Monday, February 27, 2012

Considerations When Monitoring IBM DataPower

Effective and comprehensive monitoring of SOA environments that leverage the power, and complexity, of IBM DataPower appliances requires a multi-tiered approach. The various appliances deliver enhanced security, message transformation and routing and as key components of business critical applications, require close attention to insure high availability in all regards.

Coverage for DataPower appliances must include monitoring of services and resources. While the appliances expose a good number of metrics through SNMP MIBs, monitoring of the services they provide allows proactive alerting, to both Consumer and Provider communities. Operational teams focused on the health of the appliances must be informed of any issues related to the appliances themselves, such as power, cooling and hardware errors to proactively address any degradation in performance or availability.

Service monitoring of the appliances starts with engaging the application teams, Consumers and Providers, to establish the Service Level Agreement for the deployed services. Live traffic is then analyzed and factored against the requested response times and availability levels for all subscribed Services and Operations. This additional level of validation allows more comprehensive tuning and distribution of alerts, increasing service and operation availability and reducing Mean Time To Resolve (MTTR) in case of performance degradation.

DataPower appliances are most often monitored by a combination of agents, such as those supplied by IBM as part of the Tivoli Composite Application Management for SOA (ITCAM for SOA) Platforms package. ITCAM for SOA includes agents to monitor both Service execution and Resource Usage of the appliances, supplying invaluable data and insights to Operations, Application Support and Software Development teams. The Services and Resource agents are most often deployed on the same remote server and poll the appliances via SOMA and SNMP protocols. This architecture allows more centralized deployment and reduces license cost.

The monitoring team interfaces with a variety of teams, from Network Security, DataPower Support and Enterprise Architecture, SOA Governance and Application teams. Each operational team provides their own level of expertise to the architecture and deployment effort and their involvement helps insure a solid and well supported deployment.

Application teams in SOA environments are heavily invested in the success and availability of their services, so the overall effort will benefit from their early and frequent involvement. ITCAM for SOA provides unique insight into Consumer and Provider service response time and can assist in validation of the business logic and rules driving service architecture and deployment.

Dan Kern joined Prolifics as a Senior WebSphere Administrator and Performance Tuning expert. Over the past 7 years, he has held roles as a Technical Solution Director, Practice Director for the Automation and Systems Management area and now helps customers realize the full potential of their software investments as a Solution Architect. Dan is highly regarded in the Tivoli SAPM product and business areas and continues his focus on top quality solutions.