As software-defined networking (SDN) evolves, so does assurance. Performance management of SDN requires an end-to-end view of the network and applications to ensure that customer experience does not nose-dive due to network complications. In the SDN environment, performance management faces some hurdles, such as lack of visibility, troubleshooting issues, and analyzing performance data.
Performance management — also referred to as assurance, visibility, or performance monitoring — is the task of monitoring network’s and application’s performance via various tools, such as application performance management (APM) and network performance management (NPM). It collects data from those functions and presents them on a single pane of glass for network and application administrators to review and catch issues, hopefully before those issues affect the end user’s experience with the service.
The areas of performance management and SDN are changing. The issues arising from SDN necessitate the implementation of a unified performance management tool, which converges APM and NPM, and automated solutions via machine learning (ML).
Performance Management and SDN — Issues Needing Solutions:
Pulak Chowdhury, founder and CTO of Ennetix, noted that “in [a] dynamic SDN environment, there is [an] even greater need of end-to-end performance management.”
Virtualized environments are becoming more commonplace since the introduction of SDN, and ExtraHop suggests that virtualized environments hinder the APM tools’ ability to achieve its objective. According to the company, SDN “virtualized environments may split application workloads across virtual machines that can be spun up and spun down, or moved across the data center, making it incredibly difficult to pinpoint the source of performance problems.”
Larry Zulch, president of Savvius, told APM Digest in an interview that “in the past, NPM solution providers had called their products ‘real-time’ even though their dashboards had delays of several minutes. Those delays are no longer acceptable.”
Performance management expectations before the popularity of the cloud are now outdated due to the increased network traffic from cloud computing. Service level agreements will need to be revisited to update the application and network performance baseline in the cloud to end users.
Security threats remain an ongoing need for performance management tools to detect, alert, and resolve. As SDN continues to evolve and be used in the networking sphere, the need for performance management tools to locate security threats before damage occurs will be an increasingly daunting task. The expectation is that ML and artificial intelligence (AI) will automatically notice security threat patterns and will instantly alert the appropriate personnel.