery tool. However, this was not working well in our environment and the process was being completed manually. The added time and expense of this lead us to search for something that was a closer fit to our needs.
What solutions (such as revised manual process, automated solution, and outsourcing) did CPW consider when looking for a solution?
We already had a tool in place, but we knew a better solution was needed. When trying to find a new automated approach, we looked at a large number of available systems, including, BMC, IBM, Tideway, EMC, and CA.
What criteria did you use for selecting a solution? What was the most important factor?
The solution needed to have an agentless discovery process as we wanted to find unknown dependencies, which would have been impossible with an agent tool.
The most important factors we considered included:
- Having an agentless solution
- Rapid and easy deployment
- Coverage of our infrastructure technology (Unix and Windows)
- The ability to map our critical business services
- The ability to import Cisco works Network data
- A CMDB Integration to BMC's Atrium
What technology did you select and why? How did it meet your criteria and satisfy the company's needs?
We selected Tideway Foundation and deployed it across our four data centers in order to generate a complete view of all our physical and virtual assets. The initial baseline gave us a single system of record for configuration item (CI) inventory and dependencies, while delivering data directly to our CMDB, which supports ITIL processes.
Tideway provided an automated baseline of our global infrastructure. In addition to enabling us to measure our server estate and remove redundant equipment and identify server-to-server dependencies and network configurations, the software helps us:
- Automatically populate and maintain our CMDB data to support ITIL processes
- Streamline migration of multiple virtual servers
- Inform and improve negotiations with software vendors; with an accurate inventory of installed software, we were able to achieve significant savings when faced with a vendor-enforced audit (true-up)
- Eliminate manual infrastructure audits, saving hours upon hours of labor.
Tell us about your implementation experience. For example, how long did it take (and how was this different that what you'd planned), was this a company-wide solution or used by a select group of users (which ones and why), what unexpected problems did you experience, did your IT team need special skills, did you use outside consultants?
We started this project to populate our CMDB, but soon realized that Tideway Foundation allowed us to get accurate information regarding infrastructure footprint for all our other initiatives, so the project quickly expanded. The implementation took just over a year; however, this was not due to Tideway. The main time was spent getting the CPW application teams on board to enable rollout of credentials in individual servers to enable Tideway to access them.
Through this experience, we learned to get people to see the value for them and not just the "greater good." We also learned to provide early access to the data, even if it's not complete -- Foundation helps point out the gaps and often help to fill it. Finally, we learned to solicit "quick win" improvements for the tool. Tideway Professional Services were able to create reports and extend meta-model, etc., to add immediate value.
What was critical to the success of your project?
Things that were critical to the success of our project included being realistic about our goals and having a primary objective for the project. Also, defining success criteria in advance was helpful to keep the project on track. Getting buy-in from project teams early and providing early access to the tool, as well as demonstrating "quick wins" and finding ways to add immediate value, were key to our successful deployment. Finally, capitalizing on the secondary (and probably more useful) benefits of the project not only helped achieve our initial goals with great success, but also helped us leverage other benefits for CPW.
What advice would you have to your colleagues considering this type of technology? Is there anything other IT shops should watch out for when implementing this technology?
When implementing this technology, other companies should be careful of failing to make someone accountable for enabling access to servers in the data centers. They also need to be cautious of underestimating the time needed to configure servers to enable visibility and identifying and assembling subject matter experts (the people who know what servers/applications on servers actually do). Additionally, not finding a team to "own" the ongoing scanning/reporting/reconciliation work, and allowing a project to go on too long without a quick win (you need to show immediate value) should be advised.
If you had it to do over again, what would you have done differently?
If we had to do it over again, we would assemble the CPW application teams to enable roll-out of the credentials much earlier so as not to waste time and get the project implementation started that much sooner. Additionally, we would have provided access to the data at an earlier stage, even if it wasn't complete, and we would have tried to show the immediate value by soliciting a "quick win."
Who is a candidate for this technology? At what point does it make sense for companies to look at this type of product to address their own challenges or goals?
Any company looking to gain a clear picture of their IT environment is a candidate for Tideway Foundation. However, by nature, the more servers and data centers a company has, the more they are at risk of having decommissioned servers, inaccurate counts of inventory, old software, and (in general) inefficient data centers. This is especially evident in companies that have gone through a merger or acquisition and are faced with integrating and rationalizing two or more disparate IT environments.
By employing automated discovery and dependency mapping tools, a company can get to the core of their infrastructure to see which applications are linked to which servers and where they can consolidate servers or duplicate assets.
Tideway is particularly valuable to companies that are IT dependent, have a high cost of IT downtime, need to meet compliance requirements (such as PCI, SOX, and software license management), and are looking for fast time to value.
How do you plan to expand your use of this technology?
There is still a lot to do. We plan to upgrade to Tideway Foundation version 7.2 to have better insight into MS SQL Server while consolidating and virtualizing the many SQL instances. We also still need to prioritize the applications for the service-to-application-to-infrastructure mapping. Additionally, as a means to reduce the number of OS and application version variances (to drive down costs and better automate), we need to report on standardization adherence.
Moving forward, we are planning a DNS migration project and plan to consolidate 19 different DNS servers to just one fault tolerant pair per data center. Finally, to achieve compliance for OS patching, we will implement an "independent auditor" to ensure our patching compliance.
We are also continually acquiring companies and sharing best practices with them. Tideway's footprint will continue to be extended as we integrate M&A infrastructure.
What other lessons did you learn from this experience?
Tideway is a general-purpose tool capable of performing many different tasks "out-of-the-box." Have a clear idea of the specific uses in which you will apply it to in your organization so that you focus on achieving these. Examples include server consolidation/virtualization and building application dependency models.