Demo: Increase Visibility and Control for HPC Clusters with Rockport ANM

With Rockport, your network is never just an opaque black box. The Rockport Autonomous Network Manager (ANM) provides a bird’s eye view of your active network and deep insights into network performance. In this video, Rockport Product Manager Sharon Chisholm shows how easy it is to use ANM to set up your network, monitor its health, troubleshoot issues, and gain real-time insights into the performance of your workload.

 

 

Video transcript:

Hi, I’m Sharon, a senior product manager here at Rockport Networks. At Rockport, we provide a switchless network, complete with market-leading tooling that simplifies the deployment and management of performance networks. We know many of you are tired of flying blind, unsure of what is happening in your network. With Rockport, your network is never just an opaque black box. The Rockport Autonomous Network Manager, or ANM, provides a bird’s eye view of your active network and deep insights into network performance as well as innovative time travel.

Today, I’m going to take you on a quick tour of ANM showing you how easy it is to use an ANM to set up your network, monitor its health and troubleshoot issues. And I’ll show you some of the advanced features that provide real-time insights into the performance of your workload, making your life easier.

 

Simple Deployment

The first thing I want to show you is how simple it is to deploy a Rockport network – whether large or small, deploying your new network is easy. Simply install the Rockport network card in the server, connect the cables from the network card to the shuffle, and ANM takes care of securely updating the software, then configures and manages each of the Rockport network cards.

Within a few minutes of powering on the server that hosts the Rockport network card, a secure bootstrapping process takes place. During this process, ANM provides a TLS  certificate to the newly enrolled cards. The cards will now only respond to management traffic secured with that certificate

Health

Once the new cards are enrolled, the entire Rockport Switchless Network topology and route information is automatically updated. The ANM health screen shows a view of all the cards in your network as nodes in a connected graph. Not only are the nodes and their ports color-coded based on health, the least healthy nodes are centered and enlarged to make it easier to identify potential problems in your network.

Troubleshooting

Next, to demonstrate the tooling available to help troubleshoot a problem node, we shall look at the smaller system, where I’ve made some adjustments to trigger some error conditions. If you click through, you can get some general information about the node, its ports, links on the ports, including traffic running over those links. Or you can get a list of currently active or recently active alarms. With this high optical transmitter temperature alarm, we can see we have the usual information you would expect from an alarm. But as this is a threshold crossing alert, we also get the measured values of that metric over time, as well as the threshold values that are used to raise and clear this alarm, which I artificially lowered for this demo.

Metrics and Network Transparency

ANM provides industry-leading insights into the performance of your network. We gather and make available dozens of metrics as well as tooling to help you understand them, including traffic rate, traffic profiles, packet counts (both application and network) packet size, optical port details, and general system health.

In addition to metrics of a specific card or port, you can see network-wide views, allowing you to recognize trends and outliers. Range shows a characterization of application traffic through the network. Rate shows the raw application traffic rates for each of the nodes in the network. Back pressure, the characterization of pause events, and optical shows CRC errors other metrics of interest.

In flow diagnostics, you can get details on the flow of data between two nodes, a summary of the traffic being sent, a list of the currently active paths, details of those paths, as well as details on the specific links along those paths.

Time Travel

And finally, I wanted to show you how ANM even allows you to travel back through time, whether it be just a few minutes, hours, days or even weeks. It allows you to understand what was happening within the network at a particular time that a job was running,  a problem was experienced or even just to understand historical trends. I can even change the granularity of the data that I’m looking at to gain additional insights.

 

That’s a quick overview of how you can use ANM to simply manage your high performance Rockport Switchless Network. To find out more about Rockport or ANM, please visit www.rockportnetworks.com and stay tuned for more videos in this series.

Here’s something similar we think you’ll like.