Mission-critical applications: Stability, evolution ... or both?
StoryApril 26, 2010
Tyson Moler
GoAhead Software
Due to long budget cycles and a growing deficit, defense systems must adapt to evolving threats by refreshing their current infrastructure while bolting on new capabilities to meet rising High Availability (HA) requirements. Open standards routes such as the Service Availability Forum's (SA Forum's) Availability Management Framework (AMF), along with COTS software, offer the best opportunities to success- fully meet this challenging paradox.
There is a tremendous investment in existing mission-critical applications. As they become networked, their criticality to the overall mission increases. Missions also change, raising the system status to mission critical. The net result is that aging systems are being tapped to perform at much greater levels of operational availability than their original designs intended.
The defense community has come to see open standards-based COTS technology as a solution to this life-cycle challenge. Historically, no open standard has addressed system-wide fault management and operational availability. Existing high availability infrastructures are often proprietary and difficult to adapt. COTS software helps deal with both of these issues, and also satisfies the need for unified and flexible management of system configurations.
The benefits of the open standards COTS approach to operational availability go beyond meeting current mission requirements. Future evolution is enabling – and enabled – in two important ways. First, it enables the applications to be transportable across platforms and eases the integration of future capabilities. Second, unified configuration management provides a faster and simpler approach to implementing new components and capabilities. Accordingly, the Service Availability Forum’s (SA Forum’s) Availability Management Framework (AMF) is working to support today’s mission-critical High Availability (HA) needs while also outlining numerous approaches to ease legacy software migration.
AMF and mission-critical HA
The SA Forum is a commercial software consortium that has produced a set of open standards that addresses mission-critical requirements for software-based systems with nearly real-time/real-time applications. The SA Forum has always been sensitive to the issue of legacy application support, and members of the group have identified a spectrum of support paradigms. They have additionally formalized various software redundancy schemes that support HA. The SA Forum’s work is now being implemented by the U.S. Navy and is a DoD-wide mandated specification in the DoD IT Standards Registry (DISR).
The SA Forum specifically provides an AMF service that supports subsecond software redundancy and fault management capabilities that satisfy the mission-critical requirement. AMF involves a system model that explains active/standby/spare component redundancy relationships and the policies that deterministically direct the system through fault detection, isolation, and recovery to ensure continuous service availability. System model particulars are crafted by an insightful system designer at development time. These redundant components manifest in the system as software processes or applications that could be newly developed, third-party, or legacy applications.
New applications are best served by linking in SA Forum service libraries, which provide the most time-sensitive fault detection and failover performance, as this allows for explicit application health monitoring and direct redundancy state assignment (active/standby/spare). However, many third-party or legacy applications cannot change; a nonintrusive integration approach is required.
AMF eases legacy software migration
AMF describes several approaches for legacy software migration, two of which are the Wrapper and Proxy-Proxied methods. External Passive Monitoring also provides a highly capable and nonintrusive method for providing availability services to legacy or third-party applications.
The U.S. Navy Common Processing System (CPS) program, in the context of Advanced Capability Builds (ACBs), is a great example of bringing new infrastructure and applications into an existing combat system environment. The CPS platform is a common computing environment for mission-critical applications, and it utilizes SA Forum-compliant middleware to provide software redundancy and dynamic fault management capabilities. CPS will be deployed in Aegis Modernization/ACB12 in the immediate future, and is expected to support additional combat system programs. This approach will greatly extend the life of Aegis-equipped cruisers and destroyers. The intent of the U.S. Navy Consolidated Afloat Networks and Enterprise Services (CANES) Common Computing Environment (CCE) is to achieve similar benefits.
Meeting the HA/legacy integration challenge
Overall, while legacy software migration is a challenging exercise, it is very possible to evolve mission-critical systems in a stable and evolutionary way. COTS solutions for high availability such as those provided by GoAhead Software and other vendors offer the best approach to these migrations. Open standards such as SA Forum’s AMF are paving the way for tomorrow’s mission-critical systems by providing for today’s HA needs while retaining the ability to ease legacy software migration.
Tyson Moler is director of federal operations at GoAhead Software. Prior to GoAhead, he worked in management consulting, supporting clients in business development, and program management in federal and defense markets. He holds a B.A. in International Relations from Claremont McKenna College. He can be contacted at tmoler@goahead.com.
