Reporting is a vital function for IT Services, particularly for metrics and strategic planning. Insight into service usage allows any business to make smarter decisions about the future. In the IT world, there are many specific types of reports, such as web statistics, which answer questions like "How many time was the Stanford home page read today?" and network traffic analysis, which can tell a system administrator what the network load was over the course of the last month. However, to get the most of the information being collected across an organization, reporting on a broader scale involves pulling together multiple sources of data, analyzing it, and presenting the results in a clean, easily comprehensible way. Questions like "How many undergraduate students are using mobile devices?" or "How many staff in Business Affairs at Stanford have laptops?" may appear to be simple, but because the data is contained in disparate sources and in different forms, getting the answer turns out to be a surprisingly challenging task.
In fact, a whole industry has grown around the development and marketing of "business intelligence" tools. In concept and functionality, the reporting infrastructure that IT Services is in the process of exploring is quite similar. But "business intelligence" conjures up a set of heavyweight applications and large suppliers, which is not the low-cost and open-standards approach being pursued. Hence, this particular term will be used in a somewhat different way in this vision.
At the moment, the reporting in IT Services is limited to vertical solutions for specific needs. For example, the web group uses Google Analytics and other statistics packages to analyze web data; the finance department has financial tools to keep track of how much money is spent on a given service. On an individual level in IT Services, someone might query Microsoft Access and generate a pie chart. However, there is no reporting system in place that can search across data from different workgroups; each group's data is an island unto itself. Where requests come up that require analysis across these groups, it falls to a staffer, who must expend considerable time and energy to come up with a manual technique in order to create a report that is, more likely than not, a one-off — all of which is antithetical to the automation and efficiency ideals of IT Services.
From a systems architecture standpoint, there are three subcomponents to a reporting system: a database layer, application layer, and display layer. For a robust reporting system, there is the data warehouse, then the analytics subsystem, where the data is sliced and diced; and finally the rendering of the results, which includes graphical tools that allow users to see the particular data that are interested in, in the form of graphics, text, tables, and so forth.
While complete reporting systems that provide all three elements "out of the box" are available from companies like Peoplesoft and Oracle, these products are extremely expensive and not well-suited for the data environment currently in place within IT Services; these are very powerful tools, but were designed for the business units in organizations that run large ERP (Enterprise Resource Planning) applications that manage human resources, finance, supply chain, and so forth, which are not in the systems profile of IT Services. And in the last four years, web development has evolved considerably, with open source software now underpinning many mission-critical applications. Major companies like Google, Facebook, and Yahoo! have followed the route of leveraging open-source technologies/products, a more nimble approach that does not tie organizations to a particular vendor. This approach is especially valuable as the systems inevitably have to be customized. The cost-savings of this approach is one consideration, but it is not the primary consideration.
IT Services' Metrics, Monitoring, and Reporting project, which is currently wrapping up Phase One, has developed a reference implementation of a display layer which begins to demonstrate how the reporting infrastructure for IT Services will be implemented in the near future. This display layer system is based on the OpenSocial Container specification which has, over the past two years, become extremely popular with an increasing number of software manufacturers and SaaS providers. In Phase Two of the Metrics, Monitoring, and Reporting project, IT Services will select and deploy data warehouse and data analytics subcomponents which will aggregate data from many sources and allow that data to be queried and analyzed across business functions. There are several good candidate products for these key subcomponents. In this next phase, the project will analyze the various data sources to determine how they can be best conjoined.
While this broad reporting infrastructure will provide both IT Services and campus administrators with a much more comprehensive, rich look at service usage, the personnel requirements are anticipated to be lightweight relative to ERP industry products. Rather than have a centralized staff that is devoted just to this project, the manner in which it is being developed allows current staff — who already own the data and are in a position to best understand it — to take ownership of the reporting. These core members, who will form a cross-disciplinary, virtual team, will include the Data Center's new staffer in charge of Capacity Management, someone from Financial Systems, and someone from Systems Administration (who will take on the applications integration role). The hope is that by successfully implementing this reporting infrastructure, other campus organizations may be able to take advantage of IT Services' expertise in tackling what is a very common situation: plenty of information, but no easy way to make sense of it.
- Complete Phase One of the Metrics and Monitoring project and demonstrate presentation of disparate metrics into a unified, modular dashboard display layer.
- Move forward with Phase Two of the Metrics and Monitoring project to investigate data warehouse and analytics tools that are good candidates to support the vision.
- After selecting data warehouse and analytics subcomponents, deploy these elements and inventory data feeds from the various sources of operational, service, and other business metrics.
- Design data structures and develop data dictionary for warehouse.
- Create framework within these new tools that can perform cross-functional query and data analysis for the purpose of rapid use/reuse by IT Services staff and others with authorized access.
Measures of success
- Ability to create analytical reports derived from disparate data sources.
- Provide the ability for IT Services' clients to generate their own reports in a self-serve fashion.
- A cross-group team is in place to advance the effort and meets regularly to improve and enhance reporting capabilities.
- The development of a data dictionary in order to standardize data for querying.
- Current service metrics that are readily accessible and understandable by personnel outside of IT Services.