She's At The Edge Of The Cloud (Podcast Transcript)

Listen to the Hacker Noon Podcast on Apple Podcasts, Spotify, Google Podcasts, Stitcher, or wherever you listen to podcasts.

Amy Tom chats with Mark Gamble, the Product & Solutions Marketing Director at Couchbase, about Edge Computing. They talk about being at the edge of the cloud, moving your data operations closer to your customer, and assessing your need for edge computing. Mark explains tiered data systems, setting up data centers to avoid system downtime, computing configuration for faster response time.

In this episode, Amy and Mark talk about:

• Mark's definition of edge computing - bringing data and computing closer to the applications that are being consumed (10:00)
• How to get to the edge of the cloud and closer to the customer (13:40)
• Target suffered a 2-hour POS system outage and lost $50 Million (16:25)
• What is driving the growth of edge computing - an alternative to meeting the demand of internet users (22:00)
• The balancing act between the cost of edge computing with the risk of downtime (23:58)
• How to manage analytic processing of large amounts of data while avoiding high latency when you require a real-time response of your application (30:00)
• A layered data storage approach to dictate where the data is stored (35:00)

Connect Mark Gamble:

• Email him at [email protected]

Shownotes:

• Learn more about Couchbase's Edge Computing Solution

Machine-Generated Podcast Transcript (Please Excuse The Errors)

Amy: [00:00:00] This podcast was produced by hacker noon, hosted by me, Amy, Tom, and edited by Damian. So I just read an article the other day that talked about how Nicholas Cage's new wife was FedExed a ring to her and proposed via zoom. So that being said, I'm just opening the floor up to anyone who wants to FedEx me a ring.

You can tweet us at hacker noon for my ring size and my address information. But anyways, this is the hacker noon podcast. My name is Amy Tom, and today I am joined with Mark. He is the product and solutions marketing director at Couchbase. So I'm very excited to have you on Mark. How are you doing today?

Mark: [00:00:49] Doing going great, Amy. I'm probably not as, uh, slick and cool as Nicholas cage, but. Excited, nonetheless. Great opening. That was really Funny thank you. 

Amy: [00:01:01] So can you tell me a bit about your background? 

Mark: [00:01:04] Sure. Yeah. I, uh, come to product and solutions marketing by way of sort of a technical journey. If you will. Started in high-tech many years ago at a little data company called Illustra and founded amongst others by Michael Stonebraker. Those in the database world met recognize that name. And I was doing more items kind of stuff there, but really started to cut my teeth on data and fall in love with the power of data Balestra got bought by another little company called Informix where I went into technical marketing.

Essentially, furthered my love of and appreciation for the power of data and how it is really the center point, the lifeblood of any and all applications. And from there, I went into analytics and spent many years over a dozen years at a company called actuate, which is analytics, business intelligence.

And some machine learning, AI, that kind of stuff. And they got bought by another company called OpenText. And so that's where I made the jump from technical marketing, into actual product marketing. But I come into product marketing at least informed by, you know, being a technical marketer and even a field engineer for many years at bet.

Actuate. And that's what's really, I think now I brought me full circle back to you. Now Couchbase a database company. Again, where I started, but this is a modern database company. And so is super exciting because I think we're really poised for the sort of modern era of applications that we're seeing. And yeah, that's sort of a little bit of my background.

Amy: [00:02:46] Cool. I think this would be interesting for the listeners to hear. What would you say that the difference between technical marketing is and product marketing?

Mark: [00:02:55] Great question. And a lot of it had to do with worry and headaches a lot more we're in product marketing. So technical marketing is sort of a branch of product marketing depending on the organization and technical marketing. I had particular fun with because I came out of being a field engineer into technical marketing, where I took our products. And built use case examples out of them. So what we could then take them into financial services, opportunities, or manufacturing or oil and gas or whatever it happened to be.

I, I was tasked with actually building a showcase demonstration with the technology in the vernacular of that particular industry or use case then and so that was a super fun. I got to do lots of. You know, coding and, and geeking out, working with product managers and engineers. Then when I made the jump into product marketing, you own that role still, but you then also own things like campaigns and messaging and go to market and pricing and all of the things that sort of the trappings that go with part and parcel with creating awareness and demand for a given product.

And so that's, it's definitely the difference was sort of an Epic uplift in the amount of responsibility if you will. But I found it hugely exciting, nonetheless, because I think. There are particular advantages in being technical and bringing that insight and kind of that background into the product marketing world to help inform the messaging exercise and the exercise to promote and build stories about how your products are, you know, the greatest thing since sliced bread this kind of thing.

Amy: [00:04:40] Yeah. Yeah, definitely. And what is your educational background? And if you don't mind me asking. 

Mark: [00:04:46] Yeah, sure. Actually, interestingly, I. I have a background in broadcast engineering. I was going to be the next big disc jockey in the San Francisco Bay area. And yeah, got you know, fully educated in broadcast engineering was doing towards my FCC license, but that's where for a summer I start, I worked in the same building as the company Illustra.

And I got to know some of the guys at Illustra and they said, Hey, you're kind of technical. Cause I was, you know, doing some it things. And that's how I then made the jump over to high-tech and never looked back. So interestingly, I don't have a computer science background. But I think that actually that background helps my presentation skill. I hope hopefully this is an entertaining. Listen, anyway. Anyway,

Amy: [00:05:36] yes. I kind of feel like I am in the same boat as you. My educational background is in marketing and I have a technical. I have technical knowledge, but I've definitely learned my technical knowledge as I've gone along on the job. And just via, I don't know, osmosis reading all of these hacker noon articles that everyone writes.

So that's been great. But yeah, my background I educational background is in business. So I find myself like, yeah. Probably in a similar boat to you, because I feel like with the business knowledge and the background, you have a better understanding when it comes to tech and what drives it and the thought process behind the products that you're delivering or who to talk to or who this would be interesting to you. So I think that's a really cool standpoint to come from with regards to coming into tech. Yeah. 

Mark: [00:06:32] Totally agreeing. Yeah, it's definitely, um, I think a breadth of background is better than a myopic background, if that makes sense and practical application, I think, you know, not to, you know, look down on anybody who's done top tons of higher education, but definitely I think practical application these days in the job market is hugely sought after.

And, and I think, you know, what you've learned out in the market and what I've learned out in the market and on the job are probably equivalent. Times 10. So what we might've gotten had we gone into computer science or some other higher education, I'm just guessing of course, but yeah. 

Amy: [00:07:15] yeah. And I think it also goes to show, you know, that no matter what your educational background is in or your job experience, your journey is your own journey and you have something to bring to the table no matter where you came from.

So you don't need technical knowledge per se, or you don't need business knowledge per se, but you've got something to bring to the table. I think that's important to remember also for everybody too. Well, put goodness said it better myself. Well put actually I want to know then what was your very first job ever?

Mark: [00:07:50] So I grew up in the San Francisco Bay area, specifically in Berkeley, and there is a park up in the Berkeley Hillsville Tilden park. And my very first job was a merry-go-round operator at the Tillman park. Merry-go-round. Such a good first job. It was great. It's an antique historic landmark. If you're ever in Berkeley, go to Tilden park.

I don't know if it's open now, but six should be hopefully soon, but it was just a fantastic job in high school. Yeah. I got to meet tons of people and have fun at the same time and eat hot dogs and cotton candy. Amazing. That's great. I, my first job was working at a bike store because my grandpa used to own a bike store when I was younger.

Amy: [00:08:35] So I was a very entrepreneurial child. I probably started working there at like, Hm. I think I started learning how to do the cash at like 10, so nice. All right. Going into the retail world you later though, I definitely went into marketing because I have zero in numbers, adding multiplication skills, not my thing, money.

I have issues with that. So marketing for me, for sure. But anyway, as I want it, I wanted to talk to you today about edge computing. So I think this is a very like hot, broad topic. So what would you define? Edge computing as?

Mark: [00:09:20] Sure great, great question. Edge computing really describes this notion of a distributed computing framework.

Whose essential goal is really bringing data and computing closer to the applications that consume it and doing so by leveraging servers that are, co-located sort of being able to create sort of a tiered architecture, if you will. And when I talk about edge, I really like to point out it's one of these things that's sort of soft and definition.

If you will, there's no one thing or one definition of what the edge actually is. Edge computing is a spectrum. So on one side, there's this concept of a standalone smart device where data is co-located and processed it as embedded within the application or the client that needs it like a smartphone or on a laptop or even a desktop application.

But the best example is think of an overnight delivery driver with that rugged tablet, you know, scanning packages in the back of the delivery van. That's kind of a perfect example of this end of the spectrum, where the data is inevitably co-located on that particular device, basically mitigating any reliance on internet that that driver can process packages off to the next pickup.

On the other end of the spectrum, you have this concept of edge data centers. And these could be in the form of an it closet in the back of a restaurant or, you know, a server rack and a cruise ship or in a warehouse or a big box retail store. And these would power applications and IOT devices within their local radius and where an application falls within this spectrum really depends on the scale and number of downstream clients that it serves, but really in both cases on both sides of the spectrum, this data.

It is brought closer to and process closer to the application or clients that need to use it. It's kind of a simple concept. So it's kind of, it's interesting. This is so nascent, but catching on. And like you said, there's been a ton of analyst coverage on this topic of edge technology, and it's all emphasizing the growth of edge computing or over cloud computing for faster processing and resilience to internet latency and outages.

And, you know, and again, this sort of fuzzy definition of edge can have some nuance, depending on the perspective, like, you know, a hardware vendor or a router vendor, mobile carrier might define it within their own vernacular and with their, their own devices and hardware. But we're a database company when we define the edge, as it applies again, to date it back to data again, right.

The lifeblood of any application. So by processing data closer to the consumer of the data, You would really enable some key value points for edge computing and just really at a high level, those are, um, reliability and resilience. So you can increase guarantees of business uptime, for example, by providing uninterrupted access to data, regardless of the internet.

And, you know, being able to withstand those inevitable connection issues can help you ensure better application performance. And, and when you're facing customers or your applications are facing customers, you provide them a better experience. So that's really critical. Right with edge computing. I mean, I've been in the tech industry in marketing and business writing kind of sentences for the past, like five years.

Amy: [00:12:47] So an edge computing and cloud computing and IOT are always these like big names or terms in the industry that everybody talks about. Somehow one thing is always connected to one of these things. Um, But what does being at the edge really mean? How do you get closer to the customer in terms of computing and, um, yeah.

Can you explain that process to me? Cause I think that's what I don't understand. I think like a lot of times we talk about being on the edge and being closer to the customer and it all convoluted with a lot of marketing speak a lot of the times. And I have a hard time decoding the marketing talk. What does it mean to be at the edge?

Mark: [00:13:33] Yeah, what I'll try and do is draw in the air because they were not doing a visual presentation here, but I'll kind of describe the edge computing architecture. And to start the contrast, I'll start by imagining a simple cloud computing architecture. Okay. And this would be, you know, we're central and regional data centers reside in the cloud.

So, if you imagine just a cloud and client devices and applications, let's say under the cloud are all connected to and process data housed in the cloud over the internet. So simple. Right? And that's why people are like, Oh, the cloud is elastic. I can brew and shrink it. And everybody just connects to it from wherever.

That's great. It's all over the internet. What could be easier, but that's the problem is the internet. So when the cloud data centers available, all those applications run lickety split fast and, and, uh, just as intended. But if the internet experience is slowness or latency, the client that the applications start to slow down.

And when internet connectivity is lost, like if you go out of, you know, out of bounds of, uh, connection or being able to receive a signal, your applications can access a process date. Right. And they grind to a halt and that results in business downtime. I'll talk, I'll talk about, and just to really sort of drive this home because people think, Oh, well of course you can see the impact on IOT, which might be out in the middle of, you know, on a pipeline, up in the Arctic.

But this also has to do even with, um, in huge metropolitan areas. So business wire wrote an article a while back about target. And they suffered a two hour outage to their point of sale systems on a Saturday. Like they're one of their biggest revenue generating days. Well, why did the POS system go out across, you know, almost virtually every store?

I think across the Eastern seaboard, because they were all processing data in the cloud and they lost that conductivity. I can't remember if the data center went down, but for whatever reason, Two hours of outage that actually cost them $50 million like, Whoa, Whoa. And then there was the, the added, um, impact of consumer sentiment.

Like the people that were waiting, they couldn't even have done a traditional cash transaction because all of their inventory systems were also down. Right. So they couldn't even find something in the back, even if you just wanted to pay cash. And so the, you know, they suffered consumer backlash as well and had to regain the trust of the consumers.

And so it's really, when you start to think about, Oh, you know, we're all so conditioned by connecting to the cloud and the internet all the time. We, we lose perspective until we can't connect. We've all been there like, Oh, I'm I need to, you know, whatever I'm I'm in the store. And I'm trying to get a connection to, to get that coupon for this thing, but I have no connection.

So now let's contrast cloud computing with edge computing. Right. So cloud computing has it sort of significant point of failure. So edge computing moves data prep and processing storage closer to the applications again, by the leveraging these tiered edge data centers and embedded storage directly on devices.

So the tiering is what insulates applications from central and regional data center outages with each tier, leveraging successively, more local connectivity. And then synchronizing data within an across the tiers as conductivity permits. And so imagine if all of the stores at target during that outage instead had a micro data center running in each store, then they could have continued to process.

Irrespective of that main data center outage. And then when conductivity was restored, they can then go ahead and connect back to the cloud and still sync up everything to aggregate, but they didn't suffer that downtime to begin with. And then when you start to think about, um, processing, you know, at the bleeding edge, if you will, on the device, this then insulates, even if you have a device, that's, let's say scanning barcodes.

Or taking critical measurements on an oil rig internet goes out that's okay. We can use the local data center, but if the local data center for whatever reason is also compromised or goes out, if you're processing on the device, that device just keeps on. Processing keeps on collecting that data. And then when network connectivity is restored across the tiers, it can then synchronize again after the fact.

But the, the critical component is no data was lost. And there was no business downtime. Right. Is that okay? So, yes, but when I am using edge computing and moving my data center to the closer to the source, does that not mean that I'm managing the physicality of the data center as well? Yeah, well, you definitely are, but the real key here is that, um, you're able to create a layered approach.

And so yes, you are setting up a data center, let's say ostensibly on-prem or maybe in a private cloud, but it's definitely worth that effort for the guarantees of business uptime and the guarantees of faster response time. Everything that happens within the local radius goes exponentially faster than if it had to go all the way back to a cloud data center and come back.

And so those are where making sure that you have the. The need for that kind of business continuity, very few would say they don't, but that justifies then setting up the kind of tiered data center ecosystem, if you will. But the other need there is to make sure that your database technology has the ability to sort of recognize and talk to and replicate data across every layer where there might be in, in our case, Couchbase living in the cloud and the edge data center and on the device.

Amy: [00:19:54] Okay. So the benefit is kind of this layered tiered hybrid approach, but what's the difference between edge computing and on-prem hosting of the application?

Mark: [00:20:06] Well, really the difference is in well on-prem is a part of the overall topology. So th that's actually part of edge computing. Um, and yeah, and on-prem on device and then ultimately within the cloud, you create these three layers you're then insulating yourself from outages at successively, more local kind of processing junctures. Right. That makes sense. 

Amy: [00:20:34] Okay, cool. That makes sense. I think I have a much better understanding of this now. So what do you think then is driving the growth of edge computing? Because I feel like it's really been booming.

Mark: [00:20:46] This growing community of users who expect to do everything touchless and online, this hugely growing daily day luge of data.

I think texturey estimate that we generate as a, as a human worldwide society. 2.5 quintillion bytes of data every day. And then we have this need to leverage the data more immediately at the source. Right. And so essentially that kind of the past year, if you will represents this tremendous de-centralization, you know, consumers, workforces, and the global crisis, the health crisis drove everyone to distance and stay home, but the need to work and shop and socialized in Bowie, we just turned it.

Doing it all online. So we went from being occasionally connected to pervasively connected, I think as a world community, and this began to drive immense demand on systems that weren't originally designed to handle. It left a lot of organizations scrambling to solve these problems of, you know, growing online users, creating exponentially more data and trying to use increasingly slow and unreliable applications.

And that sort of drove the need for an alternative way to meet that demand. And we didn't really talk about 5g yet, but as it is rolled out and becomes more ubiquitous, more and more automated touchless services will become enabled like, um, curbside pickup at restaurants and retail stores. Good example is, you know, a couple of, you know, a few days ago, my wife sent me to target talking about target again, but she sent me to pick up a curbside order and I asked.

You know, where do I park? I assumed I'd have to call them or sign for items. When in fact, all I had to do was click on my way on the app, on my phone. And they took it from there. You know, they, they tracked my entire trip. They knew when I pulled in, they even knew exactly which drive up spot I was in and rolled out my order and place it in my trunk.

And I was driving with less than five minutes of a ride. Right. It's kind of terrifying. I know it was, it was sort of like, Ooh, very convenient, but. But another example. Uh, uh, and again, these are not examples where you think edge computing, but I read about five, six weeks ago talking with an analyst that taco bell is looking at investing in edge technology.

And I, and I thought, what on earth for? And just yesterday at lunch, I pulled into taco bell just to get I'll get a Chalupa or something. And, um, instead of the menu kiosk in the parking lot, there's a live person. Well, the tablet walking down the line of stores taking orders. And I thought, Oh, of course his tablet is you taking the order.

It Syncs with the kitchen. And the point of sales system taking orders was faster. It was more accurate, were personal. The service was more expedient. And I have to say, well, well, my diet, it obviously isn't the best leading at taco bell. I was really impressed with the improvement overpass visits. You know, so those sorts of experiences won't go away as this health crisis subsides, you know, it's representative of a new expected level of service that relies on data to operate properly and just can't suffer latency or downtime.

Amy: [00:23:57] Yeah. Okay. Taco bell in America is so much better than it is in Canada. Like wild. I would never go to taco bell here, but I would love to go to taco bell in America, maybe again, one day. Anyways. Secondly, I'm wondering with edge computing. Is it with cloud computing? I guess in general, one of the draws I would say is that people say it's cheaper because you only pay for what you use.

And you don't have to manage your infrastructure, but with edge computing, since you do have to put in that data center closer to your customer, would you say that it's an expensive solution? 

Mark: [00:24:37] Weighed against the expense of downtime. I'd say it's a bargain for sure. And you know, it bears mentioning while there is effort involved at setting up, you know, micro data centers and whatnot.

If you choose the right database technology, like Couchbase, it has a community edition. The technology comes for free. So all of our technology that enables edge computing comes in a community edition is open source. So that's one thing to keep in mind, but I think really weighing against the impact of loss of consumer confidence, business downtime, and in cases like where healthcare is involved, I'll even talk about some of our customers, it becomes then an overriding factor that more than justifies.

The nominal extra effort to create the local data processing. So hopefully that sort of answers the question on costs. So would you say then that this is a viable solution for small businesses or is this better suited for larger companies? I don't think actually those become the factors. Really the factors become the requirement for business resilience and.

Being able to guarantee get guarantees of uptime and guarantees of real kind of real time performance. So for example, one of our customers Molo 17 in Europe, they built a, an application called Zulu emergency. So it's an interesting name, but really what it's used is by as first responders that are in remote locations, like rescuing people that are trapped in the Alps.

And what they needed was a way to actually have all of their devices for their rescue crews to be able to communicate in sync, even in the Alps. So they're using embedded data storage, but then as they get back to areas where satellite or conductivity can be restored, all of the information that they have captured is then sinked back into the emergency rooms and to the, the other first responders who were transporting these patients.

And it seems seemingly simple. They're able to actually inform various levels of the triage response. More quickly than they used to, but they they've already said that it's actually saving lives because everybody can actually prepare in real time to some of these life-threatening critical incidents that, that they're actually taking care of.

So that's just one example where it's mobile 17, not the biggest of companies, but I don't think the size of the organization plays here as much as the requirements and needs of business uptime and speed for your applications. Okay. 

Amy: [00:27:27] well, let look, can we talk about some other questions that I might ask myself? If I was a business owner? Deciding about what kind of infrastructure to set up. So one, I think would be, um, how business critical is my uptime. What's what, it's another example of some things you might ask yourself 

how much 

Mark: [00:27:45] data do you need to leverage another aspect that we. Didn't talk about is in many industries, there's a tremendous amount of data that needs to be leveraged at the source in real time.

So consider manufacturing example where there might be manufacturing lines, big machinery, and there are are sensors that are taking readings and need to actually monitor those readings and then make decisions on behalf of the entire operation in a split second. And so in these particular cases, the tremendous amounts of data can then be acquired locally process way more quickly than trying to ship all of the data out to the cloud, which actually, you know, if you need to react in a sub-millisecond type of a environment, you just don't have time to take the data, ship it to the cloud.

Wait for some analytics process, wait for the result to come down and then make a decision. These are things that need to happen in real time or else the manufacturing line could break for example, and then you're suffering business downtime. So it's really the, uh, not only the ability to, to be resilient, but also to, to harness your data better.

In real time at the source. You know, I come from a background of analytics, machine learning, artificial intelligence. And the one thing I learned is that the more data that you have for machine learning, the more accurate machine learning predictions become. And when you've got tremendous amounts of data, if you have to move it somewhere, To run your machine learning like way up into the cloud.

You're introducing a huge amount of latency. And as, as you get more and more data for better results, that latency becomes worse. So instead move that analytic processing, uh, directly to the edge. And then, uh, you, you can just take the results, just the aggregated results and put those in the cloud. At, you know, at your leisure, if you will, but you're taking better advantage of the data, more expediently at the source.

You're reducing bandwidth throughput by only pushing through, let's say just the aggregated results ultimately to the cloud. You're also reducing the storage footprint in the cloud. And so all of these things then begin to coalesce into, again, those, those overarching benefits that you, that you get from edge computing.

Amy: [00:30:15] Right. So with the performance speed, what does high latency really look like versus being at the edge? 

Mark: [00:30:24] Let's use I'll put it in the vernacular of 5g. So because 5g is, is hot 5g and edge computing are. Joined at the hip, if you will, they are integral to each other. And I'll sort of explain a little bit about that as well.

So to sort of compare, you know, on a 4g network data, travel time from a device to a cell tower and back again, it'd be 12 to 15 milliseconds with 5g. The latency level is reduced to two milliseconds. So, a dramatic decrease an increase in speed and decrease in latency. And as such, you know, 5g is expected to support, you know, 1 million connected devices per every, you know, a third of a square mile.

And, and this is, you know, several orders of magnitude more than what's possible with 4g. And it supports, you know, multi-gigabit data transfer speeds. To contrast that when you, and so within 5g they're within that radius, you're actually processing and moving information much faster where 5g, some could say falls a little bit short is that those benefits only apply to the last hop or the edge of your network from the cell tower to the device.

But then the roundtrip from the tower to a cloud data center. And back and still take up to 500 milliseconds or more depending on where you are. So we're seeing exponential growth in milliseconds as you leave this super fast radius of 5g, if you will, just as an example, but then, you know, go out into the, the, the internet at large, you know, up to 500 milliseconds or more, that's just unacceptable for applications that need to provide real time response again.

So, you know, by moving, this is again in the example of 5g, but by moving that processing within the radius of your 5g coverage, that's one example of achieving edge computing, you then are processing within that, you know, sub-second response time environment. If that makes sense.

Amy: [00:32:40] Right. Is there a limitation on how much data I can store at the edge?

Mark: [00:32:47] No limitation really. I mean, certainly not imposed by most database technology, certainly not by Couchbase, but you know, the ability to create this tiered set of layers of various data storage. If you're again using the right database technology, like Couchbase, you can also dictate where that data actually stays in the equation.

And by that, I mean, we're starting to get into data privacy and the edge benefits for that. For example, you know, within the tiered architecture that I described, data can be replicated across all the tiers. And so you've got that complete redundancy and resilience, but in the case of where you have to comply with, you know, healthcare regulations and sensitive data or GDPR, you can actually.

Specify that the data stays at the edge in the hospital, or, you know, on the person's device where it, you know, will not be actually put out into the public domain or moved into the cloud. So edge computing has huge implications on, on also data privacy and data security. For sure. 

Amy: [00:34:01] So then would you say that it's more secure than public cloud? Let's say, 

Mark: [00:34:06] Oh, certainly. Absolutely. I would think so. And if you're leveraging the right database technology, that allows you to isolate, you know, even in this sort of interconnected fabric, topology, if you will still, as you to isolate sensitive data where it. Um, needs to be then. Yeah, absolutely. You eliminate those types of problems and you don't, while you can still leverage the public cloud and sort of that hybrid, you said it, you know, hybrid cloud concept, the edge concept allows you to keep that data sensitive data on prem or on the device where it's not violating any kind of regulations.

Amy: [00:34:45] Hmm. Okay. Yeah. Makes sense to me. Cool. And so one of the last things I want to get into with you just to bring it back to Couchbase and my quest to learn more about Couchbase. How does edge computing relate to Couchbase at the end of the day? Sure. 

Mark: [00:35:05] I mean, I'll take it back to how we started the whole conversation. It's it's data. Data is the lifeblood of any application. And without it applications don't run, you can make an argument for other components that will bring an application, you know, grinding to a halt, but without data that's really the penultimate. So first, if it might make sense to answer the question, you know, what is Couchbase Catherine's is a cloud native, no SQL database.

You know, power, giant applications. If you use LinkedIn shop at Tesco, use Comcast or Amadeus, any of these you're you're using Couchbase. Um, and it's a combination of two no-sequel concepts. First is the processing of key value information in memory or hyper fast response. And second is the distributed storage Jason document data in, in a schema-less nature for flexibility.

And to that you add support for containerization, like Coopernetti's mobile and edge capabilities and a full set of built-in features. We support CQL full-text search, inventing and analytics, things that you don't have to bolt on from other technologies. So how does that play into the computing edge computing architecture?

Well, that's supported by a combination of Couchbase products. So Couchbase server is the bedrock that's, that's what goes extensively into your central data center, your regional data centers, and even your edge data centers, that's the bedrock. And then there's Couchbase Lite. This is a lightweight embeddable version of how trainings that our customers embed right within their mobile platform and or IOT devices.

And then to facilitate that data movement and, and, you know, the integrity of synchronization between all of these different repositories of data there's sync gateway, that synchronizes data between, you know, web mobile and IOT apps running on Couchbase Lite and the back end Couchbase servers. And then that cascades up between cloud and edge data centers.

So you can, yeah, so you can, this is the way in which you can deploy calculate server in central regional data centers in the cloud, local micro data centers at the edge, or directly within the client applications and then synchronize data between them, but also at the same time, because that processing is happening at each level at each tier, if any, one of those other levels goes out, And you then compartmentalized away from, you know, the other layers doesn't matter, you're still processing and capturing data.

You're still realizing that business uptime, what kind of company would be, or industry, would it be best suited to use Couchbase at the edge? It really at one time, I would've thought it would've. I would've thought that was an easy answer, like, Oh, anybody that makes mobile applications, but, but when you start to really consider the requirements, as we go back to requirements, it's no longer, is it a small business or large business?

It's does my business, can we withstand business downtime? And we withstand. Loss of consumer confidence with slow apps that don't connect. So it really kind of comes down to those factors, you know, where you're considering, what the impact of not being able to provide those services or have that awareness would be.

And so we see it, you know, there's, there's the Molo 17 example. We have others, uh, backpack EMR has a fascinating talk. They gave it our. Connect last year there, they do mobile pop-up clinics in underserved countries and they are working literally in areas that often don't even have satellite as an option for dinner, but they're able to set up these clinics where every one of their devices.

Is using Couchbase light, embedded and talk. They're all thinking with each other. And she does a great demo. So healthcare is one. We see a lot in field services. Again, I talked about oil and gas, you know, consider the oil rig that's out in the middle of the North sea. And, you know, they need to process that data that's happening on all of their IOT sensors, and they need to watch critical factors for operation.

They can't rely on the internet for being able to do this. And so they, they actually set up data centers local on the platform. But then we also have, you know, examples that you might not consider as edge computing, Louis Vuitton. Um, leverages our edge computing technology they have in their retail stores.

Um, they actually have a retail staff carrying. Tablets that have their entire catalog and they F they also have some image recognition, which is another feature ML is a feature of Couchbase Lite. So what there, the Louis Vuitton staff are able to do is actually take a picture of something that a consumer brought in.

I want to sweater just like this, or maybe they have a picture and that then will. Recognize it match it with other similar or exactly the same items. And also let the staff know, is this in stock? Where can I find it in the inventory? If it's not in stock, it's in these other stores, it will allow them to actually order.

And so they're using edge computing at Louis Vuitton to provide a premium or personal experience for their customers. So that was three examples, you know, sort of healthcare field services in oil and gas and retail, that one might not consider as your typical edge computing type of opportunity if you will. But in fact, are, are perfect examples of leveraging that architecture. 

Yeah, that's super interesting. Makes sense to me, definitely. Okay, cool. Well Mark, thank you very much for joining the podcast. I really appreciate you coming on. If the listeners want to find you and Couchbase, where can they look?

couchbase.com. It should be your first step. And then I am at mark.Gambleatcouchbase.com. So, yeah, hit me up over email with any questions. I love this, this particular topic and, uh, if you, uh, also visit couchbase.com, there is an edge computing under the solutions menu. So you can actually get tons and tons of really good stuff built by myself, yours, truly, and my colleagues on how we support edge computing.

Perfect. I will put that link in the show notes then. Great. Thank you very much, Mark. Amy. It's been super fun. You're a talented interviewer. I appreciate the time and opportunity, and I hope you get that that ring FedEx to over zoom. We'll see what else you need to flattery will get you everywhere.

Mark. You're a welcome to come back on the podcast or whatever you like. If you like this episode, you can like share and subscribe. To the hacker noon podcast, you can get us at Twitter at Hacker noon and on Instagram and LinkedIn, same address at Hacker noon. I will talk to everybody again next week. Bye .