Global supplier of cloud networking solutions, Arista Networks, has unveiled a new software solution for network observability, CloudVision Universal Network Observability. It combines data from compute and server systems-of-record with network infrastructure performance to provide deep insights into workload and application performance across campus, wide area, and data center networks.
Enabling automation of network, systems, and application/workload visibility along with AI-driven proactive analysis and prescriptive recommendations, Arista’s CloudVision Universal Network Observability (CV UNO) system would speed up issue resolution for unforeseen events, minimizes human error, and provides accurate root cause analysis of network events and their impact on application delivery. The system is available as a premium feature license on Arista CloudVision.
“Bringing together multiple network domains with full application visibility and troubleshooting will streamline network operations and improve uptime and reliability,” said Zeus Kerravala, Principal Analyst at ZK Research. “Disparate operating systems and lack of consistent data models across networks have made delivering systems with this degree of visibility previously impossible. Moreover, in an era characterized by stringent regulatory compliance, cybersecurity, and observability throughout the enterprise – is no longer optional but rather an essential imperative.”
Comprehensive Overview of Application Stack and Network
Whether the source of the issue is network, platform, system, or application-related, CV UNO would offer special systems-level capabilities to lower corporate risk, enable quick fault identification and rectification, and streamline cross-functional coordination. These features would help speed up the time to recovery. Key user benefits would include the following:
- Workload Application and Infrastructure Discovery – Across a variety of platforms and IT systems for record-keeping and inventory management, CV UNO would automatically find applications, hosts, and workloads. The combination of this comprehensive data with CloudVision’s deep perspective of the networking condition inside CV UNO’s Arista Network Data Lake (NetDL) would create a composite image of the whole network and application stack. It would also create an application-to-network graph that is time series saved and updated in real time, providing a historical record of the environment’s development and status at any given moment in history.
- Proactive Risk Analysis – CV UNO facilitates proactive risk analysis, cross-referencing, and impact analysis of network faults and abnormalities using real-time Application-to-Network graphing. Before implementing potentially disruptive network modifications into mission-critical and production networks, their effects may be evaluated.
- Realtime Network Change Impact Analysis – This composite dataset within NetDL is further enhanced by CV UNO’s deep analysis and machine learning, which can ascertain when network provisioning or status changes have impacted business and essential applications. When a network modification affects an application’s performance, CV UNO would automatically determine which change affected which workload or application, enabling the network engineering and operations teams to swiftly resolve the problem.
- Host or Application Change Impact Analysis – Without deploying any host-based agents, CV UNO can also quickly guide the operator or engineer to the precise root cause of the issue, reducing the amount of time needed for cross-functional coordination on the part of the operations team and the amount of time it takes to resolve the issue when a host or virtualization issue is affecting application performance and the network is unaffected.
- Topology-Aware Determination – Through the integration of a comprehensive perspective including all infrastructure systems, virtualized machines, systems of record, and network flow and state data, CV UNO would be able to precisely identify the underlying cause of application performance problems without resorting to the typical finger-pointing that is associated with traditional methods.
“Our customers have been demanding a composite system that provides seamless observability across network domains, prevents human errors, rapidly identifies root cause issues, and aids network engineers and operators to troubleshoot application performance issues,” said Douglas Gourlay, Vice President and General Manager, Cloud Networking Software, Arista Networks. “Arista’s Universal Network Observability, built upon Arista EOS and CloudVision platforms, fulfill this critical client need.”
CV UNO Components and Availability
The components that make up CV UNO are as follows:
- CV UNO Sensor – Gathers, transforms, and organizes flow/SNMP data before sending it to NetDL from a variety of sources, including VMware vCenter, DANZ Monitoring Fabric, and outside network devices
- CV UNO – By utilizing Machine Intelligence-based Analysis on data stored in NetDL to infer topology-aware correlations across events, changes, and anomalies, enabled via a premium feature license, it would integrate with and improve CloudVision‘s operational and network telemetry capabilities, speeding up root cause analysis and problem resolution
- CV UNO Recorder Node (optional) – expands the capabilities of packet capture, query, and replay to enable forensic, incident response, and intrusion detection use cases
- CV UNO Service Node (optional) – Allows for more sophisticated packet processing features, such as DPI-based application identification and categorization and end-to-end application delay analysis
- CV UNO Analytics Node (optional) – makes machine learning and distributed context-aware traffic analysis possible for large-scale optimization
Arista‘s CV UNO has been in active client testing, and in Q2 2024 it will become generally available on CloudVision as a premium offering.
