Juniper Apstra Demo: Proactively address network issues with intent based analytics

0:01 hard persistent failures aren't

0:03 difficult it's Shades of Gray that are

0:05 tough to find

0:06 the problem is I'm finding a needle in a

0:08 haystack it's finding a needle and a

0:10 stack of other needles

0:12 once you get the network up and running

0:14 you need to keep it running this is an

0:16 exercise in monitoring and maintenance

0:19 Tech Optics is an example in a typical

0:21 data center you're swimming in an ocean

0:23 of sfps finding the failed ones is easy

0:26 but finding the degraded ones how do you

0:28 diagnose something that is kind of

0:30 working

0:31 with appstra you can move from reacting

0:34 to a failure to acting based on a

0:36 precondition of a failure this is hugely

0:38 impactful

0:40 in this demo we're running a

0:41 multi-vendor fabric with a couple of

0:43 juniper switches and a couple of Arista

0:45 switches

0:46 apps includes a wide selection of

0:48 predefined IBA or intent-based analytics

0:50 probes

0:51 and you can create your own custom

0:53 probes

0:55 in this demo we're going to instantiate

0:57 one of the predefined IBA probes that

0:59 monitors Optical transceivers

1:04 let's look at the statistics that we're

1:06 able to collect from this probe

1:08 we can see we have Optics from different

1:10 vendors and a wealth of data from those

1:12 Optics

1:17 it's possible to filter and sort based

1:19 on that vendor field or any other

1:21 captured value

1:26 we can also drill down to view stats

1:28 within individual Lanes on multi-lane

1:30 Optics

1:34 we can see that the probe has already

1:36 raised anomalies for some of the

1:38 interfaces

1:45 we can drill down into the probes and

1:46 see exactly what's going on in this case

1:49 we're seeing a receive power threshold

1:51 warning

1:54 we can also view historical data for

1:56 each of these probes over time

1:58 in this example we're looking for

2:00 interfaces that have experienced an RX

2:02 power low warning condition in the last

2:04 hour

2:13 we can also create custom dashboards

2:15 based on these probes

2:17 so if there's a particular condition

2:18 that's important to your business it's

2:20 very easy to build dashboards that are

2:22 meaningful to you

2:23 in this example we have created a

2:25 dashboard based on this same optic probe

2:27 and can easily view any anomalies that

2:29 this probe may have uncovered

2:31 this type of insight is very powerful

2:33 when operating a network

2:35 it will help with cost avoidance by

2:37 reducing the effort needed to identify

2:38 and resolve problems it will improve

2:40 your network reliability by allowing you

2:43 to identify problems before they become

2:45 outages and it will undoubtedly lead to

2:47 happier customers by helping you

2:49 proactively address problems before they

2:51 become catastrophic failures in fact

2:53 Yahoo Japan experienced a scenario very

2:55 much like this

2:57 they were able to use these abstra IBA

2:59 probes and custom dashboards to provide

3:01 a simple way to identify these problem

3:03 Optics when the degradation began and

3:05 act on them before they caused outages

3:08 with abstra you can identify problems

3:10 early before they cause major outages

3:12 you can find that needle in the stack of

3:15 needles

3:16 [Music]

3:17 foreign

3:19 [Music]