Enterprise Architecture May 2015 Executive Committee May 20, 2015 Wednesday 1:00-2:00 p.m. 561 Smith Center
Agenda Meeting Purpose and Intended Outcomes The Interoperability Problem: Current State (15 min) The Future of Interoperability at Harvard (20 min) General Discussion (15 min) 2
Meeting Purpose and Intended Outcomes Purpose Explain current issues with interoperability and interfaces at Harvard Discuss future-state vision for operability Intended Outcomes Establish a shared understanding of our current interoperability problems Progress further toward agreement on a future-state interoperability solution Reach an understanding of financial accounting 3
Business Needs & Benefits Stakeholder Experience Today Imagine Service Users: Faculty, Students, Staff IT Systems and Policy Decision-Makers Systems and Software Engineering Operations Time-consuming, repetitive entry of biographical and other data for each application/service Inability to answer simple business questions ( how many of X does Harvard have? ) Difficult and time-consuming to access data not already provided via existing flows to user s organization Data Warehouse information is disseminated for each independent need, leading to delay and difficulty in providing large sets of data No facility for cross-university data ( what do we mean by address or student? ) Each integration is custom-coded, increasing technical risk So much time is spent on basic integration that tasks with greater user-perceived benefit are delayed Multiple ad-hoc streams are time-consuming and brittle, creating create complexity for each system Difficult to identify faults and restore service due to multiple file transfer servers/flows with hard-tounderstand interdependencies Users do not have to re-enter data apps get necessary info from a real-time service Faculty, students, and staff can ask questions on an ad-hoc basis and get reliable answers Faculty, students, and staff have flexible, timely access to data across the University Decision-makers have business intelligence based on data automatically based A common understanding of what our data means across the University Schools and departments have a set of reusable integration tools More efficient access to information frees up time to develop more user-valued services A centralized interoperation service available across Harvard, operated by a focused team No longer necessary to deploy, configure, and operate point-to-point integration services Cost-efficient, flexible access to consistent information across the University that is easy for users and less complex for engineering and operational staff. 4
The Interoperability Problem: Current State Please see the handout for a look at the current state of interoperability at Harvard. 5
The Interoperability Problem: Future-State Vision Please see the handout for a summary of our future-state vision for interoperability. THE INTEROPERABILITY PROBLEM: FUTURE-STATE VISION Establishing a series of interoperability services will provide Harvard University with a standard means of distributing data that will improve all development activities, simplify operations, and ensure that the right information is available securely based on need. The future of interoperability at Harvard will be: Fewer interfaces and more shared communications... File Transfer Consolidated interoperability services improve staff efficiency by 50% (Gartner), reuse saves approximately 200 hours per interface Standard approaches for interoperability in order to minimize developer effort A publish-and-subscribe model allowing information to be distributed on demand... using standard business objects... HBS Business MDMObjects Provides a registry of available data, Person making Enrollment it cheaper and Courseless complex for applications to get authoritative data Identification of a single authoritative Class Absence source for Seat Event Notification Assignment each information type A master data management solution for orchestrating updates to critical data definitions... distributed in a standard set of formats... Data Governance Provides a consistent definition H of the data that is managed What? Established data governance procedures When? that identify what data can be distributed and how How? Real-time, bulk-data service BI/Analytics require clean consistent data to produce accurate results... communicated using fewer protocols... Transactional Integration Consistent data transfer formats improve operations by reducing complexity by upwards of 1000% Web service protocols: REST and SOAP for real-time Asynchronous messaging: JMS, ActiveMQ queries Batch: Defined information patterns for data exchange communication x x x x... and introducing consistent security practices Integration Security Reduces the risk of compromised information H and increases confidence in data reliability Data integrity through assurance that information is valid Routine, reusable encryption methods Integrated security built into the solution Simplifying the interoperability of our applications is the only way we will be able to move forward in an agile manner, the only way to decrease our technical debt, and the only way to move incrementally into the future as we change our programs and infrastructure. Jim Waldo, Harvard University CTO Learn more about Harvard s enterprise architecture effort at http: /enterprisearchitecture.harvard.edu 6
Thank you!
Appendices
The Enterprise Architecture Vision Our Vision for Enterprise Architecture Provide a technology framework and a set of standards to enable acquisition, development, and deployment of IT services that maximize interoperation, minimize duplication, and simplify the IT environment across all of Harvard. Strategic Objectives Guiding Principles Key Performance Indicators Deliver an enterprise architecture framework that drives technology and development standards across Harvard Provide common approaches for integration across enterprise applications, processes, and data Align and rationalize technology decisions and investments Identify redundant or conflicting processes and data across organizations Ensure that EA provides active direction and delivers value to the organization Counter complexity with common solutions Enable sharing of data across organizations Preference for open-source, COTS, and programmatic interfaces both in what we obtain and what is produced Encourage, define, and ultimately provide best-practice solutions Evolve framework and solutions with advances in technology Decrease in project delivery timeframes to production Increase in the number of integrated applications using programmatic interfaces Increase in the number of funded projects that conform to an EA checklist Decrease in ad-hoc data sharing Increase in automated data exchange Increase in the number of known authoritative data sources Decrease in the number of copies of data 9
EA Program Approach 10