Guidance for Development and Management of Sustainable Enterprise Information Portals (2018)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

64 The transition to cloud-hosted EIPs entails changing culture, policies, and processes, an effort that may not be successful on the first attempt. States look to peers within state or city agencies (e.g., other departments such as housing or health) and counterpart agencies from other states, as well as to commercial sectors such as finance. This section presents three case studies on the move to an EIP environment: • TxDOT’s STARS II. • Utah DOT’s data management and storage migration. • Virginia DOT’s snowplow status service. These case studies are presented in Sections 6.1, 6.2, and 6.3, respectively. 6.1 Texas DOT Statewide Traffic Analysis and Reporting System The Transportation Planning and Programming Division is the Office of Primary Respon- sibility within the Texas DOT (TxDOT) for collection, analysis, and reporting of traffic data to meet federal requirements to develop, establish, implement, and continue operation of a traffic monitoring system (23 CFR 500 Subpart B). The Statewide Traffic Analysis and Reporting System (STARS) was developed by the Texas Highway Department and the Texas Transportation Insti- tute in the 1960s to collect and maintain data coming from a $50 million traffic data collection infrastructure. The program had statewide coverage, using • Short-term traffic monitoring data from contracted pneumatic tube axle and manual classifi- cation counts, as well as in-house special use counts; and, • Permanent counters at 362 sites that collected volume, vehicle classification, speed, and weigh-in-motion data. By 2005, the STARS system included multiple subsystems that were not integrated and used a mainframe computer that at times required parts that were no longer manufactured. This system required highly skilled programming staff to code in obsolete computer languages. Escalating system complexity and the limitation on data file types increased cost and time for new processes or minor modification of existing processes to meet new requirements. Furthermore, certain valuable new analytics and complex data could not be accommodated due to the limitations of the system. Each year, approximately $5 million in new traffic data is collected, stored, and analyzed by the STARS system. STARS was being used by the state legislature, the governor’s office, county judges, FHWA, TxDOT districts, TxDOT divisions, and the public for environmental, financial, maintenance, planning, operations, safety, and economic analyses and decision making. The DOT Portal Use Cases S E C T I O N 6

DOT Portal Use Cases 65 increasing number of users and varying user needs were outpacing the capabilities of the under- lying system, which was quickly becoming unsustainable. To address these challenges, the Trans- portation Planning and Programming Division explored how best to proceed. 6.1.1 Adopting an Innovative Strategy New to TxDOT STARS II included the purchase, customization, and installation of a web-based commercial off-the-shelf software for statewide traffic analysis and reporting, as well as the housing of data exclusively using cloud services. A business case was developed in 2007 to secure funding from the Texas State Legislature. The business case estimated a minimum financial benefit of $2.3M net present value based on a 10-year life cycle with a two-phase funding scheme with $1.9M for Phase I and $635K for Phase II. The preferred alternative was a vendor-hosted solution in the cloud that maintained TxDOT ownership of the data. System development was initiated in 2011, established stability in 2015, and, as of 2016, has received over 30 million hits per year. STARS II replaced the legacy main- frame subsystems, personal computer applications, and over 100 manual processes. In doing so, STARS II • Substantially reduces the potential for errors in data, • Reduces the time required for data analyses, • Improves business processes, • Leverages mapping and reporting tools for enhanced accuracy, and • Ensures access to a centralized traffic database free from data duplication and redundancies. The proposed solution leveraged the Texas Department of Information Resources deliverables- based Information Technology Services contract used for IT projects, rather than using the more common design-build contracts traditionally used for TxDOT construction projects. Data Trans- fer Solutions, with subcontractor Midwestern Software Solutions, was awarded the contract in January 2013. Cloud hosting has saved TxDOT approximately $310K over a 5-year period. With STARS II metadata, efficiencies were observed in the process of verifying contractor delivery of clean and complete data and in payments to contractors based on performance and delivery of quality data. 6.1.2 Challenges to the STARS II Adoption Impediments to the successful approval, implementation, and use of the STARS II system (ordered from most to least significant) included institutional, procurement, technical expertise, and implementation factors. STARS II champions succeeded in making the business case for the program by highlighting cost savings and system reliability. The fact that system data on the cloud included only public data and no data specific to individuals also eased concerns about data privacy. Prior to STARS II, TxDOT had not used cloud services to host massive quantities of data. Most significant institutional concerns were specific to data security and ownership. Migration to the cloud required a waiver from the State Director of Information Resources and the Texas Information Technology Board. Traditionally, IT services acquisition efforts favored lowest bid awards; however, this approach proved unsuccessful due to technical underperformance. A sec- ond acquisition, focused on a best-value approach, proved successful, resulting in the current STARS II. Funding for this project came from the state legislature, which operates on a 2-year budget cycle. To successfully execute the program vision, TxDOT staff navigated these complex budget cycles, remained vigilant in adhering to administrative processes, and championed the STARS II program.

66 Guidance for Development and Management of Sustainable Enterprise Information Portals The absence of enterprise information management expertise within TxDOT was recognized early by the STARS II program champion. A project manager with IT program execution exper- tise was engaged through a consultancy contract to support the planning and the migration for a term of 1.5 years. The IT program expert led the conduct of a gap analysis to determine whether business requirements were being met or not and the steps needed to ensure that all business requirements were successfully met. The IT program expert worked with TxDOT transporta- tion experts to define the requirements of the system and prepare the contract with the vendor. Significant change in work flow processes required transportation analysts to gain comfort with higher resolution of data types and quality. A collaborative approach supported by the organi- zational structure of the project facilitated process enhancements. The migration to the cloud enabled a clarity and depth of both data and analyses not feasible in the legacy system, which surprisingly led to additional challenges. The improved data and analyses required significant re-baselining of transportation system performance measures. The federal rollout of the state Traffic Monitoring Standards in 2013 compounded the complexities arising from increased data precision and accuracy in STARS II, resulting in adjustments to state data. As previously mentioned, much of the data required a re-baselining, and with the new federal standard, data merited a large correction in previous year data points. This re-baselining was initially challenged by FHWA, which required an effort to explain why the adjusted mea- sures were most appropriate. However, these challenges were overcome, and more precise and accurate data and analytics capabilities have led to far more effective operations. 6.2 Data Management and Storage Strategy for the Utah DOT In alignment with the State of Utah’s executive branch migration to use of cloud-based ser- vices for data storage and in consideration of the potential for significant cost savings, the Utah DOT (UDOT) embarked upon shifting data storage and access to a cloud environment. UDOT uses on-premise data management systems, a custom third-party-developed cloud storage ser- vice, and Google Drive services to store the significant data generated by the agency—ranging from construction-related documents and right-of-way and materials-submittal documents to road weather and traffic conditions. The cost of over 4 terabytes (TB) of data storage along with software licenses to access the data had been approximately $500,000 annually for UDOT. Transitioning to the custom, third-party-developed cloud storage service led to a 12% savings through the elimination of license and on-premise hardware and maintenance costs. At the same time that UDOT needed to find more economical ways to store its growing quan- tity of data, the agency also faced increasing complexity in its data. Drawing inspiration from state and private healthcare practices in Utah, UDOT overhauled its data governance by defining a decentralized strategy that also puts in place good data practices and integrates a team of data experts into UDOT’s business units to improve business unit value through more efficient data management. Sections 6.2.1 and 6.2.2 summarize the environment, drivers, and challenges with the evolving UDOT data management and storage strategies. 6.2.1 Keeping Data Management Close to Business Practices UDOT, like most state DOTs, has been facing challenges managing its data because data management was outsourced with the IT infrastructure management. When data management is centralized and/or outsourced, it is taken away from the staff that understand the specific needs of each business process. The focus of IT management was on software maintenance and

DOT Portal Use Cases 67 user services, not on data management. Moreover, IT structures were demanding standards and processes that were inconsistent with business unit data needs and use. Because of the situation described above, UDOT looked at the practices of other industries and found inspiration in the practices used by Intermountain Healthcare (IHC), the major health care provider in Utah. IHC sets a good example of keeping data management close to business processes. On the basis of research and experience, IHC concluded that even if IT needs to be outsourced or centralized at a corporate or statewide level, data management should be kept close to the staff who are most familiar with the data. 6.2.1.1 Adopting an Innovative Data Management Practice at UDOT After consultation and collaboration with IHC, UDOT began to consider changing its prac- tices to separate data management from IT and bring data coordinators and staff into each business unit. There was significant data generation and use at UDOT, but limited coordination across its three business groups and four regions. UDOT hired a Department Data Coordinator to coordinate data management strategies throughout the DOT. Moreover, UDOT is identifying small teams of data management experts within each business area to support data management for specific business processes. These experts will be closely familiar with the data itself, data governance, metadata, and any other specific needs for each business process. The focus will be on the whole data life cycle. The Department Data Coordinator and the embedded teams will serve as data stewards who help groups ensure correct interpretation of data, encourage data openness, and promote coordination of data among business groups and regions along the whole data life cycle. 6.2.1.2 Challenges to the Adoption of a New Data Management Paradigm While IHC dedicated three individuals within 10 business units to serve as data experts and stewards that report directly to the Data Coordinator, UDOT does not have the resources to employ dedicated data management staff in each of seven total business units and regions. Cur- rently, each business unit and region has staff who serve as de facto data experts, often taking on this responsibility informally while meeting other assigned responsibilities. One challenge the new Department Data Coordinator will face is rallying staff who do not report directly to him/her to commit attention to the data management and governance effort. Furthermore, there is the technical undertaking of understanding what data are available, the quality and completeness of data, and the value of data across business units and regions. Most significant may be the challenge of data openness, because divisions and groups are not accus- tomed to providing to other units granularity and transparency in data specific to their business processes. There is a cultural fear that showing detailed internal data may reveal data quality and business operations shortcomings that will reflect poorly on the business units. The goal of the data stewards is to illuminate the process, highlighting opportunities for growth and efficiencies by making data visible both internally and externally. 6.2.2 Migrating Data and Software to Cloud Hosting The Utah State executive branch encourages the use of cloud-based services for most of its data storage to realize cost savings. The Department of Technology Services (DTS) offers two new storage solutions: (1) a new primary storage area network (SAN) storage solution in the Salt Lake Data Center and (2) a new DTS object storage product that offers private cloud storage. While the DTS object storage solution does not enable autoscaling to demand, it does offer efficiencies through consolidating data storage across the entire DOT. Currently, UDOT stores upwards of 4TB of data.

68 Guidance for Development and Management of Sustainable Enterprise Information Portals 6.2.2.1 Innovation in Data Storage As UDOT migrates to the DTS service, it leaves behind expensive storage solutions such as SAN for less costly commodity hardware-based storage solutions used in the DTS private cloud. UDOT will also procure SaaS through third-party contracts. The migration will likely reduce costs by as much as $100K, or 20%, annually. 6.2.2.2 Challenges in Migrating Data Storage UDOT has limited control over the structure of the data storage or the hosting environment and typically relies on DTS to provide options. An alternative option, if DTS is unable to meet hosting needs, is for a consultant to purchase the product. The consultant contract can stipulate that after a certain number of years, UDOT can take over the maintenance of the product. Migration of data and software is anticipated to span 6 months while the migration is verified and tested. The migration also will require development of a new service level agreement address- ing challenges specific to metadata management, project documentation, and connectivity. 6.3 Virginia DOT Real-Time Snowplow Tracking Tool On January 25th, 2012, the Virginia DOT (VDOT) added a new web tool to the winter travel information page of its portal. This web tool allows Virginia residents to find out where snow- plows have been and where they’re going in the Northern Virginia and Fredericksburg districts. The web tool uses the 600 Northern Virginia “snow maps” that VDOT assigns to snowplow drivers and the automatic vehicle location onboard snowplows to display in real time their cur- rent and future locations. Once it snows at least 2 inches, residents can view the snowplow tracker to find out the status of plowing in Fairfax, Loudoun, and Prince William county neighborhoods at http://novasnowplowing.virginia.gov or http://vdotplows.org/. Figure 6-1 shows a screenshot of the VDOT snowplow tracker. Figure 6-1. VDOT snowplow tracker screenshot.

DOT Portal Use Cases 69 Users can enter an address in northern Virginia to see a color-coded snow map that indicates the plowing status in that area: • Green indicates a neighborhood has been plowed, • Yellow means plows are in progress in the neighborhood, • Blue indicates plows have not yet started the neighborhood, and • Gray means the area is not maintained by VDOT. Cities, towns, and some developments maintain their own roads. 6.3.1 Current System Setup The VDOT snowplow tracker is currently operated from the Microsoft Azure cloud located on the east coast of the United States. It was developed using a commercial set of software from Microsoft: • Internet Information Services web server • ASP.Net web framework • MapPoint mapping framework • Bing maps online map service • Azure blob storage cloud storage The snowplow tracking system is set up as a monolithic system running on a cloud envi- ronment virtual machine sized to accommodate the estimated demand for snowplow tracking information. 6.3.2 Challenge and Opportunity The challenge of the VDOT snowplow tracker website is that the application is not being used by Virginians outside of periods of falling snow; yet, when it is in demand, the application needs to be able to fulfill a few million requests in a few hours coming from Virginians anxious to know when their streets or nearest highways will be plowed. To handle such demand using a monolithic system, it would be necessary to provision on the cloud a large system and a high bandwidth connection capable of handling the surge of a few million requests in a few hours, even if the need for such a large system and high bandwidth occurred only on a few days a year. VDOT would have to expend significant funds to acquire and maintain a large quantity of expensive resources that would be used infrequently. This could be a difficult case to make, come budget time, despite the popularity and effectiveness of the web tool. A possible solution to this dilemma would be for VDOT to adopt a microservices approach for the snowplow tracker tool. By migrating its current monolithic setup to a set of containerized microservices that could be deployed on the current VDOT cloud provider container manage- ment service, the VDOT snowplow tracker would then become an “elastic” web tool. It would be capable of running two or three instances of the snowplow tracker container services during off times and automatically scaling to thousands of container services following a snowstorm, when demand rises. VDOT would then be charged by its cloud provider only for actual demand, that is, the actual number of containers deployed and the utilized network bandwidth. At budget time, it presumably would be much easier to garner support for a web tool that incurs charges from the provider only when it is used.