A virtualization change-out with several twists
Scott Baynes
CTO, Netgain
In the last 18 month as Netgain's CTO, Scott Baynes has faced several challenges stemming from the rapid growth of the company's private cloud computing clientele, which is composed largely of private medical groups. The firm's 300-server data center was quickly approaching N+1 capacity markers, threatening to require significant new investment in a physical plant. Baynes faced a classic technical and business case for server, network, and storage virtualization.
Within his small infrastructure team of four people, Baynes assembled a skunk-works team to vet industry-leading solutions and select a virtualization platform to support the team's goals of high availability, rapid deployment, centralized management, and financial scalability. Additionally, Baynes won admission into a relatively exclusive club of firms renting virtualization licensure, a novelty at the time, especially for companies of Netgain's size.
Baynes was able to reduce the physical space needed in its Minnesota data center facility -- even though the server count had nearly doubled. The unwelcome need for investment in physical plant was deferred indefinitely and perhaps even eliminated.
But there's more to the deployment than adopting server virtualization: Implementing server virtualization in a service provider environment was neither straightforward nor well-tested. Existing methodologies for physical-to-virtual migration centered primarily on either single-purpose servers or relied on extensive downtime. Additionally, although comparative performance metrics for virtualization products always trend aggressive, real-world results vary dramatically, as Netgain found in initial efforts.
So Baynes' team conducted significant load testing, conversion practice runs, and essentially developed a whole new test lab oriented around the long-term project requirements of large-scale conversion to a virtual environment. In practice, Netgain's use also required development of virtual hosts groups specific to phases of production, as Baynes' team discovered when moving systems into and out of test, preproduction, and live environments. This conversion activity also took place in a contractual environment in which SLA requirements for performance did not change, so processor, RAM, and other key server metrics had to remain stable (or show improvement) while Baynes worked to eke out economies of scale promised by the technology.
Using desktop virtualization to give more students PC access
Brian Beck
IT Director, Greater Educational Opportunities Foundation
The Greater Educational Opportunities (GEO) Foundation has helped foster support for Indiana's current charter school law and now sponsors three charter schools in the state, as well in Colorado‚ totaling more than 1,300 students. As GEO grew larger, its computing infrastructure strained to keep up with demands. GEO was faced with supporting multiple schools and server farms, growing storage needs with expensive cooling requirements, and dealing with the issue of desktop backups becoming more difficult and expensive every year. In light of these costs and complexity, GEO decided to implement a new infrastructure and desktop replacement system.
Some of what Beck's team did was commonplace: getting the required storage and networking equipment and integrating it with current infrastructure, and virtualizing its current servers. But Beck's team also deployed desktop virtualization via Citrix's XenServer and XenDesktop software. The technology and the money saved through the more efficient infrastructure combined to let GEO provide a 1:1 computer-to-student ratio. Previously, it could afford only one computer for every four students.
Revitalizing a broken IT staff -- just in time for a make-or-break business launch
Amy Bourne
Vice President of Technology Development, Centro
When Amy Bourne took over as Centro's technology leader in March 2009, she inherited a software development team in disarray. Trust had eroded within the group, and the team's credibility within the organization had decayed. Bourne acted swiftly to address both organizational and personnel problems. In the course of 60 days, she restructured the team and filled major leadership roles -- even convincing software engineers who had previously quit in frustration to come back. She forced teams that had been unable to resolve differences to come to the table and address them.
Among other issues, the team had been struggling with productivity and consistency in their application of agile methodologies. Amy sent key team members to Scrum training and tasked them with internal Scrum training and rollout. She provided executive-level cover to enable the team to focus on the transition. She identified the need to bring in more experienced agile practitioners and recruited adept, talented individuals.
As a result of her efforts, the team's culture and productivity has been totally transformed -- and that change proved critical in late 2009, when Centro had a window of opportunity of just a few months to get into a new business. Bourne's team was charged with taking on an aggressive role in this plan.
Bourne told her team it needed to design and develop a commercial version of Transis, Centro's media-buying automation platform, to the advertising industry on March 1, 2010. Not only was this an ambitious timeline, but the team had also decided to redesign the internally oriented system for the commercial release. Under early leadership, the team took more than a year to deploy this system, but Bourne asked them to make it happen in three months. The launch of Transis was the defining moment for Centro, and Bourne's leadership enabled success in this endeavor.
Predictive technology avoids the recession's impact
Chris Brady
CIO, Dealer Services
In the last 18 months, Dealer Services Corp. (DSC), an inventory finance provider for used automobiles and other products sold by independent and franchise dealers, deployed a predictive analytic technology to help in forecasting revenue and make real-time adjustments to the business, as part of an ERP consolidation effort led by CIO Chris Brady. She realized that the ERP consolidation provided the opportunity to do more with ERP than DCS had one previously, not just reduce costs and complexity.
One of the biggest benefits of implementing the new system was the competitive edge it gave DSC in the face of an economic downturn: Predictive analytics enabled the company to anticipate the automotive downturn well in advance, giving managers time to adjust financial reserves, tighten lending standards, and scrutinize each dealer's inventory -- minimizing risk both to itself and its customers. DSC is more profitable today because managers knew how much revenue they expected to generate.
Also, the ERP consolidation effort replaced the previous manually generated reports with documents that are automatically distributed to executives and branch office users and improve visibility into upcoming market trends, while eliminating hundreds of hours of weekly labor. The new standardized financial and operational reporting system replaced its traditional spreadsheet reporting with metrics dashboards that displayed key performance graphs, using WebFocus technology from Information Builders.
Making a BI consolidation happen in a live environment
Aris Buinevicius
CTO, ChannelAdvisor
In the last 18 months, CTO Aris Buinevicius spearheaded an effort to centralize transactional data from multiple disparate sources into an actionable business intelligence data warehouse at ChannelAdvisor, which helps retailers sell online through multiple channels such as online marketplaces, paid search, comparison shopping, and retailer Websites.
Buinevicius pushed ChannelAdvisor's BI team through an aggressive cycle to build the data warehouse in-house, given that resources for external systems and developers were few and far between during the recession. He also required the team to migrate multiple channels and retailers (while retailers were continuing to sell online), pull in various data sources, and unify all those sources and processes. Coordinating this was an incredible challenge, considering the nuances of each data source; for example, every data source had its own style of representing a transaction.
The result is that where before it was difficult for ChannelAdvisor's online retailer customers to look at the full scope of their online campaigns, now they can slice and dice the data and run different modeling to see where products are weak or strong. For example, a retailer can now quickly view channel trends on a seasonal basis. Thus, if iPods are selling out on eBay around the holidays but not moving at all on PriceGrabber, the retailer can transfer more funds and resources over to eBay around the holidays to sell more iPods. ChannelAdvisor as a company can also aggregate the data across multiple customers and verticals to view industry trends and better educate the e-commerce industry.