Vasyl - stock.adobe.com

Google Cloud introduces holistic cloud network monitoring

Spotting a gap in holistic network management in today’s multicloud business environments, Google Cloud introduces its Network Intelligence Centre

As experience becomes everything in a cloud-based world, Google Cloud has introduced a range of network monitoring, verification and optimisation tools for today’s multicloud business environments.

The Network Intelligence Centre is designed to alleviate problems that organisations embarking on digital transformation strategies are encountering as they are pressed to engage more effectively with customers, partners, suppliers and employees.

This has seen an uptake in adopting multicloud implementations and hybrid deployments to deliver the best experiences, but these are becoming increasingly complex due to the growing heterogeneity in the customer environment.

“If you are an enterprise looking to enable great consumer experiences, ultimately that requires a great infrastructure and network infrastructure,” Shailesh Shukla, vice-president of product management and networking at Google Cloud told Computer Weekly, explaining the rationale for the new product set.

“Ultimately, user experience is determined by the speed, latency, connectivity and availability uptime, and that requires a great infrastructure. If you look at how enterprises are starting to adopt public cloud, they gain advantages around speed, flexibility, agility and scale.

“But what has happened is that they have ended up not owning the infrastructure. It’s not in their datacentre in in the cloud, so the cloud is a black box. They lose visibility into their own deployment.

“Networks are also becoming more dynamic. Once you enter the cloud, every application has a different set of microservices which need to be connected, and those connections can be brought up or down based on the workload.

“So what happens is that users don’t have a view on system behaviour, and there is no holistic way of understanding connectivity and performance across the modern infrastructure in the cloud.”

To address these issues, especially as the new applications-based microservices need to talk to each other in a high-performance, low-latency dynamic way. Google Cloud believes that what is necessary is cloud-native monitoring and analysis technology, which has not existed to date.

In the first phase of fulfilling this vision for intelligent network operations, Google Cloud is offering Network Intelligence Centre with four core modules. The first two – Connectivity Tests and Network Topology – are both in beta, while Performance Dashboard and Firewall Metrics & Insights are in alpha stage, with several other modules to follow.

Network Topology lets users visualise and navigate their network with structure and insights, and is intended to simplify network monitoring and troubleshooting. It provides performance metrics for various nodes and edges, enabling metrics-driven troubleshooting, policy checks, and architecture and capacity optimisation.

It lets users visualise how those on their network are being served worldwide and discover whether they are being served optimally out of their nearest geographical regions. According to Google, users can search through networks and drill down into exactly what matters to them, therefore reducing troubleshooting time.

In the event of a network issue, users can go back in time up to six weeks to track the evolution of the network’s topology to diagnose what went wrong and when.

Explaining the fundamental relevance of the new solution to enterprise users, Shukla said that some of the key user benefits from the new monitoring solution were to accelerate troubleshooting and reduce technology overheads.

“Companies know how long they spend on troubleshooting and what the cost of an outage is – in eCommerce, this cost is massive,” he said.

“You cannot conduct any business. Companies need one place where they can look at cloud and look at on-premise and there’s no way [to date] to proactively test network configurations that would prevent an outage. Some 75% of network outages are caused by misconfiguration.

“What we can do is pre-test a given configuration in a known pattern to prevent this. Companies can now look at user patters to see if it could cause an outage before they push it out. You can now prevent outages before you have to deal with outages.

“You can ensure that the network is matching the intent of the designer. You can do this globally, which nobody else in the industry can. You can do this on a per project level, which others are not able to do.”

The solution has already been in beta testing with a number of firms, including mobile app analytics company Kochava and a real-time location data and analytics company Pivvot.

Talking about the experience of the Google system so far, Kochava enterprise architect Rob Lyon said: “Using Network Intelligence Centre, we discovered that data transferred to a particular GCE region was significantly higher than expected. By looking at Network Topology, we were able to diagnose and fix this issue and save significant costs.”

Jason Cradit, senior director of technology at Pivvot, added: “Network Intelligence Centre helps us to quickly identify and troubleshoot network problems. We’ve gained full situational awareness of our network topology, saving us time on monitoring network health.”

Read more on network monitoring

  • Many IT teams still use older monitoring tools that provide limited visibility across multiple network environments. AI and machine learning can help, but they'll take some time.
  • Datadog Network Performance Monitoring intends to ease the workload of engineering teams by automating network monitoring and providing visibility down to containers.
  • Multi-tenant public cloud monitoring can pose significant challenges. Luckily, some tools can help track these environments and even flag potential security threats.

Read more on Network monitoring and analysis