The Network is the Application, After All

The growth in web services, XML, distributed applications, and SOA in general suggest that network management and performance issues will only become more complex over time, particularly for IT organizations contracted to provide specific service levels to their customers. Indeed, not only will monitoring the traffic that occurs with respect to a complex, transactional distributed process be difficult, actually ensuring that that process performs appropriately given the phalanx of security and verbosity of the very XML-based messages sought after by the SOA faithful will be tricky as well. Thus, we expect the following:

“Further advances in technologies that consolidate and enhance data/message translation, integration and transport: As applications become more complex and chatty, the data flowing among them will continue to become more complex itself. Despite XML’s status as the intermediary content structure of record, it is quite verbose, and applications and networks will have to address an ongoing onslaught of such content. Additionally, rich media and voice will continue to infiltrate networks and become increasingly integrated with applications. This presents several issues: a need to speed transmission of the content; a need to perform higher speed translation and transformation of content; and a need to apply various forms of security ranging from authentication to digital rights management (DRM) across the network and span of applications.

“We believe that more progress will be made relative to application performance management as well as gather ing/managing the related data: There is a significant opportunity to build better management capabilities into the next generation of web-services based composite applications. Indeed, the need for better management will be exacerbated by the inherent distributed and dynamic nature of the services and the ensuing applications. Similarly, the ability to pinpoint faults and performance issues through out the network based application will be critical, and given the ever-increasing number of servers, applications and net work infrastructure components, etc., collecting the vast amounts of data and making sense of it for network man agers and administrators will be important as well.

“Relating performance to business process in such a way that business critical applications and processes will be monitored and reparable at a much higher level than previously: One could argue that “previously” doesn’t exist, though numerous vendors possess rudimentary capabilities to understand at least what hardware or software is responsible for performance issues within a particular business process. However, most approaches try to infer this from issues occurring at a device level rather than creating and analyzing a history of application performance relative to a broader set of issues (network traffic, server constraints, etc.). As more applications become process aware enabling users to define and manage (or orchestrate in composite application terms) processes at a higher level – management capabilities will be able to track more closely to elements of those processes (events, data movement, integration points, etc.). Ultimately, a performance-related view of an entire process will be available, enabling IT groups to attach specific costs and SLAs to the process (thus making ventures like utility computing more tenable)

“Appliances will play various roles in this arena ranging from performance monitoring and remediation to application and content accelertion – even as they are consolidated: Various appliances have cropped up across the network in the application management, security and performance spaces. They are attractive due to their low maintenance and implementation costs as well as their performance. Additionally, they can be placed at strategic points on a network (e.g., at the switch) to enable more specific analyses to be made regarding performance. Ultimately, just as consolidation will occur at higher levels (per the next bullet point), it will also be a factor among appliances as they combine facilities for security, data translation, etc. to offer solutions encompassing more of the stack.

“Consolidation in network management will continue as major players seek elusive “end-to-end” solutions: With IT organizations focused on ITIL-based governance, the ability to enable IT to present a pan-infrastructure view of performance – or business service management, as some term it – to business users and executives is key. Acquisitions made across a broad swath of space by Cisco, Juniper, Citrix, EMC, CA, IBM, and others helped to drive this effort in 2005, and we expect more of the same in 2006. Indeed, we believe that network management, application performance/management, configuration management, business service or activity management and other related areas (root cause analysis, for example) will increasingly be found – and integrated – under a single metaphoric roof.

Dashboard-based integration will offer system administrators and others a holistic view and analysis of overall business performance relative to infrastructure.

Clearly, the fragmentation of the overall systems management market, compounded by the increasing complexities of networks, applications and content will continue to feed a strong M&A market over the next two years.