In the last few years there has been a surge in the need for tools to support SOA governance. IT managers have sought to consolidate their SOA governance burdens by utilizing dashboards that bundle both SOA governance and monitoring solutions in one strategic package. These dashboards provide a clear and friendly interface for SOA management and governance tasks, offering more streamlined and prolific ways to carry out mundane tasks such as change management and versioning, monitoring and cataloging SOA components in an intuitive and highly organized manner. However, SOA governance runs much deeper than the obvious classification and cataloging of services for maximum reuse, transparency, version control, and functional correctness—there is a bigger picture. As organizations allocate more and more of their IT budgets to Service Oriented Architectures, they are discovering that substantial portions of their IT infrastructure will also have to undergo increased scrutiny and change in order to compliment SOA and ensure that the two evolve together in harmony. In today’s massively distributed and heterogeneous world, SOA governance must focus on the interoperability between corporate messaging systems—with a mix of Message Oriented Middleware (MOM), application servers, ESB, web services, Java, .NET and even home grown applications —and their effect on the organization’s portfolio of web-based business services.
Let’s expand on this last idea: In order to effectively manage your SOA, you need the proper metrics and measurements about every service that has been implemented. In order words, you need to understand, in a more universal sense, the IT infrastructure and architecture (as a whole) that supports and interacts with your services. SOA architects need to clearly understand:
- How services are integrating into in the larger context of enterprise technology architecture
- How well services are performing in supporting their constituent business processes
The only way to glean insight into these two critical subject areas is to establish the ability to continually monitor the flow of transactions and messages throughout the enterprise (across all infrastructures) and understand their interaction with services—from both a consumer and subscriber perspective. In most large companies, Message Oriented Middleware has evolved as the most common platform-independent way of linking systems in SOA environments.
However, just as enterprise services have become, in most infrastructures, a tangled web of siloed and undocumented methods, code, and functions, various flavors of middleware will exist throughout the corporate global infrastructure, residing in a similar state of decentralized and unwieldy discord which quickly becomes impossible to manage effectively. For just about all global business conglomerates, the current SOA reality justifies the need to monitor the performance of SOA and its supporting messaging infrastructure as a whole—on one unified tool—which offers change management, governance, and performance measurement. After all, SOA infrastructures are increasingly becoming more multifarious and complex; the result is a foregone conclusion: they are never going to become easier to manage. What still remains elusive to most companies is a true end-to-end visibility into distributed application performance—from both a transactional and business process perspective—where administrators and architects can immediately spot performance issues in a reactive or proactive mode of operation.
A useful SOA governance dashboard should provide users with configurable and customizable management options, offering plenty of holistic views of service performance from both business and IT points of view. Furthermore, it will be helpful to understand how a service’s performance (especially failures of the service) may impact various IT and business processes.
Common functional considerations of the dashboard should include:
- Web-based and user-defined (custom configurable) dashboard/operations console for real-time monitoring to ensure operational business continuity and control. Robust event monitoring should offer:
- Insight into the cause or origin of a problem
- Future impact analysis and assessment
- Historical milestoning of all previous impacts
- Categorization of events into various types or classes
- Customizable escalation paths based on incident type
- Customizable severity levels attached to incident type
- Robust meta data about all services and their relationships
- Glossary to support all service meta data and semantic references
- Simple visualization of complex IT relationships and impact on business processes at any junction in the lifecycle of a given business process.
- Service uptime, throughput, and other IT-centered performance metrics.
- A common platform where real-time performance monitoring, metrics and KPI can be collected, aggregated, and presented in legible and customizable fashion. KPI’s should:
- Give a view of how a business process is behaving and being supported by a service
- Raise alerts based on a pre-defined threshold.
- Robust, granular security, authentication and access control (supporting Kerberos and LDAP)
The “best of class” SOA governance dashboards will be able to manage different deployments and platforms of services, combining both monitoring and governance into one dashboard. Currently in the marketplace there are only a small number of companies that offer this unique bundling of functionality into a consolidated portal or dashboard. Right here in New York, I found a recognized market leader in application transaction monitoring and management applications that provides just such a SOA governance dashboard solution. Nastel’s Business Event & Metric Processing engine (BEMP) helps companies lay a sound foundation for SOA governance and includes a plethora of additional functionality which enables the monitoring, automation, visualization and fault recovery for all of an enterprise’s business integration components (middleware, application servers, rules engines, and more.)
What intrigued me the most about the Business Event & Metric Processing enginewas the tool’s correlation engine, which can combine metrics and data from many diverse SOA implementations enterprise-wide, thus providing a remarkable transparency into the service architecture from a single and consolidated locus of control. The user is able to feed information from a variety of sources and detect event and metric patterns in real-time at high speed -- processing millions of metrics & events per second. This allows IT to be more agile in relation to business requirements. Albert Mavashev, the CTO of Nastel iterates, “We provide the means for an organization to keep their entire SOA environment visible and manageable from a single pane of glass.” And with respect to my points about SOA performance metrics and KPI, Mr. Mavashev concurs, “A dashboard geared towards SOA governance must not only help make sense of complex and heterogeneous SOA environments, it must offer IT operations the ability to monitor and measure the success of how well their services are supporting the business. The key we are trying to accomplish is to reduce Mean-time-to-problem-resolution and accelerate time-to-value.”
As we can see from the detailed wish-list of dashboard functions above, good SOA governance is not just about high-level policy and best practices in coding or documenting of services, it is a twenty four-hour mission critical culture of monitoring, measuring, and management. Due to its capaciousness and direct impact to the business, it should not be difficult to grasp the potential return on investment (ROI) on SOA projects. (This at least holds true for most senior level technical managers, although some deal of evangelizing may be needed to win the hearts and minds of non-technical executives.) What is important is that everybody understands how SOA governance supports and rolls up into both IT governance initiatives and larger corporate governance agendas. In addition, it must be clear from the outset how the organization will measure SOA quality, as well as the conformity of their SOA to their governance goals; without clearly defined indicators of performance, SOA projects will become brittle and unscalable all too quickly.
About the Author
William Laurent is one of the world's leading experts in information strategy and governance. For 20 years, he has advised numerous businesses and governments on technology strategy, performance management, and best practices—across all market sectors. William currently runs an independent consulting company that bears his name. In addition, he frequently teaches classes, publishes books and magazine articles, and lectures on various technology and business topics worldwide. As Senior Contributing Author for Dashboard Insight, he would enjoy your comments at firstname.lastname@example.org
Copyright 2008 - Dashboard Insight - All rights reserved.