I want to show you how Marvis can help you with your wired issues.
What we’re going to do is we’re going to come to Marvis and we’re gonna go to Marvis Actions up here in the corner.
What I love about Marvis Actions is that it’s kind of the cup of coffee view.
It’s the first thing that you can check every morning if you want to get an idea for the general health of your network.
I’ll typically hit the SLEs in the morning and I’ll also look at Marvis to see if Marvis has found any problems on the network.
We use the power of the Cloud, AI and Machine Learning to automatically identify issues both on the wireless and on the wired side of the network.
And notice that we have an entire category here that’s dedicated to switches.
Let’s go take a look.
If we click on that…
And by the way, this is going to show us org wide.
So if you have 300 sites within your org, this will show you everything.
So if we look at our switches here, we can see that there’s a missing VLAN, there’s a bad cable, and there’s a negotiation mismatch here as well.
So missing VLANs are pretty cool.
You actually don’t need a Juniper EX switch to get these.
In fact, this is a Cisco switch that we’ve simply plugged a couple of Mist access points into, and the Mist APs are able to detect missing VLANs.
This is like one of the oldest problems in the book that we finally resolved with machine learning.
And so what you can see here is that we found this AP and this AP.
They’re missing the same three VLANs, 100, 200, 300 on port 21 and on port 13 on this specific switch.
What we do is we categorize a VLAN in one of four ways.
We categorize it as: a high traffic VLAN, a low traffic VLAN, a black hole VLAN (That’s a VLAN that we just throw unwanted traffic into so it doesn’t go anywhere, you know, maybe to do kind of a walled garden thing or something like that)
or we categorize it as a truly missing VLAN.
So we use machine learning to categorize it in one of those four ways.
The reason that we’re able to do this is because the MIST cloud is able to see how much traffic and what kinds of traffic are moving on all of our different APs network wide.
We have a holistic view over the whole network.
And then we can apply machine learning to categorize those VLANs in one of those ways.
We have a customer who is a large retailer in North America.
And they have, last time, I checked there was somewhere between 60,000 and 80,000 APs on their network.
Let’s say 70,000 APs.
There are about 70,000 APs on their network across 300 sites with non-Juniper switching.
By the way, they’re getting Juniper switching now.
We turned Marvis loose on their network to find problems.
And Marvis found 20 switches with missing VLANs.
And the accuracy was 100%.
That is the power of AI.
That is the power of machine learning.
And that’s the power of having infinite processing power and data collection capabilities in the cloud.
It helped us solve the oldest problem in the book with 100% accuracy.
We can also do bad cable detection as well.
I love this.
At my house, I made some itty bitty little patch cables the other day to neaten up my switch rack.
And a couple hours later, Marvis flagged one of the cables as bad.
And I’ve left it like that because it’s great for a demo.
But it looks like Marvis has found a bad cable on Port 10 on this switch.
And so we need to let Marvis know.
We can say, ‘Hey, we’re working on fixing this.’
And then we can mark it as resolved and the Marvis will validate that the problem has been resolved.
We use a multifold approach to do this.
We do things like, we look to see: ‘Is there a PoE being drawn on that port but no data’ Or is there a ton of transmit and no receive or vice versa’ Or are there a bunch of CRC errors on that port”
Marvis will build a case as to why it thinks that cable is bad and then it will tell you about it.
This is awesome for finding intermittent problems.
I don’t know about you but I’ve spent a lot of time running around with a link runner, running around with the toner, running around with a link sprinter, trying to find the bad cable somewhere.
Marvis does this stuff for you automatically.
It’s pretty cool.
Marvis will also find things like negotiation mismatches.
And so you can see there’s a negotiation incomplete on Port 30 here.
Lots of other things that Marvis can do.
Like, Marvis can find if it sees significantly more EAP 802.1x failures than normal, if you’d have like a radius problem or something like that, it’ll create a baseline how often do things normally fail,
and then it will alert you if you significantly deviate from that baseline.
It’ll also help you understand if there’s DHCP or DNS failures as well.
So right now, Marvis tells you about problems so that you can fix them.
And then Marvis validates that you fix the problem.
But where we’re going with Marvis is, ultimately, we think that Marvis can fix these problems for you automatically.
As we get more customers onto the MIST cloud, as we get more data for Marvis to ingest, we’re going to get closer and closer to achieving that goal.
So that’s a quick demonstration of Wired Assurance with the Juniper MIST cloud.
Thanks a ton for checking out this demo.
Hope you have a great rest of your day.
We’ll see you soon in a future video.
- Getting Started
- Wireless
- Wired Switching
- WAN Edge
- Mist Access Assurance
- Location Based Services
- Premium Analytics
- Security and Cloud Administration
- MSP
- Automation
- Product Updates
- Marvis
- Security Alerts
- FAQ