Arista goes big with network observability platform

Arista Networks has taken the wraps off a network observability platform that promises to help enterprise customers proactively manage tasks including problem resolution and root cause analysis as it works to boost application and workload performance.

The company’s CloudVision Universal Network Observability (CV UNO) system is designed to handle a data lake’s worth of network telemetry and analytical data and meld it with artificial intelligence and machine learning technologies to offer real-time network flow and application performance details, risk and incident analysis, and change impact management.

CV UNO will be available in the second quarter as a licensed premium component of Arista’s CloudVision as-a-service platform. CloudVision is Arista’s core cloud-based management platform, which delivers a view of the entire network and offers network orchestration, provisioning, telemetry, automation and analytics across data center, campus and edge enterprises.

“In the intricate dance of troubleshooting network, workload, application, and user issues, the goal is clear but daunting: prevent changes that would cause outages and cut down the time it takes to prove the network isn’t the culprit,” wrote Douglas Gourlay, vice president and general manager at Arista, in a blog about the CV UNO platform. 

“That’s where CV UNO comes in, because we’ve built a system that addresses risk analysis and impact analysis, that shifts from reactive to proactive. So, hopefully, you find a problem before it’s created,” Gourlay said. “And then when you need to be able to defend yourself and say it is or is not the network’s fault, CV UNO will give you accurate and informed root cause analysis based on the data that shows exactly where a problem is.”

Once turned on, CV UNO can discover all the networked components in the enterprise, including applications, hosts, and workloads across various platforms, IT systems of record, and inventory management systems, according to Praful Bhaidasna, product management leader at Arista.

CV UNO includes a software-based sensor that collects, normalizes, and curates flow/SNMP data from various sources such as VMware vCenter, DANZ Monitoring Fabric, and third-party network devices as well as specific MAC and IP addresses.

All of this data is sent and stored in Arista’s Network Data Lake (NetDL), which is a component of CloudVision. NetDL includes other network and system data gathered over the years from a variety of Arista and third-party systems, which ultimately helps with historical analysis of problems and threats and can be used by CV UNO as well, Bhaidasna said.

From that data, CV UNO builds an application-to-network graph on CloudVision’s central monitor; the graph is continuously refreshed and stored in time series to show a record of the environment’s evolution and state at any point in time.

CV UNO’s ML and AI-based capabilities use this data to infer topology-aware correlations across events, changes, and anomalies, thereby accelerating root cause analysis and expediting issue resolution, Bhaidasna said.

“With CV UNO, customers can see both ends of a transaction, from the user to the VM, based on the application, the identity of the user, the network packet. [CV UNO] could then determine a response to a problem and accurately decide if it’s in the network segmentation or the application,” Bhaidasna said. “It is the keystone that pulls all of our system information and capabilities together to empower the network operator and network engineering team to be able to make very informed decisions about what to do next with problem resolution or troubleshooting application performance issues,” Bhaidasna said.

Another important capability of the system is gauging the impact of a network or host system change before that change is implemented, Bhaidasna said. For example, a customer could determine what the impact of shutting off a particular port on a router or switch would be on the downstream network and its applications before the change is implemented.

“It’s amazing to me that it’s taken 25 years for this industry to figure out that when somebody is making a destructive change to the network, they might want to know what that change is going to impact before it goes out. And it’s almost laughable, but it’s excessively true,” Gourlay said.

All eyes on network observability

Network observability tools are a hot topic in the industry right now, as vendors look to help organizations manage a growing, complicated mélange of distributed network and applications.

According to research firm IDC, networks and network teams are under tremendous pressure to keep pace with business demands and technology advancements. “All the while, expanding connectivity requirements, management toolsets, service dependencies, and security threats are complicating the design, deployment, operation, and evolution of the network infrastructure — both as a distinct IT domain and as a vital component of an entire digital infrastructure,” IDC stated in its report, “Top Trends in Network Observability for 2023.”

Network observability tools provide detailed intelligence and insights that not only heighten network integrity and innovation but also contribute to management efforts focused on other IT domains, such as security operations, cloud operations, and system reliability engineering, and broader IT automation efforts such as AIOps, IDC stated.

“Understanding the state of the network at any given time, delivering consistent and high-quality network services, maximizing the productivity of both users and IT staff, protecting networked resources, and adapting readily to new requirements become key tenets of digital business success,” IDC stated.