Navigating Networking in the Hybrid Cloud: Insights From Kyle Baxter

0:00 [Music] coming to you from The Windy City Welcome to Let's Talk Shop a podcast

0:07 about all things cloud and Enterprise Tech listen to insights and guest

0:12 interviews with it thought leaders and professionals now here's your host Elias

0:19 Can Acer welcome back everyone to another episode of let's talk shop today

0:25 is different we are going to move away from all my ex Gartner colleagues from all the cloud the strictly Cloud talk

0:32 and we're going to talk to talk to a vendor that is near and dear to my heart

0:37 I grew up with this vendor during my entire time in the data center they've always been a leader when it came to

0:44 networking so I'm I'm super excited to have a guest from Juniper Networks for

0:51 the first time Kyle Baxter senior product manager for Juniper abstra

0:56 welcome to the show thank you for coming on thank you it's a pleasure I'm excited to

1:01 be a part of your show today I am too so we're going to talk about Juniper abstra

1:07 I haven't been in the data center for a very long time so before I give this

1:13 episode a little bit of a setup tell me about your role at Juniper but tell me more importantly about Juniper abstra if

1:20 I didn't know anything about it summarize it for me go excellent yeah let's do that so as you

1:27 said my name is Kyle Baxter I'm a senior product manager here at Juniper

1:33 um so as my title indicates I work with the engineering team to to lead our

1:38 product and and vision and what we're going to be doing with our solution going forward so what Juniper abstra is

1:48 um let's start from the beginning it was actually a startup that was founded by

1:54 several industry experts former Juniper and Arista engineers and experts that

2:01 decided that the network and especially the data center can be operated in a

2:06 different in a new way there's there's got to be a better way to do it than just driving through the CLI and

2:13 depending on the CLI for everything so that was how it was started in the premise it was started with is to

2:19 completely change how data center networks are managed and drive it through what we call intent so after

2:26 invented the the term called intent based networking instead of driving your network from the CLI the goal is now

2:34 you're driving it from a business intent so what do you want your business outcomes to be and after figures out and

2:40 turns that into the CLI so you no longer need CLI experts you just need General

2:48 Network experts that can know your business outcomes and drive to the

2:54 desired outcomes and it's all done in based on stamped reliable designs so

3:00 that's one of the major Keys is we have what we call Blueprints and so whether it's a small Edge data center or a

3:08 three-stage ipclow fabric or even a large five-stage ipclow fabric we have

3:15 industry standard designs that are repeatable and you know will work every

3:21 time that you just you stamp out and you go off and run on your your business so

3:26 it changes how people think of the data center instead of looking at it as a as

3:32 a cost center it's truly just a pure cost center now it changes it to a way

3:38 where they can drive business outcomes and accelerate their business to new outcomes going forward

3:45 uh I love it and and this is one of the things that I always remember Juniper with is intent based and to be honest

3:52 with you intent based is it it's prior to AI it is the AI right that's you know

3:57 that that was exactly right so the whole point is to just instead of trying to code something or script something or

4:04 integrate something hey just do this for me this is my end result do it for me and all of this gen

4:10 AI type I think is going in that direction I think a lot of the providers be it cloud or some of the traditional

4:16 data center vendors are all moving in into that AI but Juniper was is always

4:21 there first that to be honest with you during my time at Gartner one of the things that I was thinking about is what

4:27 is the maturity model look like for clouds and at one point I was talking about the the kind of the end result of

4:35 where you want to get to from a cloud perspective is being is having an intelligence layer where you can simply

4:41 State my business outcome I'm trying to do this and I want everything else to then

4:46 automatically happens now when I was talking about it it was hey you're kind

4:52 of way into the future but you know with Chad GPT and with you know some of the

4:57 Gen AI it almost now feels like we're pretty close to that so I'm super intrigued and I hope Juniper doesn't

5:04 doesn't let go of the intent based term and the business outcomes and jump on

5:09 this AI stuff because I think you have it exactly right but I want to dwell a little bit more so

5:16 we live in a world where infrastructure is everywhere right it's not just in the data center it's not

5:22 just in the cloud anymore it's everywhere it's at the edge it's in the data center it's in the cloud it's in

5:28 multiple different clouds you have to be able to operate your on-premises

5:33 environment your Edge locations with the same ease the same user experience that you do in the cloud

5:40 so I'm curious from that perspective tell me a little bit about does abstract give me that type of ease that type of

5:48 functionality that I would find with an AWS or an Azure or Google for that matter it absolutely does

5:55 um and and the way we've done it is we've built the the product on on three major pillars that we believe are

6:01 absolutely key to driving a Data Center and to be able to do it at scale and at

6:09 speed the first pillar is is reliability and so while you know it sounds obvious

6:14 you got to have a reliable data center we all know that but it's it's how do you do that and how do you do that at

6:20 speed because everybody wants speed but they also want reliability and the

6:25 way we do that is with those blueprints that I talked about so we have you know validated designs that we can stamp out

6:33 and and repeat and do over and over and you know they're going to work every

6:38 single time we also have continuous validation so as as part of the the

6:44 solution we constantly gather Telemetry and data from all the managed devices and we can constantly and continuously

6:52 compare that against your intent so something deviates or goes out of bounds

6:57 we can alert you immediately and so that way you can take actions and and

7:02 hopefully reduce you know that mean time to repair or or even in cases where the

7:08 network always gets blamed first you can reduce that mean time to innocence

7:14 the the second pillar is the The Experience so Juniper is very focused

7:19 and it's a whole from experience first networking if you go look at any marketing campaigns that's you know

7:24 where we lead and abstra absolutely goes into that that same piece

7:30 um and and like I was talking about from the beginning the the whole goal was to to change how data centers are managed

7:37 and to create more Simplicity ease of use and drastically reduce the the

7:43 complexity because no one no one likes a solution that is complex and expensive to operate and maintain no one wants

7:50 that right so with with intent based networking you can remove that complexity because you no longer have to

7:56 focus on the the CLI and that's a huge win we have you know several customers

8:02 that are you know large Enterprises out there that love that capability because

8:07 now they no longer have to depend on you know highly skilled and certified CLI

8:12 experts and in today's age where you know people are moving and leaving and

8:17 going to other opportunities and they lose that tribal knowledge right that is

8:23 absolutely key and after can keep that you know that single source of Truth and know exactly what's going on in your

8:29 network and so that way you can you can reduce that need one of our customers even said that you know the bottom line

8:35 is you no longer need an engineer that knows Cisco Juniper Arista or Dell or whatever vendor you just need a guy that

8:41 knows how to click with a mouse but well it's maybe a little dramatic yeah yeah I wish it was that easy but I

8:48 get it yeah but but the the point there is that you you have that flexibility

8:54 now to hire um differently you don't have to look at you know what certification do you have

9:00 you just have to see can you you know um understand General Network Concepts

9:05 um and then that blends that that expertise from from your private Cloud to your public cloud

9:11 the the third pillar um is is around multi-vendor um and so this was something that was

9:18 started when when Appster was a startup before it was acquired by Juniper and it's absolutely a key that's going to to

9:25 continue um and and it's been a you know a big value that we've been able to leverage

9:31 um because obviously Supply chains are getting better but Hardware lead times are still sometimes painfully long and

9:39 so what that multi-vendor strategy allows is is now it freeze um companies and and Enterprises from

9:47 that tie-in to a specific vendor so now they can you know play vendors against each other

9:53 um and and do it you know get into a bidding war or who can get them the right Hardware the fastest because you

10:00 know Hardware has become you know more of a commodity yeah it's a commodity you know in a way

10:05 um yes everybody has their own special tweaks but you know the difference between a juniper hardware and Aristo hardware and a Cisco Hardware

10:11 there's very minor differences the difference is how you operate that and that's what app store brings with that

10:18 multi-vendor capability is now no longer what you know switch you can get if you

10:23 need 96 ports you just find whatever Hardware you can get and then you stick it in there and you can know it's going

10:28 to work because of those validated designs so how so tell me a little bit

10:33 about that I'm just so I can understand it so let's say can I get a Cisco switch and then have abstra manage it like

10:40 would it become an abstraction layer or virtualize it or absolutely absolutely so when you're building your

10:47 data center within within appstra you don't ever pick what vendor you want until the very end so you build your

10:54 whole design on your on those intent based business outcomes okay and saying

11:00 for example I want it to be you know a three-stage IV glow fabric with X number of spines and X number of leaves or

11:05 whatever subscription level you want I want you know these mini vlans for for virtual networks and routing zones and

11:11 whatever it is you need and then at the very end you say oh and by the way I want it on Cisco Hardware or Juniper

11:17 Hardware or you know if vendors work together you can mix within a fabric but

11:23 likely you probably don't or you keep certain you know layers in the fabric at

11:28 the same vendor um for just interop ease of mind and

11:34 then it apps or figures out and deduces what the actual CLI configuration is needed based on your intent so it really

11:41 frees you up from you know having to be you know a Cisco July expert or Juniper CLI expert or even narista or whatever

11:48 vendor you just have to know General vendor networking or just general networking terms

11:54 so would that also be the case for let's say a VMware environment how would it work with like let's say NSX would it

12:01 abstract would it integrate like I'm curious now yeah yeah um so we don't necessarily manage you

12:08 know the vcenter in a 60 but we have we definitely have Integrations okay because there's a definite tie-in

12:14 um apps just focus on managing the switches but there's there is specific

12:19 Integrations we have where we can get that what's going on in your virtual networks that are hosted in your racks

12:26 for the switches that are managed by apps are located um so that way you can see is what you

12:31 configured on the virtual side also matching what's configured on the switches because that's a really common

12:38 use case and for errors um is you know mismatching config between you know what's on the virtual

12:44 side versus what's on the actual switch for the network inside and so it can

12:50 flag those and even have you know simple one-click resolutions to resolve those is to fix mismatching configuration

12:58 so what about the edge so you know today there's a lot of these smaller data

13:04 centers let's say um that are spread out um I I'm assuming I can also deliver an

13:10 appliance um virtual or physical to the edge yes yes

13:16 um we have either both use cases Astra is a is a virtual machine um so you can deploy it anywhere

13:23 um so we have use cases of of companies that deploy one instance of Appstore say

13:29 in their headquarters and they you know have you know they have Wan networking to their remote site and you know they

13:36 deploy what we call a like a collapsed Fabric or maybe it's a you know small data where there's you know two two

13:43 switches and a you know basically a very you know like a collapsed spine architecture I mean you can use that

13:48 after instance to manage multiple data centers so you can use it to manage all the ones in your headquarters all your

13:55 remote data centers or Edge data centers we also even have use cases where customers have uh deployed abstra in

14:03 those remote locations if they don't have the right networking or they just don't have the the bandwidth to get the

14:09 right um you know like round trip time because there is a a limit or sure if you get if

14:15 you get over it then it's hard to to manage it especially when we're talking about that continuous validation that's

14:21 where we get stuck sometimes but um but with that we can we can have ways also where we can consolidate those

14:28 views so if you do have multiple instances of abstra you can have a single powerful view on top

14:35 so I'm sure you've heard a lot of folks are continuously optimizing their

14:41 environment there's a lot of focus on costs today there are some applications that were rushed to the cloud that

14:47 shouldn't have been in the cloud in the first place so folks are trying to figure out what to do with them and some

14:52 of the options are to bring them back so I'm curious if you're seeing some of these and how abstra is helping if

14:59 there's any customers that you can name and you know if you can't that's fine I understand but if there are any use

15:05 cases um that are in the public that you can that you can bring up I'd be curious you

15:10 know kind of how they're using abstract in in the real world yeah and there's absolutely use cases

15:16 like that um not saying you know it's a specific customer but you know very public use

15:22 case is uh is zoom everybody got very familiar with with Jim a couple years ago when uh you know the the video

15:29 conferencing exploded and and zoom used the public Cloud to expand fast because

15:34 they couldn't get enough Hardware quick enough um especially as all sudden Supply chains you know Screech to a halt they

15:41 couldn't get their hardware and so they used the public Cloud to expand and what they found and what many other you know

15:48 companies have found as they went to the cloud their margins decreased especially

15:53 when you're talking high intensity you know data types of applications like video conferencing or now it's coming

16:00 with with AI That's become you know the Hot Topic and so we absolutely see lots

16:06 of of customers moving to the back and repatriating applications you know back

16:13 to an on-prem Data Center and Appster can can absolutely help there because

16:19 we've seen a lot of return on investments on you know reducing the time to to stand up a data center

16:26 Forrester even published a report and that's out there publicly on the ROI of

16:32 abstra and what we've seen with a lot of customers that have been doing this kind of activity of standing up new data

16:39 centers to bring applications back on Prem is they sell you know over a 320

16:44 percent ROI when they used abstra 320 percent and that was you know full

16:51 payback with less than six months and they saw savings across the board from from Day Zero where it was you know

16:57 designing you know they were saving 60 percent of time day one it was turning from 24 hours to deploy every single

17:05 device to less than two hours because they could they could they could design everything ahead of time and they just

17:10 could plug the device in and it would get the config and automatically start running and then even to even the more

17:17 powerful part is the day two plus where you're continuing to operate and manage

17:22 that Data Center and they sell you know additional you know 60 plus percent savings there for reducing that mean

17:28 time to repair and in meantime to innocence like we mentioned earlier now is are these numbers compared to the

17:35 cloud or is are these numbers part of like a data center modernization like it's a little more more context yeah

17:41 exactly yeah good good question um so it's compared to using just what

17:46 tools they had today so typically that would be some kind of like ansible or python scripting if they were trying to

17:52 get some scale or just natively within the the CLI um so there was several tools as part of

17:58 that savings that they were able to get rid of so they could get rid of you know the all the ansible scripts they had

18:05 um if they still wanted to keep them they could um they could get rid of you know all the python or other operation tools to

18:11 manage and monitor the devices they were able to consolidate and get rid of a lot of those tools

18:18 so if I were to put you on the spot and ask you about some best practices for

18:24 Network optimization on-premises using abstra what what can you tell me

18:31 yeah yeah so um what we've seen and I'll give you

18:36 some real customer examples I'll start with a a large retail customer so to

18:43 kind of set the stage they used to see hundreds of of cases every day on failed

18:51 shopping carts okay and as we're all familiar in the networking world everybody would blame the network first

18:57 the shopping cart failed because the network you know was slow yeah it was slow had a problem whatever so it's the

19:03 Network's fault so they were always the first to look at cases um and so when they once they deployed

19:09 abstra and it does more than just the the design and build and deploy that's

19:14 where that day two um savings comes into because we were constantly Gathering Telemetry and

19:21 information from the devices to compare against that intent what we were able to do with that that customer is reduce

19:28 their troubleshooting time from from hours down to minutes or even be able to

19:34 provide dashboards because we have dashboards as well that have historical information where they could even show

19:40 it to other teams and they could say see if if the dashboard was green at the

19:46 time the shopping cart failed it wasn't the network wasn't the network let's go look somewhere else

19:51 um so that was that was one um another one that was also really exciting was a large internet hosting

19:58 company and so they used to to have all sorts of outages due to failed Optics

20:04 because they had a fairly large deployment and so they had Optics all over the place and they just couldn't keep up with the the Optics and

20:11 replacing them in time and so they'd have various outages all the time and so with abstra and that information we were

20:18 able to to gather we were able to look at various statistics on the Optics from

20:23 things like power voltage temperature and more to understand when an optic is

20:29 beginning to degrade so we were able to proactively help them reduce and prevent

20:35 issues so they could replace those objects before they failed so then they no longer had outages due to

20:43 um due to failed Optics and so that's where you know they were able to to Leverage leverage what we call

20:49 intent-based analytics and so that's that that constant in continuous validation that I've been talking about

20:55 against the the intent of their design

21:00 fascinating um I I got a probe so

21:06 for a while for a while I thought the innovation has moved completely just to

21:12 the cloud providers but this is like I said the intent base was always interesting to me but now you're making it even more interesting but

21:18 what's coming down the pipe what can you tell me a little bit about the future of

21:25 abstra in particular some of the capabilities that maybe we can expect not holding you to it but maybe we can

21:32 expect yeah yeah so one I'll I'll talk about that's that's quite interesting um you

21:39 kind of hinted at it a little bit um with with AI well you know we want to keep the the core what we've done with

21:46 the intent based networking we also want to explore what we can do with AI to

21:53 enhance the operator's experience and guess what if you're familiar with the Juniper portfolio we have an AI solution

21:59 if you're familiar with our Mist um offering they have a Marvis AI

22:04 component so one of the things we're doing and this isn't just just abstra in the data center but company-wide is how

22:11 can we Leverage What what mist is done with the Marvis Ai and they're they're predictive and proactive analytics and

22:19 Ai and ml models is how can we also Leverage What they've done and bring that to to Data Center use cases or even

22:27 Wan use cases um because because Juniper has a full portfolio of solutions is how can we

22:34 continue to leverage what we have in-house across our portfolio and make more and more Better Together stories

22:41 and so that is something that be on the lookout for what we're going to be doing is we'll augment what we've talked about

22:48 the power of intent based networking and the intent-based analytics and continuous validation with Marvis AI

22:57 and so that's not just to inform right so you mentioned earlier when we were talking that you know one of the the

23:04 advantages is it will be able to notify you very early on so that you can take action right but it's also true that you

23:11 can automate the responses to certain things that happen so that you're always ahead of the

23:17 the situation is that correct correct yeah we have we definitely have that in the product today where we have

23:24 um we don't necessarily Auto remedy the solution um but but we provide you with a button

23:30 because we don't want to push any configuration without you as the operator acknowledging

23:36 um but we'll provide you easy easy buttons for things like you know let's take an easier example

23:41 um miscabling um so say somebody miscabled or you know

23:46 went and changed the cabling um Appster will immediately determine that hey the cabling is wrong and it

23:52 will pull the lldp information and you can in one click say remedy my intent

23:57 with what's actually you know been plugged in and then you can then start

24:03 working again as intended so if any of our viewers or listeners

24:09 are interested to learn more by the way I'm going to have links and a whole bunch of stuff in the show description

24:15 below in the show notes but is there like a live demo environment how do they

24:20 demo app store like give me a little bit of a sense if someone wants to kind of play with it uh how does that work yeah

24:27 yeah exactly we have that we have both a YouTube channel where we have several

24:33 short you know like two to five minute long videos and we'll we'll make sure that's in the in the links as well

24:39 um that that you can go out and watch and see true business outcomes so you can see those things that I was talking

24:45 about and the value that you can get and provide with with abstra and then we also have a solution called abstra Cloud

24:52 Labs where you can sign up and get access to to Labs or pre-deploy

24:59 deployments so whether you want to see you want to see one that's actually already deployed or you want to build it

25:04 from scratch we have labs and lab guides that get give you detailed step-by-step instructions that you can see and then

25:12 also point out that we have these Juniper validated designs that take in

25:17 more than just after and they combine our entire portfolio of solutions so you can you know if you wanted to look at

25:22 how would it actually work with combining uh you know the qfx hardware with abstra with the xrx for security

25:30 and firewalls into a full full-blown solution we have that in Juniper validated designs where you can go look

25:37 at a very detailed white paper interesting So based on your experience

25:43 um how quickly do you can you get abstra to be deployed in an environment realistically like if you know if I'm a

25:49 small business if I'm a medium-sized business if I'm a large business give me all three but how easy would it be to

25:56 get started now you can always optimize you can optimize you know forever right but I'm talking initial deployment what

26:04 would you guesstimate yeah yeah so you know say a user starting from scratch they know nothing

26:10 about about abstra we have um training on our Juniper education

26:16 University Services um we have those three in like five day courses and that's usually enough to get

26:22 somebody um knowledgeable enough that they can start playing with the product um and so then you know maybe another

26:28 you know week or so playing with the product and then once they're ready to deploy it's it's incredibly fast I've

26:35 done demos where I've deployed an entire three-stage ipclow fabric with you know

26:41 with thousands of lines of configuration in minutes in just a few clicks

26:47 um so you know once you've because that's that value of those validated designs and those Stamped Out blueprints

26:52 is in just a few clicks you can say I want this design with these switches and

26:58 off it goes and deploys it and you have now a running Fabric in just minutes

27:05 so I'm also curious is there an offering for a managed service so can someone have Juniper or a juniper trusted

27:12 third-party manage this environment or is it a DIY

27:18 solution so to speak um most of our customers do it

27:24 themselves okay but we do have a handful that wanted to uh to reduce their their

27:31 overhead and so they've engaged whether it's Juniper Professional Services or we have several Partners out there that

27:37 both offer it as various levels of managed services from all the way from you know we'll completely host the

27:44 virtual machine and everything for you to you know we'll we'll help you get

27:49 started and give you training um in anywhere in between so we do have offerings via whether it's Juniper

27:55 Professional Services or or our partner community interesting so I'm I'm super intrigued

28:03 um I hope everyone that's watching is as well and I'd love to hear from you from you all if if we should bring Kyle back

28:10 for a demo of abstra and maybe he can show us the product because you know as you were talking I couldn't help you

28:16 know but but Wonder Hey you know this would be really great if I could see some of these things so yeah maybe in

28:22 the next few months here as we get into conference season for everyone maybe you

28:28 and I or You and I and someone else maybe we can we can do a live demo of some of the things that you're talking

28:33 about hopefully there'll be some newer announcements also that we can uh that we can show especially the AI uh portion

28:39 that you mentioned I think that would be that would be incredibly fascinating um absolutely I'd love to come back on

28:46 and do that what else didn't I ask you what are you hearing from customers what else should folks know about abstra

28:53 about about the the environment in general today any interesting insights that you want that you'd like to share

29:00 yeah one thing we didn't touch on was the developer experience um as you know the the scale and

29:07 evolution of data centers people are looking to do things faster and and even

29:13 get away from even just clicking in the UI they want to take it to the next you know the next stage and completely

29:19 automate it and why we have a full set of of apis our UI is actually a

29:25 first-class citizen of our of our API so every API or UI is actually utilizing

29:31 um but where we're seeing is is these customers want to take that to the next level and so what we're doing is things

29:37 with uh with things like Postman Collections and sdks to give examples

29:42 and ways to get started plus what what everybody's getting to is terraform

29:48 terraform's kind of the the new devops you know preferred tool and so we have a

29:54 recently launched terraform abstra provider that's that's out there that

29:59 that people can go out get get started with and try because that's where

30:05 probably we see a lot of the evolution of where a lot of our customers and

30:11 Enterprises in general are going to be moving towards is how can they take that to the next level and and bring in more

30:18 automation um to it you know think of being able to as a user being able to open up a case

30:25 and say Hey I want a new service deployed and it's going to require a you know a new rack and some new virtual

30:30 lands and servers deployed and you could potentially automate all that from creating a ticket and service now to

30:37 fully deploying your your service and nobody has to touch anything maybe some approvals in there but nobody would have

30:44 to actually touch any UI or do any of the configuration it could be fully automated based on you know how you want

30:51 your your environment set up that's where we see a lot of um promise and and where we see a lot of

30:58 the future going towards I'm gonna throw one more at you I can't help it you're you're just you're you

31:04 know you keep asking you know you keep saying things and my mind's going in a million different direction I love it how how long before or maybe we're

31:13 already there before abstra becomes application aware so if

31:18 the application let's say is having a hard time from a network perspective is you know maybe it's too

31:25 much Network traffic maybe there's a lot of latency is it able to rectify the

31:31 problem automatically or is it just going to notify me as quickly as possible so that I can take action and

31:38 if it's the latter will it have recommendations or is it just flagging that the application is

31:45 is having a networking issue great question so what we're doing is we're

31:51 bringing more flow data analytics later this year to begin to get that that type

31:58 of information so we can get that Network observability and deep insights into what's going on

32:03 inside your network traffic and we can you know be able to to start with

32:09 layering on you know the alerting what you talked about to let you know when there is issues or say there is a

32:15 network issue what applications are impacted or who's consuming all that bandwidth or who's hitting your Cloud

32:22 egress cost or um or or you know security vulnerabilities you know are you getting

32:28 you know attacked from the outside um all sorts of use cases where then we

32:34 can get to your other point is can you have more you know actionable um

32:40 insights or options that's where potentially that that Marvis integration

32:46 and what we'll be doing in you know likely more next year is can we then take all that information leverage the

32:53 capabilities that Marvis has within Mist which you can already do some of those kinds of capabilities in the missed

33:00 cloud with campus infrastructures and we can bring that to the data center and be

33:05 able to provide you know actions and recommended and maybe even one day proactively resolve issues before before

33:13 they go bad exactly yeah a Minority Report type of situation

33:19 exactly I've I've been saying for a while that I think the world's going to be very different six months from now

33:25 than it is today across the board all Technologies are going to to integrate

33:31 to leverage to take advantage of this this AI hype that just exploded and I am

33:37 super excited to see what Juniper does with it and based on some of the things that you've mentioned here today I think

33:43 when the next time we talk uh this is going to be a superintendo thing but let's see what happens with the demo

33:49 maybe we do a demo uh real soon and then maybe we'll do something else towards the end of the year that'll be super

33:54 interesting especially how you're going to integrate um Marvis and Mist into

34:01 um into this I think that'll be that'll be fascinating Kyle exactly thank you so much for

34:07 making time thank you so much for coming on the show I I learned a heck of a lot and um I'm gonna keep an eye on you guys

34:14 see see kind of how this is uh this is going to progress definitely it was a pleasure thank you

34:19 for letting me join and hopefully we'll get a chance to to do it again and we'll get some live demos and see this in

34:25 action absolutely all right and with that we'll see you all in the next show

34:32 thank you [Music] thank you