Juniper Apstra Demo: Root Cause Identification

0:01 [Music]

0:07 this demo will show how appstr uses

0:09 intent-based analytics and root cause

0:11 identification to reduce mean time to

0:14 resolution

0:15 juniper appstra leverages intent-based

0:17 analytics to manage data center fabrics

0:23 it manages a network fabric as a single

0:25 entity rather than a collection of

0:27 individual devices

0:33 in this view i'm showing the dashboard

0:35 of an operating data center fabric in a

0:37 remote location

0:40 earlier one of the devices in my network

0:42 was experiencing problems and a

0:44 technician replaced that device

0:48 this display reflects the status of the

0:50 network after he has the new device

0:53 you can see that i have bgp anomalies

0:56 some cabling anomalies

0:59 i've got route table anomalies

1:03 all the individual devices appear to be

1:05 operating in a healthy state

1:08 but in this node status view i've got

1:10 different size red dots that indicate

1:13 the number of anomalies associated with

1:15 each category for each device

1:19 for example border rack leaf 1 has 2 bgp

1:22 anomalies 1 cabling anomaly and nine

1:25 route table anomalies

1:27 let's go to the active network tab and

1:29 see the topology

1:33 from this view we can see all the

1:35 devices in the fabric and a color-coded

1:37 status for each one

1:41 on the right side i have the telemetry

1:43 coming from all devices in the fabric

1:45 this reflects the same information we

1:47 saw in the dashboard view

1:51 if i want to drill into this specific

1:53 leaf device and see more details

1:56 i can see there are two route anomalies

2:00 meaning that i have two routes missing

2:02 from the table

2:03 apstra knows from the intent model what

2:05 routes should be present and it compares

2:08 the current state to the intended state

2:10 this is what is meant by intent-based

2:12 networking

2:14 if i go into this spine device i can see

2:16 that i have a number of additional

2:18 anomalies cabling bgp and routes

2:22 a graphical map indicates something is

2:24 wrong with the cabling in this device

2:27 this is the device that was swapped

2:29 earlier by the technician in a remote

2:31 location

2:36 if we go to the root causes table we can

2:38 see that there are two root causes have

2:40 been identified based on the anomalies

2:42 that are present in the blueprint

2:50 i can see three devices associated with

2:53 root causes

2:54 spine two border rack leaf one and

2:57 border rack leaf two

2:58 the tabular view is going to give me

3:00 details about how these root causes were

3:02 identified

3:04 root cause identification is telling us

3:06 these two spine leaf links are down it

3:09 determined this because of the

3:10 relationship of the associated anomalies

3:13 to the intent-based model

3:18 traditionally the technician in the

3:20 remote location would have to go to the

3:22 location in person and examine the rack

3:24 and the device and the cables that he

3:26 installed

3:29 but appstore gives me a more efficient

3:31 way to resolve this problem

3:35 by going into the stage tab and into the

3:37 links table

3:40 i can pull the current lldp data from

3:42 the devices this is the same information

3:44 that was used by the root cause

3:46 identification process to identify the

3:48 problems

3:49 i can update this cabling map

3:52 from that data

3:56 and by updating i'm going to have the

3:58 system configure the ports for the

4:00 appropriate devices on the other end

4:05 the new configurations will show up in

4:07 the uncommitted tab for review

4:10 we can see that the tabs turn yellow

4:13 this is a pre-change analysis that

4:15 allows me to verify what's going to

4:17 happen before pushing the changes to the

4:19 production network

4:20 in this case these links are going to be

4:22 swapped all looks good to me so i'll

4:25 commit that

4:27 i will write a note

4:38 when i commit staged commands appstr

4:41 sends instructions to the agents and the

4:43 agents determine the appropriate command

4:45 line syntax to apply to the devices this

4:48 is what we mean when we say you specify

4:50 the what and appstra will take care of

4:53 the how

4:54 it only takes a moment for each agent to

4:56 detect the new state and the anomaly

4:58 indications are resolving

5:00 you can see how they are all turning

5:01 green

5:06 let's go back to the dashboard view and

5:08 you can see that all the anomalies are

5:10 cleared and the network is operating in

5:12 the intended state

5:14 [Music]