- User experience monitoring
- Real end-to-end data
- Zero touch deployment
- Dynamic and scalable
Once upon a time, network troubleshooting used to be easy. Just assume that the user is the problem, tell him/her to turn the client off and on again and in 90% of the case, that worked wonders. Well, those times are over. Microsoft manages to build software that works (mostly without 24h reboots), the network has become truly mission critical and support people that just blame the error on the user are no longer fashionable.
Are your users sometimes complaining about network performance and you can’t give them a verified response because you’re missing a holistic end-to-end network performance monitoring? Is the problem on the user end side? Is is the network link or the switching infrastructure? Pinpointing the root cause isn't easy and even thought there are many tools out there, they either focus on the network part or on the user side. Now, what if you could run the same test scenarios from within the network as well as from the user's client? This even without installing anything on the user's computer! And what if you could aggregate all this information on one simple screen and get a quick overview of what's going on? Does this sounds like a fairytale?
... we just built a simple, yet very powerful solution for this problem! UXmesh is an end-to-end user experience monitoring tool, that brings real-life monitoring to a whole new level. Get data from the client's browser, network components and even further! So, don't rely on your client's subjective opinion. Give him a tool to deliver live, unbiased data you can work with.
With UXmesh you can easily test the end-to-end network quality throughout your network up into any user’s browser. Thanks to the new Docker support on the Cisco Catalyst 9000 Switches your network devices can run the same code that is executed in the user's browser! Ohh my... you see where this is going. All intermediate network nodes, combined with the users in one single mesh where every individual connection can be seen.... this gives you a hell of an overview and helps you quickly locate the root cause.
Try it out and use the UXmesh Docker image to turn your Cisco Catalyst 9000 Switches into performance monitoring nodes across your whole network!
Cisco Catalyst 9000 Switches with Docker support! Thanks Cisco!
When a client runs the UXmesh code, it calls home to the UXmesh server and joins the waiting room. As soon as a second client joins, a webRTC channel is established directly between the two clients. The quality of this channel is now evaluated by the code running on either client and the data is sent back to the UXmesh server.
As soon as a third client joins, this client establishes channels to either already active client, and so forth.
The new capability of the Cisco Catalyst 9000 Switches to host and run Docker images now allows to not only monitor between client browsers and servers, but to monitor each network segment by itself! Browser to switch, switch to switch, switch to server! And if necessary, that can be done in full mesh!
Docker image has been installed and "running" on the Catalyst 9000 instance in the "IOS XE on Catalyst 9000 16.12 EFT Code" Sandbox.
Caused by the sandbox restrictions that disallow internet and interdevice connectivity and the need of the project to be deployed on at least three switches to make sense, the next phase would be to validate the project on physical instances of Cisco switches.
Yay, Proactive Monitoring! Use the application recognition capabilities of Cisco NBAR2 to identify webRTC traffic and its way through the network to proactively monitor this route! Where and when needed! Scaleable, dynamic and end-to-end!
- etablishes a connection to the server, to retrieve a list of all clients in the network
- connects to each client in the network directly with webRTC
- can be executed as a Docker container on a device or as client-side-code in the browser
- aggregate data received from clients
- holds a list of all clients, which are connected to each other
- returns the list of all clients to each conneting client
- provides a http endpoint and delivers the graphical UI to see fancy animations of the collected network data
to see a showcase of an infrastructure with 4 uxmesh-clients and 1 uxmesh-server, visit https://uxmesh.k8s.zebbra.ch
if you want to setup the szenario in your own network, then please continue reading...
prerequisites: --> install node in version 10.x or higher
brew install node
npm install
with debug output:
DEBUG=* npm run server
regular:
npm run server
docker run quay.io/zebbra/uxmesh:latest npm run server
or
docker build -t uxmesh .
docker run uxmesh npm run server
type in your browser: https://uxmeshserver.your-domain.com/?connectionUrl=https://uxmeshserver.your-domain.com
apart from the graphical interface your browser operates as a client as well and starts to send peer information accross the network to all registered peers.
Prerequisites: --> install node in version 10.x or higher
brew install node
npm install
With debug output:
DEBUG=* npm run client
Regular:
npm run client
Use optional parameter to set uxmesh-server address (default is http://localhost:3001):
npm run client https://uxmeshserver.your-domain.com
type in your browser: https://uxmeshserver.your-domain.com/?connectionUrl=https://uxmeshserver.your-domain.com
connectionUrl defaults to http://localhost:3001
apart from the graphical interface your browser operates as a client as well and starts to send peer information accross the network to all registered peers.
docker run quay.io/zebbra/uxmesh:latest npm run client
or
docker build -t uxmesh .
docker run uxmesh npm run client
- chrome://webrtc-internals/
- https://github.com/feross/simple-peer (see perf dir for simple perfomance test example)
- https://d3js.org/
- https://gist.github.com/d9bf021c395835427aa0.git
- https://blogs.cisco.com/developer/code-for-catalyst-challenge-is-now-open