With the increasing complexity of Enterprise networks, there is a need for self-correcting and self-healing mechanisms that learn, predict, and plan. Cisco is announcing our newest SD-WAN innovation with Predictive Path Recommendation (PPR) powered by Cisco ThousandEyes WAN Insights. This is a significant capability to simplify network operations by leveraging recommendations from Cisco’s Predictive Networks. Predictive Path Recommendations provide proactive guidance for maintaining network stability and improving the performance of critical Application Groups distributed across the SD-WAN fabric. IT defines applications that require a specific SLA into groups so that PPR can predict which paths will meet those criteria.
Cisco SD-WAN provides IT with scalable, secure, cloud-managed WAN fabrics with extensive capabilities for visibility and troubleshooting of day-to-day network operations. The simplicity of management and exceptional Application Quality of Experience (AQE) are the key driving factors for all innovations underpinning Cisco SD-WAN.
AQE is achieved by constantly monitoring application path metrics and making intelligent choices among all the available paths. Cisco SD-WAN leverages existing capabilities of Application-Aware Routing (AAR) to adapt to unexpected degradation or outages by switching to the most optimal path. This ability to react quickly and automatically to changes in network KPIs provides an optimal Application Experience.
PPR, in combination with AAR, is a powerful tool that helps organizations optimize the performance of their wide area networks. One of the key benefits of PPR is its ability to generate long-term recommendations for network optimization. Rather than simply reacting to network issues as they arise, PPR takes a proactive approach, continuously monitoring the network and issuing recommendations whenever a better path is available. This helps to ensure sustained improvement over a long period of time. Figure 1 illustrates the three phases of the Predictive Path Recommendation cycle.
SD-WAN continuously monitors application behavior in relation to characteristics of all available paths within the WAN fabric and then generates long-term recommendations for paths that will reduce the probability of experiencing an SLA violation.
As changes to the WAN occur, the predictive models evaluate historical path metrics and usage to provide an early-detection system by warning of potential SLA violations before they occur and providing recommendations for alternate network paths per Application Group.
Network Admins/Operators can leverage the visualizations that are available in Cisco ThousandEyes and SD-WAN to view, monitor, and validate the effectiveness of the predictive model recommendations.
Operators select which policy changes that are recommended by the predictive models to apply in the SD-WAN fabric.
Workflow-Review & Application of Recommendations
PPR generates recommendations on a per Application Group per Site basis and these are available to visualize, explore and review before applying policy changes to the Network. From Cisco SD-WAN vManage UI, administrators can launch the Predictive Networks tab to view and explore all available recommendations.
SD-WAN administrators can find additional insights into the historical performance of the current path versus recommended path in terms of path quality and impacted users specific to an Application Group at a specific site. In addition, the aggregated metrics for the entire site are also available, which helps Admins identify circuits and paths which are problematic. This view is helpful in understanding the impact of policy change based on model recommendations.
Path and Quality of Service (QoS) details for path endpoints help admins verify the path recommendations. The visualization helps compare and correlate the historical Network KPI information presented with path quality variations, number of users, and application experience over time.
The Future of Connectivity Relies on Self-Healing Networks
Integrating Cisco ThousandEyes Predictive Path Recommendation with Cisco SD-WAN vManage provides IT with a proactive solution with actionable recommendations to reduce disruptions in network fabric while simplifying network operations. The predictive solution helps to improve the application experience by avoiding network degradation before it happens. It enables operations personnel to work more efficiently and to focus on strategic activities rather than reactive triage. Moreover, Predictive Path Recommendation provides the foundation for intelligent closed-loop network automation.
Learn more about Cisco SD-WAN, Analytics, and WAN Insights:
Cisco SD-WAN vAnalytics Animated Explainer Video
Tomorrow’s Network Today: Announcing ThousandEyes WAN Insights
Analytics and Automation: Driving SD-WAN Success at the Network Edge Podcast
The Network Is Core to Digital Transformation Success (Cisco ThousandEyes)
Keep up with the latest in networking news, subscribe to Networking Blogs email list, and get curated content from networking experts at the Networking Experiences Content Hub.
Hello, are there any plans to incorporate observed transport bandwidth into the dynamic AAR? Transport speeds / bandwidths are not fixed and can fluctuate in realtime depending on transport type (LTE, broadband, etc.) We have seen behavior where Cisco SDWAN prefers a slightly less latent or jittery circuit over one that actually has much greater bandwidth.
Thanks for your comment and it is very relevant. Yes team is working on the observable bandwidth as the parameter in AAR. It will be more a configurable option to have this in AAR algorithm as different deployment have different needs.