WEBVTT 00:00.000 --> 00:12.000 Thank you very much, Wilson. 00:12.000 --> 00:16.000 Thank you very much, Wilson. 00:16.000 --> 00:19.000 You can hear me at the back of K. 00:19.000 --> 00:21.000 This kind of volume, I'm glad to... 00:21.000 --> 00:24.000 I'm just saying we should start every talk like that. 00:24.000 --> 00:26.000 Just launching sweets at people. 00:26.000 --> 00:27.000 That's awesome. 00:27.000 --> 00:28.000 I take that away. 00:28.000 --> 00:30.000 That's my takeaway from Boston. 00:30.000 --> 00:35.000 But yeah, documenting you event architectures with Open API and 18KPI. 00:35.000 --> 00:37.000 That's what we're going to be speaking about. 00:37.000 --> 00:40.000 Before we get started, it's just myself. 00:40.000 --> 00:41.000 My name is Dave Boone. 00:41.000 --> 00:45.000 I'm a core maintainer of Open Source Project called Event Cartelog, OK? 00:45.000 --> 00:48.000 So I left AWS this last year now. 00:48.000 --> 00:51.000 It's just focused on Open Source Free Software full-time. 00:51.000 --> 00:53.000 Now, of course, it doesn't pay the bills. 00:53.000 --> 00:56.000 So I did some kind of consultancy stuff on the side. 00:56.000 --> 01:01.000 Fun fact, I was at NDC a couple of days ago, and the speakers did it there during the talk. 01:01.000 --> 01:04.000 And the someone says to me, wow, is this a picture of you? 01:04.000 --> 01:09.000 And I looked and I said, go, this is what full-time open source does to you. 01:09.000 --> 01:12.000 I'm two kids. 01:12.000 --> 01:16.000 So yeah, previously, it was eight to two years. 01:16.000 --> 01:20.000 I was a developer, I became from an architectures. 01:20.000 --> 01:22.000 It's an architecture type that I love. 01:22.000 --> 01:24.000 I actually worked at 18 KPI. 01:24.000 --> 01:29.000 I were Lucas and fan over there for just under a year, doing some open source stuff, which is cool. 01:29.000 --> 01:33.000 And I realized that's almost almost coming up to 20 years, 01:33.000 --> 01:37.000 doing startups and enterprises, and oh my god, I can't believe that. 01:37.000 --> 01:39.000 But yeah, I left open source. 01:39.000 --> 01:42.000 I'm in the right company and the right thing. 01:42.000 --> 01:45.000 The right conference is actually my first time at Pholstum, 01:45.000 --> 01:49.000 so I come here yesterday, absolutely overwhelmed, 01:49.000 --> 01:51.000 just like turning up and being like, oh my god. 01:51.000 --> 01:55.000 There's absolutely awesome to be here talking to you all. 01:55.000 --> 01:59.000 As these three things we're going to cover today, 25 minutes, 01:59.000 --> 02:01.000 that's going to be quite cool run through. 02:01.000 --> 02:05.000 The device between open API and 18 KPI and whites there, 02:05.000 --> 02:09.000 can we use both of them together, better together? 02:09.000 --> 02:12.000 Can we also, like, how does event catalog fit into this? 02:12.000 --> 02:16.000 Why is it a thing and maybe something you can help with? 02:16.000 --> 02:21.000 So, and then at the end, I've got loads of resources and slides you can dive into. 02:21.000 --> 02:25.000 So, a few people don't know actually, I actually spent eight years, 02:25.000 --> 02:30.000 eight to nine years actually working with mere cats, but even though not. 02:30.000 --> 02:33.000 So is anyone work with mere cats before? 02:33.000 --> 02:35.000 Yeah, no one, okay. 02:35.000 --> 02:39.000 I felt like this is kind of how I feel about it, obviously. 02:39.000 --> 02:43.000 But unfortunately, it's not these type of mere cats. 02:43.000 --> 02:47.000 I actually used to work for a company in the UK, but this kind of mere cat, 02:47.000 --> 02:50.000 which is like an electronic weird mere cat. 02:50.000 --> 02:53.000 So as a company in the UK, I could compare the market. 02:53.000 --> 02:55.000 We sold a lot of insurance as to people. 02:55.000 --> 02:57.000 Quite a comparison website, right? 02:57.000 --> 02:59.000 I was there for eight years, I joined at the time. 02:59.000 --> 03:01.000 Thought works were consulting for two years, 03:01.000 --> 03:05.000 and I kind of felt I hit the jackpot there because they taught me a shit load 03:05.000 --> 03:08.000 about software engineering, which is really, really cool. 03:08.000 --> 03:11.000 At the time we were doing a lot of this legacy system, 03:11.000 --> 03:13.000 and a big moral of application. 03:13.000 --> 03:15.000 This was in 2013 and whatever it was. 03:15.000 --> 03:17.000 It was just the way we built applications, right? 03:17.000 --> 03:20.000 And it thought works, Martin found everyone coming bored, 03:20.000 --> 03:23.000 and they said, yes, we got to do guess what, micro services. 03:23.000 --> 03:25.000 Okay, so we're like, yes, you know, 03:25.000 --> 03:28.000 it's like CD driven development at this point in time. 03:28.000 --> 03:30.000 So we'll do some micro services, 03:30.000 --> 03:33.000 split this core into certain things and scale it. 03:33.000 --> 03:35.000 I just actually took a five-year project. 03:35.000 --> 03:37.000 I was there for the most of that. 03:37.000 --> 03:39.000 And there's an interesting project. 03:39.000 --> 03:42.000 And what it tends to look like is something like this. 03:42.000 --> 03:45.000 These split things up, that kind of makes sense. 03:45.000 --> 03:47.000 And we were doing that. 03:47.000 --> 03:50.000 I was mainly focused here on the right, which is the UI kind of elements. 03:50.000 --> 03:53.000 React was a thing, Angular 2 was a thing, back then. 03:53.000 --> 03:55.000 So I was mainly kind of focused over here, 03:55.000 --> 04:00.000 but working very much in hand-in-hand partnerships with people doing APIs, 04:00.000 --> 04:03.000 right, people building micro services. 04:03.000 --> 04:07.000 Now, the architecture type that tends to be the trend there, 04:07.000 --> 04:10.000 but still kind of is, I guess, in some ways, 04:10.000 --> 04:12.000 is this kind of synchronous communication. 04:12.000 --> 04:15.000 Okay, so we did a lot of synchronous communication, 04:15.000 --> 04:17.000 synchronous chaining in the architecture. 04:17.000 --> 04:20.000 So you'd have, like, the client sends a request to the server, 04:20.000 --> 04:23.000 the server goes to the next service and the other service. 04:23.000 --> 04:25.000 And, yeah, you get a response back. 04:25.000 --> 04:28.000 So please get me something here with a response back. 04:28.000 --> 04:31.000 And this is kind of what that kind of looks like, okay? 04:31.000 --> 04:34.000 And now you scale this over five years. 04:34.000 --> 04:37.000 Now there are questions that start to emerge from 04:37.000 --> 04:39.000 people in the organization, right? 04:39.000 --> 04:43.000 And these questions tend to be like these, right? 04:43.000 --> 04:47.000 So like, what APIs do we have in this organization? 04:47.000 --> 04:49.000 And if you guys are free for whatever, 04:49.000 --> 04:52.000 it's just building APIs because it's so easy to do that. 04:52.000 --> 04:54.000 How can I use them or version them? 04:54.000 --> 04:57.000 Was that versioning strategy looked like in our organization? 04:57.000 --> 05:00.000 And then, finally, who's even consuming these APIs? 05:00.000 --> 05:03.000 What about telemetry and all this kind of stuff, okay? 05:04.000 --> 05:07.000 Now, luckily for us, there is something that can help. 05:07.000 --> 05:09.000 Okay, this is back in 2011. 05:09.000 --> 05:11.000 Swagger come out, okay? 05:11.000 --> 05:13.000 Swagger open API specification. 05:13.000 --> 05:16.000 So hands up if you're using open API. 05:16.000 --> 05:18.000 Yeah, most people, okay? 05:18.000 --> 05:23.000 Open API is now generally a great standard for building APIs. 05:23.000 --> 05:26.000 Most people use open APIs to do this. 05:26.000 --> 05:29.000 And it works very well for us. 05:29.000 --> 05:34.000 So Lucas is going to kill me because now I'm showing Yama on a screen. 05:34.000 --> 05:36.000 Times up, okay? 05:36.000 --> 05:39.000 Yeah. 05:39.000 --> 05:41.000 But yeah, we have machine readable code. 05:41.000 --> 05:44.000 Yama all that we can look at represents something. 05:44.000 --> 05:47.000 When we have standards in these, this kind of sway, 05:47.000 --> 05:51.000 we can also build tools and visual representations of things. 05:51.000 --> 05:53.000 Everyone loves it, fantastic. 05:53.000 --> 05:54.000 This is great. 05:54.000 --> 05:57.000 So I'm going to extend where we kind of API first. 05:58.000 --> 06:01.000 People will come back into that in two seconds. 06:01.000 --> 06:03.000 What tends to happen is this. 06:03.000 --> 06:06.000 Yeah, so every organization, every team here is like, yes, 06:06.000 --> 06:09.000 we'll have an open API specification. 06:09.000 --> 06:10.000 Thumbs up, great. 06:10.000 --> 06:13.000 If you want to learn about our API, 06:13.000 --> 06:16.000 I will point you to the API specification. 06:16.000 --> 06:18.000 And everyone's like, yes, this is awesome. 06:18.000 --> 06:20.000 This is working wonders. 06:20.000 --> 06:25.000 To the point, right, that if you wanted to introduce an API, 06:25.000 --> 06:29.000 you cannot do it without a specification in an organization. 06:29.000 --> 06:33.000 And this is where we are today in most enterprises that you'll see out there. 06:33.000 --> 06:35.000 Okay, most enterprises will turn around. 06:35.000 --> 06:39.000 If you go to any large scale, say, no, you cannot introduce an API. 06:39.000 --> 06:42.000 Where is the business case? Where is the specification? 06:42.000 --> 06:48.000 So open API is very well established over the last 15 odd years. 06:48.000 --> 06:53.000 Now fast forward to where we are today. 06:53.000 --> 06:58.000 And yeah, API is so interesting. 06:58.000 --> 07:01.000 Okay, so this is Stripe Documentation Stripe. 07:01.000 --> 07:04.000 Had like, maybe some awesome docs. 07:04.000 --> 07:08.000 So that last 5 or 6 years, they tend to be the lead in this docs space. 07:08.000 --> 07:12.000 These docs tend to bind to some kind of API specification. 07:12.000 --> 07:15.000 We can actually use these docs and check out what's going on 07:15.000 --> 07:20.000 and make pull requests, get requests, post requests, and everything else like that. 07:21.000 --> 07:24.000 So there's maturity happening in this space. 07:24.000 --> 07:28.000 But there was a problem in another space, right, history, 07:28.000 --> 07:32.000 repeating itself in a venture of in architectures, right. 07:32.000 --> 07:37.000 That everyone I speak to when I was at AWS large enterprise large customers 07:37.000 --> 07:42.000 is that governance with event architectures is always an after for. 07:42.000 --> 07:46.000 Okay, and it's an interesting thing when you start to understand this 07:46.000 --> 07:52.000 and dive deeper into this topic, is that the adoption, why is that? 07:52.000 --> 07:56.000 It's because the adoption of an architecture is growing. 07:56.000 --> 08:01.000 Now, this is an interesting thing because the venture of an architectures isn't anything new. 08:01.000 --> 08:05.000 You can go back 25 odd years as an enterprise integration pattern book. 08:05.000 --> 08:09.000 Okay, if you go pass back before then, white papers about using messages 08:09.000 --> 08:13.000 in Linux and low-level kernels and all this stuff, messages aren't new. 08:13.000 --> 08:18.000 So why is it the rage and the in-thing or whatever that might look like? 08:18.000 --> 08:24.000 And realistically, just because a venture of an architectures is just more accessible to everyone. 08:24.000 --> 08:28.000 Okay, so like as sure, we have Google Cloud doing the Pub sub AWS, 08:28.000 --> 08:31.000 where I was at we had like SNS, SQS, Event Bridge, 08:31.000 --> 08:35.000 Confluent, massive people in the Kafka space. 08:35.000 --> 08:38.000 We've got solid and gravity doing API management. 08:38.000 --> 08:41.000 There's a lot of VC money and money being pumped into this. 08:42.000 --> 08:45.000 We've got Cloud Flair, by the way, just coming up from the ranks at the moment, 08:45.000 --> 08:49.000 absolutely in everyone's top food in this space, which is very interesting. 08:49.000 --> 08:54.000 Cloud Flair doing things like durable functions with ingest and restate in temporal 08:54.000 --> 08:56.000 which are startups in this space. 08:56.000 --> 09:00.000 Durable functions give us a nice patterns we can use with events. 09:00.000 --> 09:04.000 Actually, saying, here's my code, and I'm now going to do orchestration in the code. 09:04.000 --> 09:08.000 So when an event comes in, this bit is triggered a year later, 09:08.000 --> 09:12.000 this bit triggers is in the same durable execution in the same server, 09:12.000 --> 09:15.000 as function as an example, which was impossible before. 09:15.000 --> 09:19.000 So there are things happening there, there are things happening in the space. 09:19.000 --> 09:24.000 And the problem is we get like this as easy as much easier to publish a message 09:24.000 --> 09:27.000 onto a broker onto a channel onto a queue. 09:27.000 --> 09:31.000 Okay, you can get SDK installed today within two minutes into a cloud 09:31.000 --> 09:33.000 and you're publishing messages. 09:33.000 --> 09:36.000 This is great, this is awesome. 09:37.000 --> 09:41.000 So when we start building our architectures, this is kind of what it looks like. 09:41.000 --> 09:45.000 We have producers and we have messages and we have channels, 09:45.000 --> 09:48.000 this is queues and what kind of stuff and then we have consumers 09:48.000 --> 09:51.000 putting this information off. 09:51.000 --> 09:54.000 Now when you dive deeper into a message type, you know, 09:54.000 --> 09:57.000 there are things like commands and queries and events. 09:57.000 --> 10:00.000 So a command will be a representative intent. 10:00.000 --> 10:03.000 A command is, please do something, no, like, 10:03.000 --> 10:05.000 I cannot do that here as a response. 10:05.000 --> 10:08.000 Okay, so commands, you might have seen this in kind of rest APIs 10:08.000 --> 10:09.000 and things like that. 10:09.000 --> 10:12.000 Query is the same kind of thing, please get me something, 10:12.000 --> 10:15.000 here it is, an event is a multiple fact, 10:15.000 --> 10:17.000 something in your system has happened. 10:17.000 --> 10:19.000 So an example here is my dog. 10:19.000 --> 10:22.000 You can't really see it here, but my dog's tongue is kind of hanging out 10:22.000 --> 10:25.000 as my dog is weird, but my dog was saying, 10:25.000 --> 10:26.000 like, I want a toy. 10:26.000 --> 10:29.000 Like, I can take that command and reject it and say, 10:29.000 --> 10:31.000 no, you cannot have a toy or yes, 10:31.000 --> 10:34.000 you can eat another toy that I have to purchase. 10:34.000 --> 10:36.000 So what toys are available, 10:36.000 --> 10:39.000 so I start to look in these kind of queries, okay? 10:39.000 --> 10:44.000 So in your architecture, you have these types of messages. 10:44.000 --> 10:47.000 Now what tends to happen is developers and architect 10:47.000 --> 10:49.000 as we build a venture of architectures, 10:49.000 --> 10:51.000 as we go to whiteboards and we draw boxes 10:51.000 --> 10:53.000 because we're loosely coupled, right? 10:53.000 --> 10:55.000 So we start doing this fantastic. 10:55.000 --> 10:57.000 And as someone pears into the room and says, 10:57.000 --> 10:59.000 I like this, this is really cool. 10:59.000 --> 11:01.000 I'm going to add some cues and other channels. 11:01.000 --> 11:03.000 Now we're kind of reliable, 11:03.000 --> 11:05.000 and this thing's looking fantastic. 11:05.000 --> 11:07.000 And then the other thing comes along, 11:07.000 --> 11:08.000 says, we want a piece of this. 11:08.000 --> 11:10.000 We want to continue this. 11:10.000 --> 11:11.000 And we're all like, yes, 11:11.000 --> 11:14.000 we've got loosely coupled distributed system here. 11:14.000 --> 11:16.000 And we look at it and we go build it, 11:16.000 --> 11:17.000 and it's great. 11:17.000 --> 11:21.000 And what happens is, with event architectures as this, 11:21.000 --> 11:23.000 as you get that value over time, 11:23.000 --> 11:24.000 okay? 11:24.000 --> 11:27.000 We have this perception of value that we have this bump in value. 11:27.000 --> 11:29.000 I've just produced my first message. 11:29.000 --> 11:31.000 I've got my first consumer. 11:31.000 --> 11:32.000 Awesome. 11:32.000 --> 11:33.000 Look how cool this is. 11:33.000 --> 11:34.000 Let's keep going. 11:34.000 --> 11:36.000 And that value kind of goes up. 11:36.000 --> 11:38.000 You know, we've got producers and consumers 11:38.000 --> 11:42.000 were adding things and more producers and consumers. 11:42.000 --> 11:44.000 Now at this point, 11:44.000 --> 11:48.000 something interesting happens in this journey that you go on, 11:48.000 --> 11:52.000 is that the value tends to plateau somewhere in the middle. 11:52.000 --> 11:54.000 Now why is that? 11:54.000 --> 11:57.000 Because there's the complexity on to it. 11:57.000 --> 11:59.000 And what you'll see in most people, 11:59.000 --> 12:01.000 most of the times is this. 12:01.000 --> 12:03.000 Is that complexity at some point, 12:03.000 --> 12:04.000 at some scale, 12:04.000 --> 12:09.000 just rapidly goes up with a distributed kind of big ball of mud. 12:09.000 --> 12:10.000 Why is that? 12:10.000 --> 12:12.000 Because assumptions are made as like a standards, 12:12.000 --> 12:15.000 like a governance, lack of documentation. 12:15.000 --> 12:16.000 Okay? 12:16.000 --> 12:17.000 I want you to talk. 12:17.000 --> 12:20.000 I've got this at the end. 12:20.000 --> 12:22.000 But I want you to talk in January, 12:22.000 --> 12:24.000 which is fantastic talk about software complexity, 12:24.000 --> 12:26.000 by Peter Van Hardenberg. 12:26.000 --> 12:28.000 And he says like complexity is not linked to size, 12:28.000 --> 12:31.000 but when we start to interact with things. 12:31.000 --> 12:35.000 And this is the essence of complexity when it comes to event architectures. 12:35.000 --> 12:38.000 It's because we design something like this 12:38.000 --> 12:40.000 and whiteboards in our teams. 12:40.000 --> 12:41.000 That is loosely coupled. 12:41.000 --> 12:42.000 But when we implement it, 12:42.000 --> 12:44.000 it is a completely different story. 12:44.000 --> 12:46.000 And there are things that happen. 12:46.000 --> 12:49.000 There are big balls of mud that you will come across. 12:49.000 --> 12:50.000 It's not if it's when. 12:50.000 --> 12:53.000 I cannot state the amount of people and conversations I have 12:53.000 --> 12:57.000 in the open source community that experience the same problems. 12:57.000 --> 13:00.000 These problems, some of these problems, 13:00.000 --> 13:03.000 lack of standards, assumptions, 13:03.000 --> 13:07.000 ownership, documentation, governance. 13:07.000 --> 13:09.000 So guess what? 13:09.000 --> 13:11.000 There are questions that start to emerge 13:11.000 --> 13:13.000 when you start building event architectures. 13:13.000 --> 13:15.000 These questions are, 13:15.000 --> 13:17.000 well, what events do we have? 13:18.000 --> 13:20.000 Well, okay, so what's the answer to that today? 13:20.000 --> 13:22.000 Here's a GitHub repo. 13:22.000 --> 13:23.000 Go look. 13:23.000 --> 13:25.000 Okay. 13:25.000 --> 13:27.000 What are the payloads of the events? 13:27.000 --> 13:30.000 I told you, here's a GitHub repo. 13:30.000 --> 13:32.000 Go look, right? 13:32.000 --> 13:35.000 So you say, okay, I'm going to make a change in a poor place. 13:35.000 --> 13:39.000 I'm going to add something or change something in a schema. 13:39.000 --> 13:41.000 By the way, who am I about to break? 13:41.000 --> 13:43.000 Well, we don't know. 13:43.000 --> 13:46.000 The reason we don't know is because 13:47.000 --> 13:50.000 we are told as we build a venture of 13:50.000 --> 13:53.000 an architectures that producers do not know about consumers. 13:53.000 --> 13:56.000 This is a whole point where building distributed systems 13:56.000 --> 13:58.000 and we're building as wonderful things. 13:58.000 --> 14:01.000 That producers don't know about consumers, 14:01.000 --> 14:03.000 which is an interesting thing. 14:03.000 --> 14:06.000 And this causes chaos in my opinion. 14:06.000 --> 14:09.000 So on the right, on the left, 14:09.000 --> 14:12.000 we have open API, one to one. 14:12.000 --> 14:14.000 It's quite simple things to think about 14:14.000 --> 14:16.000 what kind of response world. 14:16.000 --> 14:18.000 On the other hand, we have one to end. 14:18.000 --> 14:20.000 Think of a spiderweb now. 14:20.000 --> 14:22.000 Producers are coming and they're going, 14:22.000 --> 14:24.000 how can we manage all of this? 14:24.000 --> 14:26.000 Now, luckily for us, 14:26.000 --> 14:28.000 there are specifications and those are 14:28.000 --> 14:30.000 standards up that that can help. 14:30.000 --> 14:33.000 And this is ACENCAPI. 14:33.000 --> 14:35.000 And so for those who are here earlier, 14:35.000 --> 14:37.000 Luke has done a talk on ACENCAPI. 14:37.000 --> 14:39.000 But for those who do not know what it is 14:39.000 --> 14:41.000 and just recently joined us again, 14:41.000 --> 14:43.000 some YAML, 14:43.000 --> 14:45.000 because I don't show YAML. 14:45.000 --> 14:46.000 That's really bad actually, 14:46.000 --> 14:47.000 it's black. 14:47.000 --> 14:48.000 Yeah, you can't read that. 14:48.000 --> 14:49.000 But anyway, 14:49.000 --> 14:50.000 ACENCAPI, 14:50.000 --> 14:52.000 I just took this from the web, you know. 14:52.000 --> 14:53.000 It's a specification allows you 14:53.000 --> 14:55.000 to document channels messages, 14:55.000 --> 14:56.000 different bindings, 14:56.000 --> 14:57.000 protocol agnostic. 14:57.000 --> 14:59.000 You can actually have an ACENCAPI specification 14:59.000 --> 15:01.000 in your project from now 15:01.000 --> 15:04.000 from now so you can actually go into a service 15:04.000 --> 15:06.000 and see what messages are sent, 15:06.000 --> 15:07.000 what messages are received, 15:07.000 --> 15:09.000 and how that all works. 15:09.000 --> 15:11.000 And so no longer just going through the code 15:11.000 --> 15:13.000 and working all this out. 15:13.000 --> 15:16.000 Why I recommend ACENCAPI personally, 15:16.000 --> 15:18.000 because as standardization, 15:18.000 --> 15:20.000 standardizations allow us to have 15:20.000 --> 15:22.000 machine readable things, 15:22.000 --> 15:23.000 machine readable things that allow us to build 15:23.000 --> 15:26.000 developer tools and ecosystem around this. 15:26.000 --> 15:27.000 As I mentioned, 15:27.000 --> 15:29.000 there is protocol agnostic. 15:29.000 --> 15:30.000 So you can use anything you want, 15:30.000 --> 15:32.000 SNS, SQS, Kafka, 15:32.000 --> 15:34.000 RubberMQ, MQTT, 15:34.000 --> 15:35.000 it doesn't matter. 15:35.000 --> 15:37.000 So it goes with what you're working. 15:37.000 --> 15:39.000 There's tooling, it brings discoverability, 15:39.000 --> 15:42.000 and documentation to schemas. 15:42.000 --> 15:44.000 So again, like OpenAPI, 15:44.000 --> 15:47.000 we can do stuff and render it into 15:47.000 --> 15:50.000 some nice views and visualizations and representations, 15:50.000 --> 15:52.000 and most people like to read something like this. 15:52.000 --> 15:54.000 Some like to read like that, 15:54.000 --> 15:57.000 but we can get both of the same. 15:57.000 --> 15:58.000 So the question is, 15:58.000 --> 16:00.000 when you're building event architectures, 16:00.000 --> 16:01.000 which one do you use? 16:01.000 --> 16:04.000 Do you use OpenAPI or ACENCAPI? 16:05.000 --> 16:07.000 And it's not really a question of, 16:07.000 --> 16:08.000 which one should I use? 16:08.000 --> 16:09.000 The question is, well, 16:09.000 --> 16:11.000 what is your architecture under the hood 16:11.000 --> 16:12.000 and what kind of messaging 16:12.000 --> 16:14.000 palins are you using? 16:14.000 --> 16:16.000 So I said earlier, 16:16.000 --> 16:18.000 most of the palins that I was previously doing 16:18.000 --> 16:19.000 years ago, 16:19.000 --> 16:21.000 was more synchronous kind of requests, 16:21.000 --> 16:23.000 whereas we have this synchronous chain going 16:23.000 --> 16:25.000 on and on and coming back. 16:25.000 --> 16:27.000 And something like you might see some commands 16:27.000 --> 16:29.000 and queries over here. 16:29.000 --> 16:30.000 But in other cases, 16:30.000 --> 16:32.000 you'll see ACENCAPI, 16:32.000 --> 16:33.000 like, 16:33.000 --> 16:34.000 if I order a pizza, 16:34.000 --> 16:35.000 I'm not going to wait around 16:35.000 --> 16:36.000 and wait until I get a response, 16:36.000 --> 16:37.000 things happen, 16:37.000 --> 16:38.000 and when I order things off, 16:38.000 --> 16:39.000 Amazon, 16:39.000 --> 16:40.000 things are happening. 16:40.000 --> 16:42.000 So what tends to happen nowadays, 16:42.000 --> 16:43.000 again, 16:43.000 --> 16:44.000 with companies I'm working with, 16:44.000 --> 16:45.000 in the open source, 16:45.000 --> 16:46.000 things, 16:46.000 --> 16:47.000 and speaking to a lot of people, 16:47.000 --> 16:49.000 is people have a mixture of both. 16:49.000 --> 16:50.000 Okay? 16:50.000 --> 16:51.000 So here's an open API, 16:51.000 --> 16:52.000 specification, 16:52.000 --> 16:53.000 we have a rest API 16:53.000 --> 16:54.000 to do whatever else. 16:54.000 --> 16:55.000 Here's a message based, 16:55.000 --> 16:56.000 kind of application, 16:56.000 --> 17:00.000 and some people have both within their application. 17:01.000 --> 17:02.000 But I think there is, 17:02.000 --> 17:03.000 like, 17:03.000 --> 17:04.000 it's taken another step further. 17:04.000 --> 17:05.000 I think there are problems with 17:05.000 --> 17:07.000 specifications, 17:07.000 --> 17:08.000 okay? 17:08.000 --> 17:09.000 So on the left, 17:09.000 --> 17:10.000 we have standards, 17:10.000 --> 17:11.000 tooling, 17:11.000 --> 17:12.000 schemas, 17:12.000 --> 17:13.000 implementation details, 17:13.000 --> 17:15.000 which is fantastic. 17:15.000 --> 17:16.000 But on the left, 17:16.000 --> 17:19.000 I think we're still missing a whole lot of pieces, 17:19.000 --> 17:20.000 okay? 17:20.000 --> 17:21.000 And I do a lot of, 17:21.000 --> 17:23.000 like, a work with a demand of design, 17:23.000 --> 17:24.000 people are there in, 17:24.000 --> 17:25.000 dive into demand of design. 17:25.000 --> 17:26.000 But I think there is, 17:26.000 --> 17:27.000 a lot missing, 17:27.000 --> 17:28.000 things like semantic meaning, 17:28.000 --> 17:31.000 there's a schema of an order that has been placed. 17:31.000 --> 17:34.000 What does this mean in our organization? 17:34.000 --> 17:35.000 Why have the fields changed? 17:35.000 --> 17:36.000 What about versioning, 17:36.000 --> 17:37.000 and lack of ownership? 17:37.000 --> 17:38.000 Who owns these things? 17:38.000 --> 17:39.000 Okay? 17:39.000 --> 17:41.000 And lots of value on different, 17:41.000 --> 17:42.000 like kind of personas, 17:42.000 --> 17:44.000 understanding these schemas. 17:44.000 --> 17:45.000 It's completely different from, 17:45.000 --> 17:46.000 like a business analyst, 17:46.000 --> 17:48.000 or a product owner looking at events, 17:48.000 --> 17:50.000 versus something else. 17:50.000 --> 17:53.000 So what can we do? 17:53.000 --> 17:56.000 And this is kind of where I focus a lot of my time, 17:56.000 --> 17:57.000 and energy at the moment. 17:57.000 --> 17:58.000 There's a free, 17:58.000 --> 18:01.000 an open source tool called Event Catalog. 18:01.000 --> 18:02.000 Okay? 18:02.000 --> 18:04.000 So is anyone heard of this? 18:04.000 --> 18:05.000 Just curious? 18:05.000 --> 18:08.000 So a few people, 18:08.000 --> 18:10.000 which is really cool. 18:10.000 --> 18:11.000 So this is, 18:11.000 --> 18:12.000 not funded, 18:12.000 --> 18:13.000 it's MIT open source, 18:13.000 --> 18:15.000 this is what I do for the laws 18:15.000 --> 18:17.000 and for life at the moment. 18:17.000 --> 18:19.000 But how it works, 18:19.000 --> 18:20.000 right? 18:20.000 --> 18:21.000 So we create an Event Catalog. 18:21.000 --> 18:23.000 I'm a super geek when it comes to documentation. 18:23.000 --> 18:24.000 I just love it. 18:24.000 --> 18:25.000 I love docs as code, 18:25.000 --> 18:26.000 and all this kind of stuff. 18:26.000 --> 18:27.000 And the value it can add. 18:27.000 --> 18:28.000 We can add Mark down. 18:28.000 --> 18:30.000 We can add somatic meanings, 18:30.000 --> 18:31.000 NDX components. 18:31.000 --> 18:34.000 We can create beautiful visual representations 18:34.000 --> 18:36.000 of information in your architecture. 18:36.000 --> 18:40.000 And then we can use Git as a source of truth 18:40.000 --> 18:42.000 and create pull requests, 18:42.000 --> 18:44.000 whatever else to change this and deploy it. 18:44.000 --> 18:45.000 So what does this look like? 18:45.000 --> 18:48.000 What do you get out of the box is something like this? 18:48.000 --> 18:51.000 So now teams can actually start to understand 18:51.000 --> 18:52.000 what services we have, 18:52.000 --> 18:53.000 what events do we have, 18:53.000 --> 18:54.000 what are the schemas, 18:54.000 --> 18:55.000 what are the brokers, 18:55.000 --> 18:56.000 what are the versioning. 18:56.000 --> 18:58.000 Actually apply some of these to the domains. 18:58.000 --> 18:59.000 But you know, 18:59.000 --> 19:02.000 we actually start to use some of the components 19:02.000 --> 19:04.000 and DX components and start to build 19:04.000 --> 19:06.000 a representation of our architecture 19:06.000 --> 19:09.000 that people all can hopefully understand. 19:09.000 --> 19:12.000 So facing KPI and Open API 19:12.000 --> 19:14.000 can actually use these, 19:14.000 --> 19:16.000 to actually feel event catalog. 19:16.000 --> 19:17.000 And actually feel documentation 19:17.000 --> 19:19.000 and reasoning for your organizations. 19:19.000 --> 19:20.000 Download schemas, 19:20.000 --> 19:21.000 view things, 19:21.000 --> 19:24.000 actually when you're versioning your schemas, 19:24.000 --> 19:26.000 we can actually do something interesting here. 19:26.000 --> 19:28.000 Versioning is an interesting thing 19:28.000 --> 19:32.000 because why has this name been added to the schema 19:32.000 --> 19:33.000 to the payload, 19:33.000 --> 19:35.000 to the event, to the API? 19:35.000 --> 19:36.000 Where has it come from? 19:36.000 --> 19:37.000 Why is it like that? 19:37.000 --> 19:39.000 The people that make these changes 19:39.000 --> 19:40.000 are probably long gone. 19:40.000 --> 19:41.000 Okay? 19:41.000 --> 19:42.000 New people are coming on board. 19:42.000 --> 19:44.000 Why does this thing look like that? 19:44.000 --> 19:46.000 All this history is lost. 19:46.000 --> 19:48.000 So what happens if we can just add 19:48.000 --> 19:50.000 some kind of change log to this stuff? 19:50.000 --> 19:52.000 Or to make change logs in this area? 19:52.000 --> 19:53.000 And this is again, 19:53.000 --> 19:56.000 what event catalog can do? 19:56.000 --> 19:58.000 By the very visual person, 19:58.000 --> 19:59.000 I like visualizing things, 19:59.000 --> 20:01.000 but why can't we just have visual representations 20:01.000 --> 20:03.000 of schemas, 20:03.000 --> 20:05.000 of data, of queries, and commands, 20:05.000 --> 20:07.000 and events, and channels? 20:07.000 --> 20:08.000 And we can, you know, 20:08.000 --> 20:10.000 here is a service in the middle 20:10.000 --> 20:11.000 and on the left and right, 20:11.000 --> 20:13.000 we have queries and events 20:13.000 --> 20:15.000 that are being pushed into this service. 20:15.000 --> 20:17.000 So this is what this service accepts. 20:17.000 --> 20:19.000 On the right, we use in Kafka 20:19.000 --> 20:21.000 to represent something else. 20:21.000 --> 20:23.000 We can use standards 20:23.000 --> 20:24.000 and specifications 20:24.000 --> 20:26.000 to build these visualizations in these tools. 20:26.000 --> 20:28.000 So you can take your 18KPI document 20:28.000 --> 20:29.000 and actually visualize 20:29.000 --> 20:32.000 some of this help to help other visual learners 20:32.000 --> 20:34.000 die deeper. 20:34.000 --> 20:36.000 And again, as I said, 20:36.000 --> 20:38.000 many teams and many people 20:38.000 --> 20:39.000 are like mixing matching things. 20:39.000 --> 20:41.000 So why can't your teams just have their own 20:41.000 --> 20:43.000 event catalog, as an example, 20:43.000 --> 20:44.000 or their own spectrum, 20:44.000 --> 20:46.000 just feel the documentation? 20:46.000 --> 20:48.000 And I think there's huge value in documentation. 20:48.000 --> 20:50.000 I created this project three years ago 20:50.000 --> 20:52.000 when I was working at 18KPI 20:52.000 --> 20:53.000 as a side project, okay? 20:53.000 --> 20:55.000 And then I joined AWS. 20:55.000 --> 20:56.000 I was there for a couple of years. 20:56.000 --> 20:59.000 And this project is just slowly getting traction 20:59.000 --> 21:00.000 because people are out there 21:00.000 --> 21:02.000 in this like, a vendor and shitshow. 21:02.000 --> 21:04.000 And it's like, how can we get value 21:04.000 --> 21:05.000 out of what we already have? 21:05.000 --> 21:08.000 People are leaning into these open source projects 21:08.000 --> 21:09.000 to help them. 21:09.000 --> 21:11.000 So with that, you know, 21:11.000 --> 21:14.000 I'm just one dude with an idea. 21:15.000 --> 21:17.000 And as I said, it's not funding. 21:17.000 --> 21:18.000 I'll just do this because I enjoy it. 21:18.000 --> 21:20.000 If you're interested in any of that, 21:20.000 --> 21:22.000 you know, I recommend come check us out. 21:22.000 --> 21:24.000 We've got like almost 1,000 people on Discord now. 21:24.000 --> 21:28.000 20,000 people create catalogs in that 60 contributors. 21:28.000 --> 21:30.000 If you're interested in helping out in that space 21:30.000 --> 21:33.000 or you're a geek like me in documentation 21:33.000 --> 21:35.000 and things like that, 21:35.000 --> 21:38.000 I recommend just come check out the project. 21:38.000 --> 21:41.000 Because, you know, taking a step back, 21:41.000 --> 21:43.000 the only reason I can do this 21:43.000 --> 21:45.000 and invest a lot of time in this 21:45.000 --> 21:47.000 because what is the alternative? 21:47.000 --> 21:49.000 The alternative literally is this. 21:49.000 --> 21:52.000 I haven't seen anyone else kind of avoid 21:52.000 --> 21:55.000 this complexity that comes with event architectures. 21:55.000 --> 21:57.000 The only thing we can really do is manage it 21:57.000 --> 21:58.000 and push it down. 21:58.000 --> 22:00.000 And we can do that with open API 22:00.000 --> 22:02.000 and A-Sync API. 22:02.000 --> 22:04.000 And I believe we can't have standards in tooling. 22:04.000 --> 22:07.000 And this is where I'm at with event catalog 22:07.000 --> 22:09.000 and hopefully event catalog might be able to help you 22:09.000 --> 22:11.000 in some of that. 22:11.000 --> 22:15.000 So to summarize everything I spoke about here, 22:15.000 --> 22:18.000 I said at the start, you know, 22:18.000 --> 22:21.000 the open API specification is a standard now. 22:21.000 --> 22:23.000 You'll go out there in many enterprises 22:23.000 --> 22:25.000 and organizations and many people are using this 22:25.000 --> 22:26.000 to build stuff. 22:26.000 --> 22:29.000 Events of an architectures is completely different. 22:29.000 --> 22:30.000 We draw systems like this. 22:30.000 --> 22:32.000 We decoupled real resilient or whatever else. 22:32.000 --> 22:34.000 There will be a point in time. 22:34.000 --> 22:36.000 That complexity will come in 22:36.000 --> 22:38.000 and you've ended up in some big fall of mud. 22:38.000 --> 22:39.000 All right? 22:39.000 --> 22:41.000 Because it's not really a one-to-one mapping 22:41.000 --> 22:42.000 between rest APIs. 22:42.000 --> 22:45.000 It's like this spider web that we're trying to control. 22:45.000 --> 22:47.000 So how can we control it? 22:47.000 --> 22:49.000 I highly recommend if you haven't already, 22:49.000 --> 22:51.000 check out A-Sync API. 22:51.000 --> 22:53.000 It's a fantastic specification, 22:53.000 --> 22:55.000 fantastic community behind it. 22:55.000 --> 22:57.000 So I recommend just diving deeper. 22:57.000 --> 23:01.000 If you want to go further and look at other specifications 23:01.000 --> 23:03.000 out there, there are things happening. 23:03.000 --> 23:06.000 Cloud events, specifications is a fantastic one 23:06.000 --> 23:09.000 that mainly driven by folks at Microsoft. 23:09.000 --> 23:13.000 Cloud events is like here is an envelope 23:13.000 --> 23:15.000 and here is a message inside the envelope. 23:15.000 --> 23:17.000 Now the envelope is standard for everyone. 23:17.000 --> 23:20.000 Okay, why create envelopes all the time. 23:20.000 --> 23:22.000 So I would recommend looking back 23:22.000 --> 23:23.000 if you're using event architectures. 23:23.000 --> 23:25.000 There's a lot of those specifications, 23:25.000 --> 23:27.000 which is fantastic all about workflows and things 23:27.000 --> 23:28.000 like this. 23:28.000 --> 23:30.000 And then we've got X registry down there, 23:30.000 --> 23:31.000 which again is more around 23:31.000 --> 23:33.000 discoverability of your event architectures. 23:33.000 --> 23:35.000 Again driven by Microsoft folks. 23:36.000 --> 23:37.000 So there are things happening, 23:37.000 --> 23:39.000 and it's really cool space to be in. 23:39.000 --> 23:41.000 Personally, I feel like we could do more 23:41.000 --> 23:43.000 than just specifications for our organizations. 23:43.000 --> 23:46.000 I think we can actually have business value there. 23:46.000 --> 23:48.000 There's a lot of like stuff we can take 23:48.000 --> 23:51.000 from domain different design and actually combine the two. 23:51.000 --> 23:53.000 And that's it. 23:53.000 --> 23:54.000 That's what I do. 23:54.000 --> 23:56.000 I'm trying to do with event cataloging 23:56.000 --> 23:58.000 if you're interested, come hang out. 23:58.000 --> 24:00.000 Last year I created a book. 24:00.000 --> 24:01.000 This is free. 24:01.000 --> 24:02.000 I've been saws thing. 24:02.000 --> 24:03.000 I've done loads of whatever. 24:03.000 --> 24:05.000 There's 120 pages on everything. 24:05.000 --> 24:06.000 Event driven architectures. 24:06.000 --> 24:08.000 You can just go and help yourself. 24:08.000 --> 24:10.000 And then here at the end, 24:10.000 --> 24:11.000 everything I spoke about. 24:11.000 --> 24:14.000 This is resources from another talk 24:14.000 --> 24:16.000 but everything is pretty much in there. 24:16.000 --> 24:18.000 So if you want to dive deeper in your own time 24:18.000 --> 24:20.000 or free, you can hang out with me on X. 24:20.000 --> 24:22.000 I'm not sure who uses X anymore. 24:22.000 --> 24:23.000 So I'm an OG. 24:23.000 --> 24:25.000 OG handle here on blue sky. 24:25.000 --> 24:27.000 I haven't changed that yet. 24:27.000 --> 24:28.000 And then LinkedIn or whatever. 24:28.000 --> 24:29.000 And I'm here for the rest of the days 24:29.000 --> 24:31.000 if you're seeing me walking around 24:31.000 --> 24:32.000 that awesome. 24:32.000 --> 24:34.000 And I hope there's some value in that for you today. 24:34.000 --> 24:35.000 So thanks for spending your time with me.