DJ Spry, Director of Product Management, Apstra, Juniper Networks

The Automated Data Center in Action

Demo Drop Data CenterNetwork Automation
DJ Spry Headshot
Title slide has a green box with the title of the demo, “Live Demo: The Automated Data Center in Action,” written in white text on top The Juniper and Apstra logos are at the bottom of the green box. The presenter’s name, DJ Spry, is also included.

See first-hand how quickly and reliably the automated data center works.

This live demo with Juniper’s DJ Spry takes you step-by-step through key concepts of the Automated Data Center. By the end, you’ll see how automation can drastically simplify the user experience and increase business uptime.  Sign up for a full length demo.

Show more

You’ll learn

  • How the automated data centers enables new business and network services

  • How to recover data center fabric to a previous known good state 

  • How real-time and historical analytics and provide better network insights 

Who is this for?

Network Professionals Business Leaders

Host

DJ Spry Headshot
DJ Spry
Director of Product Management, Apstra, Juniper Networks
Transcript

0:06 hi i'm dj spry from juniper networks and i work on abstra intent based networking software and today i'd like to go over a

0:13 live demo of the automated data center in action with you a quick agenda what we'd like to do is

0:19 to show you how you can quickly and reliably enable new business and networking services and how you can

0:24 quickly and reliably you know recover data center fabrics to a new previously known good state also we're going to

0:30 cover how you can identify network service level objective issues and reduce your mean time to resolution your

0:35 mean time to innocence as well as have better network insights with real time and historical analytics

0:41 all of which with a drastically simplified user experience to increase your business up time with that i'd like

0:47 to hop directly into the demo okay i'm logged into appstr and you can see

0:53 that there is a data center in austin texas being managed by appstr so the first thing i want to do is

0:59 create a new tenant or security zone or vrf in order to place my new application in or my new service

1:05 so i'm going to call this prod and i'm going to hit create it's as simple as that there's additional

1:10 options that you can choose from but you can do it as easy as in two clicks after

1:15 at this point in time is going to provide some checks and say hey in order to establish this new security zone we

1:21 need to establish you need to give some additional information in this case it's uh some

1:27 loopback ip addresses for the network resources right the next thing we need to do is

1:33 create a new virtual network that we can place this application in in this case this application is an amazing new app

1:40 that we want to put into production we want this quickly and reliably into service we expect a lot of demand for this and

1:46 everyone's really excited and you can specify addresses manually like i'm doing here

1:52 or you could allow appstra to auto assign these resources through a resource pool

1:58 that is available as an internal ipam capability i'm going to enable these as vlan tagged

2:04 interfaces across every interface in the data center again because we believe the demand for this application is going to

2:09 justify that so i've just provisioned this new virtual network everywhere in the entire data center in just a few

2:15 short clicks and if i wanted to look at the configuration

2:21 that appstr automatically renders i can do that here and scroll all the way down as an example and i say hey there's

2:27 being there's a new vnvn3 and if i look back

2:33 the virtual network for my amazing new app i can see the vn3 on leaf three

2:41 is for my amazing new app so again it's a very high level declaration interface that's presented to the user

2:48 and abstra automatically through its data model renders all of the configuration necessary for the

2:54 underlying network devices and operating systems so looking at the uncommitted tab here is where you review what you're

3:00 about ready to or what you're allowed to push into production in this case there's no errors

3:06 and i'm reviewing that yes in fact i do want to provision prod in my amazing new

3:11 app and i can get some additional information before i push this in but this is exactly what i want to deploy and i'm going to go

3:17 ahead and commit this to deploy a new app

3:26 what happens at this stage is appstore also takes a snapshot or a copy of the

3:31 entire network state and of the graph database and saves it every time on every successful commit and i can use intent

3:39 time voyager to look back at all of the past commits successful commits that's happened inside the network and jump to

3:45 any of these revisions so here unfortunately what happened is that the

3:50 application team found an issue and marketing team wanted to delay the launch of our amazing new app so what i

3:56 want to do is go ahead and jump back to this revision and again this could be the same people deployed it or it could

4:02 be completely different people have no idea of the context of what exact configuration was made

4:09 and or what the individual command line elements for every individual device they don't need to know that you can

4:15 jump back and restore state across the entire data center fabric and in one click so this you can go from broken to

4:22 unbroken or roll back you know application deployments in single clicks you get the same validation engines

4:29 the from appstra with intent time voyager as you do with initial config

4:35 initial configuration deployments through staged so here i can see that i want to remove prod or remove my amazing

4:42 new application i'm going to go ahead and commit this and say removing

4:47 amazing app vn and security zone

4:53 so this is dj

4:58 all right and just like that everything is removed configuration is being deployed you can

5:04 see that there is nothing pending if i go to my staged review there is no more amazing new app

5:10 and if i were to look the physical configuration interface of leaf three that we just looked at

5:17 you can see that bn3 is now removed all right so i've logged into another

5:22 instance of app sure managing another data center in austin texas but very quickly i can determine that there are

5:28 issues affecting the service in this data center so i'm going to go ahead and click into it

5:33 you have a dashboard right off the bat that shows you the health of the data center across many different facets from

5:39 both configuration deviations interfaces cabling bgp and route

5:44 so slow responses to service disruptions or end user experience issues can result in a loss of customers corporate

5:51 reputation and certainly revenue the abstraint based design in closed loop analytics builds a feedback loop

5:58 between what the architects who built the system and the operators who are responsible

6:04 for ensuring how that service or is that service working in production and because appstra is used from both day

6:12 zero all the way to day three the documentation is always up to date and the closed loop or live query capability

6:18 allows real-time notification for any deviations in this case you can see that there's a

6:24 configuration deviation issues with interface cabling in bgp scrolling on down there is this node

6:30 status and i like to refer to the blast radius to see what devices are most impacted and how they're impacted i'm

6:37 going to focus on leaf three and it's a bit like google maps is the further you drill down the additional

6:42 bit of information you receive so here it's not just validation of

6:48 configuration but of networking states and networking elements and because the policy

6:54 is declared through the user through appstra we understand the network behavior and continually audit

7:01 the network to see if it's in accordance and behaving as it should be in this case even though that there is a route

7:07 to 100 32 it's not the route that is expected to be there from the service

7:12 that was instantiated so here we are receiving a route with the next hop is

7:17 172.16.0.10 where it should be four so there is a deviation here another bit of information is that we

7:24 append or astra appends data that is useful to the operators and

7:29 here it's just not a route but it's a route from spine2 as an example so again this has presented a lot of

7:36 contextualized data from bgp ldp and interface but focusing on

7:42 configuration as abstra is now the single source of provisioning

7:48 uh any deviation that happens to the network is immediately raised and flagged to the user so here you can see

7:54 that there's been an interface xc00 0 that has been disabled on this particular juniper device so this could

8:01 be something innocuous or it could be malicious but the point is is that you always have confidence in 100

8:07 accountability to of what is running in your network at all times

8:14 last thing i want to show is how operations teams can leverage apps identify and isolate

8:20 issues in the routing protocols themselves that may be impacting network availability or service level objectives

8:26 so it's impossible for any person to mentally keep track of all the evpn route types

8:32 and the control protocol behavior for every network provisioned in the data center but after rather than trusting the

8:38 network in the protocol behavior is codified into an iba probe or an intent-based analytics probe that is

8:45 constantly validating the health and accuracy of the ebv evpn control plane rib and flooding

8:51 domains on every switch and raises anomalies for any deviation and here is a deviation

8:58 the abstract has identified for the evpn vxlan type 5 validation

9:03 rather than looking at this individual probe what i'd rather show is that you can create dashboards of probe and

9:09 analytic data most important to you in your business as an example this dashboard is showing east to west

9:15 traffic server traffic traffic exiting the data center in the last 12 hours the relationship in the

9:22 graph database along with iba

9:37 hi i'm dj spry from juniper networks and i work on abstract intent based networking software and today i'd like

9:43 to go over a live demo of the automated data center in action with you a quick agenda what we'd like to do is

9:49 to show you how you can quickly and reliably enable new business and networking services and how you can

9:55 quickly and reliably you know recover data center fabrics to a new previously known good state also we're going to

10:00 cover how you can identify network service level objective issues and reduce your mean time to resolution your

10:06 mean time to innocence as well as have better network insights with real time and historical analytics

10:12 all of which with the drastically simplified user experience to increase your business up time with that i'd like

10:17 to hop directly into the demo okay i'm logged in to appstre and you can see

10:24 that there is a data center in austin texas being managed by appstr so the first thing i want to do is

10:29 create a new tenant or security zone or brf in order to place my new application in or my new service

10:36 so i'm going to call this prod and i'm going to hit create it's as simple as that there's additional

10:41 options that you can choose from but you can do it as easy as in two clicks after at this point in time is going to

10:47 provide some checks and say hey in order to establish this new security zone we need to establish uh you need to give

10:53 some additional information in this case it's uh some loopback ip addresses for the network resources

11:02 right the next thing we need to do is create a new virtual network that we can place this application in in this case

11:08 this application is an amazing new app that we want to put into production we want this quickly and reliably into service we

11:15 expect a lot of demand for this and everyone's really excited and you can specify

11:21 addresses manually like i'm doing here or you could allow appstra to auto assign these resources

11:27 through a resource pool that is available as an internal ipam capability

11:33 i'm going to enable these as vlan tagged interfaces across every interface in the data center again because we believe the

11:38 demand for this application is going to justify that so i just provisioned this new virtual network everywhere in the

11:44 entire data center in just a few short clicks and if i wanted to look at the

11:49 configuration that apps automatically renders i can do that here and scroll all the way down as

11:55 an example and i say hey there's being there's a new vnvm3

12:01 and if i look back the virtual network for my amazing new app

12:08 i can see the vn3 on leaf three is for my amazing new app so again it's

12:14 a very high level declaration interface that's presented to the user and abstra automatically

12:21 through its data model renders all of the configuration necessary for the underlying network devices and operating

12:26 systems so looking at the uncommitted tab here is where you review what you're about ready to or what you're allowed to

12:32 push into production in this case there's no errors and i'm reviewing that yes in fact i do

12:39 want to provision prod in my amazing new app and i can get some additional information before

12:44 i push this in but this is exactly what i want to deploy and i'm going to go ahead and commit this to deploy a new

12:49 app

12:56 what happens at this stage is appsure also takes in a snapshot or a copy of the entire network state and of the

13:03 graph database and saves it every time on every successful commit and i can use

13:09 intent time voyager to look back at all of the past commits successful commits that's happened inside the network and

13:15 jump to any of these revisions so here unfortunately what happened is that the

13:21 application team found an issue and marketing team wanted to delay the launch of our amazing new app so what i

13:27 want to do is go ahead and jump back to this revision and again this could be the same people have deployed it or it

13:32 could be completely different people have no idea of the context of what exact configuration was made

13:39 and or what the individual command line elements for every individual device they don't need to know that you can

13:45 jump back and restore state across the entire data center fabric and in one click so this you can go from broken to

13:52 unbroken or roll back you know application deployments and single clicks you get the same validation

13:59 engines the from abstra with intent time voyager as you do with initial conf

14:05 initial configuration deployments through staged so here i can see that i want to remove prod or remove my amazing

14:12 new application i'm going to go ahead and commit this and say removing

14:17 amazing app vn and security zone

14:23 so this is dj

14:28 all right and just like that everything is removed configuration is being deployed you can

14:34 see that there is nothing pending if i go to my staged review there is no more amazing new app

14:41 and if i were to look the physical configuration interface of leaf three that we just looked at

14:47 you can see that bn3 is now removed all right so i've logged into another

14:53 instance of appsure and managing another data center in austin texas but very quickly i can determine that there are

14:58 issues affecting the service in this data center so i'm going to go ahead and click into it

15:04 you have a dashboard right off the bat that shows you the health of the data center across many different facets from

15:10 both configuration deviations interfaces cabling bgp and route so slow responses to service disruptions

15:17 or end user experience issues can result in a loss of customers corporate reputation and certainly revenue

15:24 the abstract based design in closed loop analytics builds a feedback loop between

15:29 what the architects who built the system and the operators who are responsible

15:35 for ensuring how that service or is that service working in production and because appstra is used from both day

15:42 zero all the way to day three the documentation is always up to date and the closed loop or live query capability

15:49 allows uh real-time notification for any deviations in this case you can see that there's a

15:54 configuration deviation issues with interface cabling in bgp scrolling on down there is this node

16:00 status and i would like to refer to the blast radius to see what devices are most impacted

16:06 and how they're impacted i'm going to focus on leaf three and it's a bit like google maps is the

16:11 further you drill down the additional bit of information you receive

16:16 so here it's not just validation of configuration but of networking states and networking elements

16:22 and because the policy is declared through the user through appstra we understand the network

16:29 behavior and continually audit the network to see if it's in accordance and behaving as it should be in this case

16:36 even though that there is a route to 100 32 it's not the route that is expected

16:41 to be there from the service that was instantiated so here we are receiving a route with the next hop is

16:48 to 172.16.0.10 where it should be 1.4 so there's a deviation here another bit of information is that we append or astra

16:55 appends data that is useful to the operators and here it's just not a route but it's a

17:01 route from spine2 as an example so again this is presented a lot of contextualized data from bgp ldp and

17:09 interface but focusing on configuration as

17:15 appstra is now the single source of provisioning any deviation that happens to the

17:20 network is immediately raised and flagged to the user so here you can see that there's been an interface

17:27 xc000 that has been disabled on this particular juniper device so this could be something innocuous or it could be

17:33 malicious but the point is is that you always have confidence and 100 percent accountability

17:39 to of what is running in your network at all times

17:44 last thing i want to show is how operations teams can leverage apps identify and isolate

17:50 issues in the routing protocols themselves that may be impacting network availability or service level objectives

17:57 so it's impossible for any person to mentally keep track of all the ebp and route types

18:02 and the control protocol behavior for every network provisioned in the data center but after rather than trusting the

18:09 network and the protocol behavior is codified into an iba probe or an intent based analytics probe that is constantly

18:15 validating the health and accuracy of the ebv evpn control plane rib and flooding

18:22 domains on every switch and raises anomalies for any deviation and here is a deviation

18:28 the abstract has identified for the evpn vxlan type 5 validation

18:33 rather than looking at this individual probe what i'd rather show is that you can create dashboards of probe and

18:40 analytic data most important to you in your business as an example this dashboard is showing east of west

18:45 traffic server traffic traffic exiting the data center in the last 12 hours

18:52 the relationship in the graph database along with iba presents you with this rich contextual data that is always

18:57 running reducing the number of tools and interfaces you have jumped here's that evpn type 5 anomaly raised

19:03 by abstra that is impacting the 11.0.0.24 network while others remain

19:08 unimpacted so here if i go and look i can see the 1100 24 is impacting my red

19:14 database this is analysis you can't get from traditional network monitoring tools or existing products and appstore

19:19 does this dynamically for every virtual network provisioned i have purposely deactivated the routing

19:25 and bridging interface for virtual network 2 or red database 01 on the juniper device leap 3

19:32 which is part of an evpn vxlan fabric here in order to force this loss of the route

19:37 type 5 and to trigger this anomaly so now that we see how quickly you can

19:43 locate and isolate issues let's see how simply you can restore the entire network as a single system back to

19:49 compliance and restore the network services to the database all you have to do to put the three back

19:54 into compliance is navigate to the active tab go to telemetry and find the configuration deviation

20:02 and here you see this apply full config apply full config puts on confirm

20:08 will take the golden configuration and render it back to the device so

20:13 so here is the configuration that is inactive and if i run this command again you can see already that abstra has put the

20:21 configuration the golden configuration back on and that everything is okay

20:26 going back to the dashboard you can see that anomalies are clearing

20:32 red things are going to turn green from evpn routes to cabling bgp and anomalies but you can already see the

20:38 configuration deviation is gone as well as the cabling so this portion of the demo shows how

20:44 anyone can quickly and reliably identify service level objective issues and reduce the mean time to resolution and

20:51 mean time to innocence using closed loop intent based analytics and now for the final part of the demo for today

20:58 so here i'm logging in to a new instance of a abstra

21:03 which is managing a data center in switzerland and if you navigate to the active tab be the full topology there is

21:09 a layer option that overlays real-time data usage and interface statistics for you directly onto the topology

21:16 so this is real-time information and the more that you click on the further you drill down very similar to the previous

21:22 it's like google maps the more detail you get so here i've navigated down to this particular device and if i hover

21:29 over i can get real-time statistic information and optic information uh how much data

21:35 so there's actually one step better where you can click on headroom and you can choose any endpoint in

21:41 uh in the network and look at the live traffic utilization across all points of

21:46 the network for which that data which reverse so here i'm going to choose rack 2 and you can see that the data traffic is

21:52 a bit different additionally there is a dvr-like capability this time series database

21:58 which allows you to use go back in time and view the data as a traverse across

22:03 those same endpoints in real time so here i can see what data happened to

22:08 be traversing across this link i can move it around i'm going to move it to this day so this is really useful for

22:15 forensic or analysis if you want to go back in time and say for instance the database team said that we moved a you

22:23 know we did a database replication but it didn't seem like the data was traversing the network or we had an issue so here you can you have the

22:29 ability to go back in time choose the aggregation of how you want and prove whether or not that the data actually

22:35 traversed in the network at that moment in time and again you can look at this from a topology standpoint uh across any point

22:43 in the network and any device okay well that concludes the demos today

22:50 so to recap what we've done is uh shown how you can quickly and really reliably enable new businesses and

22:56 networking services recover and enable data center fabrics to previously known good states and to identify issues in

23:03 the network identify those service level objective issues and reduce your mean time to innocence to gain a better network insights with

23:10 real-time and historical dvr-like capabilities of analytics that dramatically you know simplifies the

23:17 user experience and increases up time appstra is an intent intent-based networking solution that provides a

23:23 unified consistent and seamless experience from the architects day zero to the operators on day two plus

23:30 uh and it does so in a closed loop in an automated fashion

23:36 to re to really restate is appstore delivers on business outcomes so you can see here that there are some pretty um

23:43 shocking numbers the 99 and improvements and agility and infrastructure agility

23:48 but really when you look at what you can do with abstract in seconds and minutes as composed as compared to weeks and

23:55 months you can really quickly understand or easily understand where those 99 numbers come from

24:01 if you're interested you would like to learn this or try it out yourself we have a few links on the screen here

24:07 there's appstore academy youtube playlist and even a cloud-based lab environment sandbox at juniperbeadlabs

24:13 where you can try this out for free and with that i'd like to open it up for some q a

24:32 all right hello all uh thank you for joining me live sorry about a little glitch there

24:39 at the beginning uh i just felt that uh there was so much demand to see the first demo we decided we'd rerun that

24:44 again but we are live so anyone has any q a feel free to ask them now

25:05 all right does anyone have any questions here all right so i've got one says uh

25:12 can i incorporate the appstore solution into a larger network orchestration um summarize here apis etc

25:19 so yes the new game through the api the same validations

25:25 design principles and high-level declaration through the apis who did as you do the ui

25:30 appstore is even the ui is 100 a client of our api system

25:36 so for those who are looking at integrating or using this technology and to a broader

25:43 um orchestration environment inside their inside their business it is possible many of our customers do it um

25:50 and whatever that tool of choice for your for your api orchestration is be that hands-full or terraform etc all of

25:56 those sort of capabilities are available uh to you the user

26:02 um scroll through here if you have any again if anyone has any questions feel free to pop them across the channel

26:10 um scrolling through it says um i have a question is what methods can

26:16 app store use to collect telemetry from device so a good question abstra has an agent

26:23 model we have both an on box and off box agent and with this agent comprised of a couple different packages

26:30 for which to get uh telemetry on and off the box in case of an on box we are

26:35 using what's locally available but always be a programmatic interface where we can and a lower level something like

26:41 cli where we must but uh there's numerous

26:46 you know ways that we can collect numerous bits of data and telemetry off of the

26:52 device if you're curious please reach out to us and contact your se and we can go into

26:57 more detail and more depth there but the way that we collect that is through an agent

27:02 uh and in the off box say something like rest comp and unbox locally to device be a programmatic interface

27:09 so thank you for that question let's see

27:16 oh look and see anyone have any other questions see i've got one sorry one just popped

27:21 in here that says um what benefit what benefits does the

27:26 single source of truth provide for me in the operations team so it's a valid question what it allows

27:32 you to do is have a single repository of truth if you will

27:37 that is always in sync with the operational network and as appstore was also the tool to

27:43 design it it understands the deviations or the changes in um

27:50 expectations from that from that business so the way that we you know sort of

27:55 understood that is now you have this repo which is a staple repository that says hey you want to get some bit of

28:00 information on the device you don't have to go to five different tools to get that information so there's context with

28:06 regard to telemetry the configuration um the analytics everything is concatenated

28:12 inside of this of the apps technology it's database single source of truth and data store that's always in sync

28:19 with the network so from operational standpoint you don't have to go to 15 different tools to get the information

28:24 that you need so it's a much more streamlined and accurate bit of information

28:31 um scrolling through

28:39 the other waiting for a few more questions to pop across my board here

28:48 so what is um [Music] if operational state differs what are my options beyond alerting and

28:55 can some form a lot of remediation or suppression be enacted

29:01 yeah it's a great it's a great question there are depending on what the deviation is or

29:07 what did it the what this difference between the running state and the operational state that appstore has detected there's a few different options

29:13 that are available to use an operator at your disposal or at your ability to use

29:20 an example of this in our integrations with vmware their our ability to perform

29:26 auto remediations on differences between vlan configuration mismatches that exist in the fabric

29:31 um as opposed that we discover through the programmatic interfaces through the vmware api

29:37 so you know as an example if uh if someone configured a virtual network in the in the network and that virtual

29:43 network doesn't exist in the hypervisor uh or in the virtual infrastructure then you have an ability to auto remediate

29:49 that uh the same thing happens in reverse if someone uh in the virtual infrastructure team provision say mvlan

29:56 and the virtual network and you want to auto remove it remediate that by configuring that network inside the

30:01 fabric that exists so depending on what that what that is you may have may or may not have auto remediation options

30:08 um and beyond alerting we can also stream um you have this ability to also stream the

30:16 analytics off um in google protocol buffer so a standard based way of streaming this to additional platforms

30:22 be it splunk or kafka or whatever your tool de jour is uh as well as we have um

30:28 [Music] syslog capabilities directly from appstore

30:33 thank you for that question take a few more here

30:40 there's a fair number of people so i knew the question just popped in

30:47 uh can you give further examples of what benefits the apps are in the vmware integration yeah absolutely so

30:53 we have integrations with both vmware and the vsphere as vmware on nsxt and we do a number of validations in

31:00 real time around in mtu stranded vms in the virtual

31:08 infrastructure side that may not be backed by the proper virtual network in the fabric side or a lack of virtual network in the fabric side

31:15 we have the ability just from a visibility standpoint through the app store interface you can

31:21 see all of the virtual networks uh that exists all the vms that exist that are both

31:26 attached to the fabric is but are also part of that part of that domain and then you can

31:33 you know search for mac addresses ip addresses and vms holistically through the single interface you can see where

31:40 those vms live where they attach to etc um so yeah there's a really a number of

31:45 um really great useful integrations with the vmware stack and as we move forward we're adding

31:51 additional integrations we've done some mvps or tech previews of features with ground group based policy whereby we

31:59 were reflecting the security policy instantiated by nsx into the physical network for the customers who may have

32:04 some uh vm to bare metal use cases um so that again those are other things that

32:09 we are looking at adding into the product later in the year

32:16 so that i'm checking the time

32:22 just one more how can i integrate layer 7 security into the fabric

32:29 yeah it's a it's a good question this is the way that you would integrate layer seven into the into your data

32:35 center into your fabric really probably doesn't differ a whole lot from how you would do that today whereby you know you still

32:41 want some sort of layer seven boundary point or some sort of security device capable of player seven and and there is

32:48 you know the supported connectivity from your data center fabric from the networking devices to those security uh platforms are 100 supported in app store

32:56 today so we're not um yet wrapping any of the intent based networking capabilities around these layer 7

33:03 security device devices but we do integrate with them and allow you to move that traffic

33:08 and those applications get them um through these layer seven security you know staple engines and that

33:15 security uh and there are a number of options that that in the juniper portfolio to this end that uh

33:27 any other questions make sure that i am looking in the right places here

33:38 all right so uh one more how does this help customers align

33:45 so i'm gonna sort of summarize here is how does this uh solution help customers align day zero to day two

33:53 uh so as abstract as this single source of truth but you design the network so you

33:58 can build a network through our data model without even having access to the to the act to the physical hardware so

34:04 you can create this data center that matches precisely what you want to deploy into production and

34:09 the people doing this designing you know are typically architects and engineers etc but that same platform is going to

34:16 be used when it's delivered to the operations team and so you can model the network of day zero you can even build

34:23 what you want your network to look like at day 100 even before you ever deploy the first and so you can impart

34:30 standardizations at the very beginning of the cycle so you know a great network starts with great design and that's

34:36 where apps begin and then anything from a day to operation new virtual networks that are added that state uh in that

34:43 declaration of policy between your operations teams is always

34:49 running through the singles running through the single source of truth and this closed loop tent based analytics that identifies if there's any

34:55 deviations so it really bridges that gap between people are building things and people are operating things and allows

35:01 them to work on the same the same platform with the same goal in mind and for everyone to have you know

35:06 documentation that's up to date uh and validation of whether or not the services that they uh instantiated are

35:13 accurate and that the network is is actually delivering those services appropriately and if not giving the operators contextual information on how

35:20 to solve those problems faster than they ever have before and we do so uh in you know an open

35:27 model we have an open date or we have a data model and everything we do uh that appstore

35:33 builds and enforces from an inside is with open standards or standards-based protocols so you can interface with our

35:40 data model and the vendor an operating system and hardware that you choose underneath is a business

35:46 decision but from an operational perspective everything is multiplied

35:53 so i'll hang on here for another few seconds if you have any questions please throw them over

36:04 see how familiar do i need to be with evp and vxlan do you only support vpn vxlan

36:11 you don't have to be familiar with evp and vxlan at all appstar insurers through our reference

36:18 design that there is interoperability all of the configurations all the routing policies

36:24 whenever virtual networks are instantiated all the individual configurations you know that make that happen so you

36:30 don't have to be concerned with the underlying protocols underneath um you don't have to be an expert in that at

36:36 all you really need to know is that you want you know a service in the network from a to b

36:41 and additionally you if you're in a shop or you don't

36:47 need say ebp and bx land at all perhaps you're running something like hadoop which is an application player three

36:53 awareness cluster um or you're an entire vmware and sxd shop with 100

36:59 virtualization and the overlays are starting from the compute side you also have the option of deploying just a pure

37:05 layer 3 fabric in the same value proposition intent based analytics and et cetera are

37:10 available for both of those reference designs ebp and vxlan or ip fabric so that choice is

37:16 is yours as a business but you don't have to be dramatically an expert on any of them

37:24 uh so a good one coming in from adam says i have a network with several smaller three stage

37:32 clo fabrics would i be able to manage multiple fabrics with the same instance of abstract

37:37 yeah that is a very good point so uh appstore is purely in the management plane and what we

37:43 the software requires is ip reachability to the management interfaces of the devices

37:48 uh what that means is that you can use the same software deploy multiple different data centers under the same

37:55 bits so i can't recall if that was as part of the demo where i had multiple different data centers running but

38:01 you know this is definitely an advantage that we have over say traditional sdn um

38:09 products that are in the market that allow us to create or allow the customers to create something like

38:14 availability zones so imagine that you have a couple colos located in the same vicinity and you wanted to control them

38:20 all ubiquitously through the same interface that's 100 uh capable of within the appstore product

38:28 and thank you for the question i should say i can't recall in the demo we do have support for three stage we

38:34 have support for five stage we have support for multi-plane super spots in the five stage so in addition to vpn and

38:40 ip clove fabric we also port three five multi-plane five-stage

38:52 okay is there any other questions

38:59 all right with that i think we're start to wrap up here i appreciate everyone for

39:05 joining i appreciate for those who joined early and got to see a bit of the demo twice so that was fantastic

39:12 but if there's any other questions please don't hesitate to reach out to ourselves or anyone on the juniper team

39:18 if you'd like additional information please follow the links thank you all for joining everyone have

39:24 a great week and i hope to talk to you soon cheers

40:10 you

Show more