1 00:00:00,019 --> 00:00:04,189 Michael: Hello, and welcome to Postgres fm, a weekly show about all things Postgres. 2 00:00:04,640 --> 00:00:08,990 I am Michael, founder of PG Mustard, and this is co host Nikolay, founder of Postgres ai. 3 00:00:09,320 --> 00:00:10,879 Hey, Nikolay, what are we talking about today? 4 00:00:11,389 --> 00:00:12,349 Nikolay: Hi, Michael. 5 00:00:12,349 --> 00:00:15,289 Uh, why Postgres and why Postgres in the past? 6 00:00:15,334 --> 00:00:20,238 Why Postgres had success still is having success, obviously. 7 00:00:20,288 --> 00:00:27,649 Let's also a little bit, talk about what, we should expect in the future in terms of Postgres popularity, usage. 8 00:00:27,679 --> 00:00:28,289 And so on. 9 00:00:28,424 --> 00:00:30,384 Michael: Yeah, I'm really looking forward to this one. 10 00:00:30,624 --> 00:00:32,694 We have a couple of inspirations. 11 00:00:32,694 --> 00:00:37,944 I wanted to give a shout out to Ryan Booz, who is starting a community blogging event. 12 00:00:37,974 --> 00:00:40,104 And this "Why Postgres?" 13 00:00:40,134 --> 00:00:42,204 is gonna be, I think, the first topic there. 14 00:00:42,209 --> 00:00:47,361 So maybe we can consider this our entry, our audio entry to that, event. 15 00:00:47,861 --> 00:00:55,491 And, also I saw a tweet not long ago, somebody asking who or what made Postgres cool, which, had a lot of interesting answers. 16 00:00:55,491 --> 00:00:59,301 And I think it'd be, I'm looking forward to hearing your takes on this as well. 17 00:00:59,611 --> 00:01:07,886 But first, actually quickly, I did want to apologize because last week, I said that, amcheck doesn't guarantee no false negatives. 18 00:01:07,886 --> 00:01:09,816 And I was, completely the wrong way around. 19 00:01:09,876 --> 00:01:11,346 It's no false positives, 20 00:01:11,396 --> 00:01:12,146 Nikolay: a huge mistake. 21 00:01:12,176 --> 00:01:12,686 Very huge. 22 00:01:12,836 --> 00:01:13,196 Yeah. 23 00:01:13,316 --> 00:01:13,646 Well, I 24 00:01:14,156 --> 00:01:14,846 Michael: don't, trust me. 25 00:01:14,851 --> 00:01:15,066 Basically 26 00:01:15,066 --> 00:01:15,266 Nikolay: joking. 27 00:01:15,956 --> 00:01:16,676 I'm joking. 28 00:01:16,676 --> 00:01:22,399 I also, I also need to confess, , I also made mistake and I actually made it in reality. 29 00:01:22,612 --> 00:01:29,882 I, only recently understood, some, case when, like amcheck when we check a lot of indexes, we usually do it in one 30 00:01:29,882 --> 00:01:35,802 single thread and if it's some temporary server, not production, of course it does make sense to parallelize it. 31 00:01:35,802 --> 00:01:36,792 And, uh, only in po. 32 00:01:36,792 --> 00:01:41,712 August 14, it was actually implemented in, in CLI tool, pg_amcheck. 33 00:01:41,732 --> 00:01:48,693 So dash J appeared so worth checking and try to use even the, for older post version. 34 00:01:48,693 --> 00:01:54,036 I personally never tried, but I suspect it should work as well for, for old versions of I checking post. 35 00:01:54,547 --> 00:01:56,107 , you can use a newer client. 36 00:01:56,107 --> 00:01:56,347 Right. 37 00:01:56,679 --> 00:01:57,189 Okay. 38 00:01:57,281 --> 00:01:58,031 Enough confession. 39 00:01:58,271 --> 00:02:01,541 I also want to thank you everyone again for feedback. 40 00:02:01,661 --> 00:02:07,751 This is super important and I also want to send everyone, uh, like some words of support. 41 00:02:07,751 --> 00:02:12,311 So in hard times, I hope work distracts from bad news and so on. 42 00:02:12,731 --> 00:02:18,131 And I hope Postgres isnt well in, well conditioned everywhere, well maintained and so on. 43 00:02:18,563 --> 00:02:19,373 Michael: absolutely. 44 00:02:19,986 --> 00:02:21,966 Well, so why Postgres? 45 00:02:21,996 --> 00:02:24,366 Should we start with a little bit of history? 46 00:02:24,496 --> 00:02:30,976 So maybe we could do a quick summary or do you want to start with where things started to change or, , where would you like to start with this? 47 00:02:31,606 --> 00:02:34,730 Nikolay: Well, being a human, I always have some right? 48 00:02:34,751 --> 00:02:40,692 When, uh, I, uh, stopped and needed to choose some free database system. 49 00:02:41,152 --> 00:02:43,822 , I, I moved away from Oracle and SQL Server. 50 00:02:44,288 --> 00:02:55,152 Um, MySQL was considered bad very quickly in my mind, and I chose Postgres because it, behave much better and closer to what I learned at university. 51 00:02:55,269 --> 00:03:00,249 So yes, unlike my sequel, this is, was my reason why Postgres, but was very long ago. 52 00:03:00,249 --> 00:03:02,169 Almost 20 years actually. 53 00:03:02,570 --> 00:03:10,829 Michael: So are we talking, SQL Standard Compliance or are we talking around, acid compliance or how would you summarize that? 54 00:03:11,053 --> 00:03:15,623 Nikolay: SQL Standard compliance is one thing and indeed a very good reason. 55 00:03:15,623 --> 00:03:25,260 And it was so, but also, like logical, predictable behavior, you expect something, but, , with my sequel, for example, this terrible thing we had 56 00:03:25,320 --> 00:03:36,594 with my sequel that time, if you have, February 28 and add one day, you end up having March 3 or something like this, or date 0 0 0, uh, behavior of nulls. 57 00:03:36,954 --> 00:03:39,284 Although, although Postgres behavior of nulls also interesting. 58 00:03:39,614 --> 00:03:48,149 So in MySQL, I started to encounter too many things and some, not only a SQL standard, but for example, principles, A C I D principles. 59 00:03:48,366 --> 00:03:49,292 In MySQL. 60 00:03:49,420 --> 00:03:52,930 If you need the full check search, you needed to use my sum. 61 00:03:53,110 --> 00:04:01,333 So my sum required repair table all the time because it was not, , didn't follow a C I D principles in terms of wall, records and so on. 62 00:04:01,727 --> 00:04:03,437 So pos behave much better. 63 00:04:03,723 --> 00:04:08,463 But it had limitations and lack of some functionality, of course, like replication at that time and so on. 64 00:04:08,733 --> 00:04:11,533 So we needed to deal with slowly then long debt. 65 00:04:11,673 --> 00:04:14,596 It was not fun, but still, this, this is my path. 66 00:04:14,764 --> 00:04:22,075 But, at that time, I felt clearly that POGO is considered as not number one choice for most people. 67 00:04:22,675 --> 00:04:32,134 And I felt it clearly, like until like 2014, I was thinking like, PO this is for spec. 68 00:04:32,164 --> 00:04:36,034 It's like free bsd, you know, most people choose Linux, but you choose free bsd. 69 00:04:36,364 --> 00:04:38,854 Or even, even not, maybe it's not good comparison. 70 00:04:39,094 --> 00:04:43,174 It's like Linux compared to Windows for personal computers. 71 00:04:43,543 --> 00:04:47,263 You choose it, you know why, But most people still prefer Windows, right? 72 00:04:47,743 --> 00:04:49,993 This is, this was my feeling at that time. 73 00:04:50,503 --> 00:04:52,828 But of course, uh, I thought it's okay. 74 00:04:52,858 --> 00:04:54,405 Like, we have strong reasons. 75 00:04:54,448 --> 00:04:56,338 we won't switch to my sequel. 76 00:04:56,428 --> 00:04:56,931 Never. 77 00:04:57,052 --> 00:05:01,762 We know why, but it's okay that most people around are preferring different database system. 78 00:05:01,852 --> 00:05:03,742 Well, poor of those people, right? 79 00:05:03,982 --> 00:05:06,581 This was our, way of thinking. 80 00:05:06,731 --> 00:05:10,183 I mean, our, in terms of like community, I remember this. 81 00:05:10,607 --> 00:05:16,216 And sometimes of course, we were even blamed that we, advertise pogs too much. 82 00:05:16,238 --> 00:05:18,397 It's like, sect, almost sec, you know? 83 00:05:18,685 --> 00:05:25,225 And it was until 2000 13, 14, 15, when suddenly things changed. 84 00:05:25,362 --> 00:05:35,305 They changed earlier, I think, but around 2013, maybe 12, we had some point, when Pogo suddenly became a winner compared to MySQL. 85 00:05:35,325 --> 00:05:36,380 But why, right. 86 00:05:36,867 --> 00:05:41,085 Michael: Yeah, I want to go back, quite far to introduce when some things happened. 87 00:05:41,085 --> 00:05:43,576 I think we'd want to talk about, in a moment. 88 00:05:43,812 --> 00:05:46,932 But for people that don't know, I think some dates will be quite useful. 89 00:05:47,172 --> 00:05:50,266 I've done some research and just going right back. 90 00:05:50,791 --> 00:05:58,239 If we go back to when a lot of this all started, there were a lot of research projects in the, in the seventies and eighties that, a lot of these projects 91 00:05:58,239 --> 00:06:03,329 came out of universities, Ingres and Oracle came out in the seventies. 92 00:06:03,329 --> 00:06:04,750 Those, actually have similar roots. 93 00:06:05,000 --> 00:06:12,437 In the eighties, Postgres started as a post ingress project, and even from the very beginning it had a lot of 94 00:06:12,437 --> 00:06:15,407 the foundations I think have proved very important today. 95 00:06:15,412 --> 00:06:18,467 So things like its extensibility and vcc. 96 00:06:18,756 --> 00:06:22,677 Then we've got a few other important dates here that I think we might come back to. 97 00:06:22,677 --> 00:06:25,977 We've got Microsoft SQL Servers first version in 89. 98 00:06:26,177 --> 00:06:34,252 then we've got Postgres being released under an MIT style license, which I believe is the Postgres license in, as 94. 99 00:06:34,642 --> 00:06:36,202 I think that's been really important. 100 00:06:36,682 --> 00:06:43,344 Then, Postgres 95 added SQL support or sql support feels like a really important date. 101 00:06:43,944 --> 00:06:45,531 And then, I've got a bit of a gap. 102 00:06:45,531 --> 00:06:49,671 In fact, actually another important thing in 95 was my SQL version one came out. 103 00:06:49,905 --> 00:06:51,165 So I think that lays a lot. 104 00:06:51,165 --> 00:06:53,775 Nikolay: 95 was absolutely amazing year. 105 00:06:53,775 --> 00:06:57,615 Actually, internet started to work actually that year, I think. 106 00:06:58,005 --> 00:07:07,861 And not only you mentioned PSUs and my, but if we look at languages, Java, Java Script, PHP and Ruby, all were born in 95. 107 00:07:08,491 --> 00:07:09,091 Can you imagine? 108 00:07:09,271 --> 00:07:09,811 Michael: Wow. 109 00:07:09,991 --> 00:07:10,171 Yeah. 110 00:07:10,171 --> 00:07:12,751 And Windows 95 was a massive release as well, wasn't it? 111 00:07:12,751 --> 00:07:12,871 Nikolay: Yeah. 112 00:07:12,871 --> 00:07:20,322 So it was like, I think 95 and maybe 2014, some changing years, like in many areas. 113 00:07:20,676 --> 00:07:20,916 Michael: Yeah. 114 00:07:21,396 --> 00:07:25,075 So I was gonna say that, , it's probably massively overgeneralizing. 115 00:07:25,075 --> 00:07:32,455 And I think your free BSD example is great because for any small system there are always going to be some really passionate 116 00:07:32,455 --> 00:07:35,315 people that there have been people that have been using Postgres. 117 00:07:35,795 --> 00:07:41,245 Nikolay: I want to give this whole time, I want to take my words back, not FreeBSD, but Linux versus Windows. 118 00:07:41,305 --> 00:07:43,045 This is a better feeling actually. 119 00:07:43,135 --> 00:07:43,435 Michael: Okay. 120 00:07:43,668 --> 00:07:44,268 Yes. 121 00:07:44,374 --> 00:07:50,764 I guess the thing here though is we've got, with Postgres, we've got a, a changing of the guard I feel a little bit, whereas I don't think 122 00:07:50,764 --> 00:07:54,614 that's happened with Linux and Windows for personal computers yet, 123 00:07:55,244 --> 00:07:56,384 Nikolay: maybe Okay. 124 00:07:56,637 --> 00:08:05,270 Slowly this technology is winning more and more hearts because it's just better because it follows better principles. 125 00:08:05,270 --> 00:08:08,180 For example, very strong CLI everywhere. 126 00:08:08,240 --> 00:08:17,002 PSGs has excellent cli and Linux has excellent cli and principles like don't write too much if everything is fine, better keep silence, right? 127 00:08:17,032 --> 00:08:21,250 Like if success just keeps silence, exit code, like return code, and that's it. 128 00:08:21,507 --> 00:08:26,137 While other tools can very chatty, right? 129 00:08:26,438 --> 00:08:34,148 This is small things and these principles, uh, very like Linux is, is all about a lot of CLI tools, so many tools. 130 00:08:34,574 --> 00:08:40,506 And Postgres has very strong psql and a lot of things surrounded there are many common things here. 131 00:08:40,686 --> 00:08:44,912 So FreeBSD maybe it's too different compared to Linux. 132 00:08:45,619 --> 00:08:46,891 Michael: Yeah, fair enough. 133 00:08:47,221 --> 00:08:58,454 So, my understanding is that for a lot of these early years, if you had a serious workload, chances are you're gonna be running on a paid for product. 134 00:08:58,514 --> 00:09:03,254 Probably something like either Oracle, maybe even ibm. 135 00:09:03,584 --> 00:09:12,156 , but then as time went on, SQL Server or Microsoft SQL Server became the main competitor to Oracle for these big serious workloads. 136 00:09:12,726 --> 00:09:21,523 And then for smaller workloads, or for hobby projects, lamp Stack became popular with MySQL as the database. 137 00:09:21,823 --> 00:09:28,153 Uh, I think things like WordPress picking MySQL as its backend seemed really important for its adoption. 138 00:09:28,483 --> 00:09:37,040 So I think we had like a, a phase where if you had money and you were a serious company, you'd go for one of these paid for proprietary systems. 139 00:09:37,162 --> 00:09:44,446 And if you didn't or you were a startup, maybe if you, if you wanted sequel database, you'd go with MySQL. 140 00:09:44,751 --> 00:09:48,141 Is that roughly your understanding, like how you saw it as well? 141 00:09:48,501 --> 00:09:58,028 Nikolay: Yeah, and also important availability and for, for those who don't want to pay some websites and like apps and so on, startups 142 00:09:58,155 --> 00:10:06,645 for them it's important availability and how easy, like the cost of maintenance when you start, like how easy to start and so on. 143 00:10:06,645 --> 00:10:12,071 And for Postgres I remember very well, it was for a long time it was blamed for high barrier of entry. 144 00:10:12,345 --> 00:10:16,215 So it's very difficult for new users to start with it. 145 00:10:16,701 --> 00:10:19,491 For example, I installed it and I cannot connect to it. 146 00:10:19,551 --> 00:10:22,281 This was quite long problem for podcast. 147 00:10:22,472 --> 00:10:27,632 Something for maybe smart people, but barrier so high, why should we spend time for it? 148 00:10:27,825 --> 00:10:35,400 MySQL is, it's easy to start and easy to maintain in the beginning, but I remember actually in 2009 or 10, 149 00:10:35,460 --> 00:10:38,880 I attended MySQL, uh, user group in Moscow, Russia. 150 00:10:39,450 --> 00:10:48,869 And I remember that it was, about performance and they, at the time already, they compared themselves with Postgres and they admitted that 151 00:10:48,869 --> 00:10:52,639 Postgres is winning in performance for one node at that time very well. 152 00:10:52,639 --> 00:10:59,990 Like they said, you know, Postgres is better here, here, here, here, because at that group, MySQL developers were, I mean, hackers. 153 00:11:00,080 --> 00:11:07,678 They also were there and they are discussing like we compare my SQL with and we are losing what to do about. 154 00:11:08,106 --> 00:11:13,376 . And it was interesting because also Postgres had the image that, uh, it's also slower than MySQL. 155 00:11:13,656 --> 00:11:16,633 It still has in many minds actually, but it's not so. 156 00:11:17,111 --> 00:11:24,125 And it, it's interesting, I realized that, Postgres is good even among MySQL users and fans. 157 00:11:24,368 --> 00:11:28,418 So it's controversial, but I think for popularity it's super important. 158 00:11:28,638 --> 00:11:34,488 how accessible it is, for example, does your operational system already have it installed or it can be installed in one line? 159 00:11:34,801 --> 00:11:43,952 and, uh, also providers, if you create some website, uh, you have some hosting, Is MySQL available there or Postgres available there? 160 00:11:44,002 --> 00:11:49,392 If only MySQL, and it was so in many cases, you will prefer MySQL because it's available, you don't need to 161 00:11:49,392 --> 00:11:52,551 install it and provide gives it to you very quickly. 162 00:11:52,551 --> 00:11:53,421 It's important, right? 163 00:11:53,421 --> 00:11:55,191 For small, small projects. 164 00:11:55,386 --> 00:11:56,376 ? Michael: Yeah, I think so. 165 00:11:56,376 --> 00:12:05,281 I think developer experience is important, but, but I also might argue that Postgres hasn't changed that much from a database point of view, But 166 00:12:05,281 --> 00:12:14,046 I think a lot of frameworks and cloud providers and things have really helped there in terms of making it easier to spin up or to connect to. 167 00:12:14,076 --> 00:12:15,576 So I, I'm not, Yeah, I don't know. 168 00:12:15,576 --> 00:12:20,110 Have you, did, have you noticed a change in the, tooling, for example, like some of the issues you mentioned, 169 00:12:20,115 --> 00:12:22,783 like being able to connect to it becoming more easy? 170 00:12:22,783 --> 00:12:23,713 Nikolay: Well, I don't know. 171 00:12:23,713 --> 00:12:30,658 Like, it's hard to me to, like, I, I understand like, like availability is super important, but I remember when 172 00:12:30,658 --> 00:12:33,088 I started Postgres, of course, I compiled from source. 173 00:12:33,198 --> 00:12:34,578 I could not install it from package. 174 00:12:35,004 --> 00:12:39,867 Maybe packages were not available, but over time it became much more available. 175 00:12:39,927 --> 00:12:49,897 Like all Linux distributions started to have packaging, up to date packaging with all contr models, all popular models of course, and so on. 176 00:12:50,197 --> 00:12:58,517 Of course, ye and up systems are most important here, and they both have POEs available there, but you don't need to compile it. 177 00:12:58,560 --> 00:13:05,453 and this means that most admin on hosting, they start to just add it because why not? 178 00:13:05,453 --> 00:13:06,683 It's, it's there, right? 179 00:13:06,683 --> 00:13:11,383 So somehow it was, I think it was slow revolution and, how to connect. 180 00:13:11,505 --> 00:13:13,551 we have some small documentation. 181 00:13:13,641 --> 00:13:20,999 For example, I remember I always end up, when explaining people how to install and connect Pogo on Linux, on, on Bo, on CentOS. 182 00:13:21,189 --> 00:13:21,959 I, I always. 183 00:13:22,185 --> 00:13:30,704 Found the digital ocean documentation like already, like maybe up to 10 years ago, they had excellent to date documentation, how to upgrade a new version. 184 00:13:31,064 --> 00:13:40,603 And uh, so this howto documentation actually pause this documentation official documentation is still lacking this halto, types of documents 185 00:13:40,993 --> 00:13:45,345 describing for particular linear distribution how to do particular tasks. 186 00:13:45,499 --> 00:13:48,229 But others like digital ocean, they fill this gap. 187 00:13:48,589 --> 00:13:52,601 And, this is important I think, and uh, it slow process. 188 00:13:52,788 --> 00:13:59,358 Michael: I think this takes us really nicely onto some reasons why Postgres did then succeed in spite of those things and why 189 00:13:59,358 --> 00:14:02,778 it has becomes and is becoming more and more popular each year. 190 00:14:03,178 --> 00:14:07,768 I wanted to take, take us back time wise to a few other important dates. 191 00:14:08,218 --> 00:14:12,238 Um, I looked at the, I think PostGIS has been really important for Postgres. 192 00:14:12,268 --> 00:14:18,266 I think a lot of people describe it as even compared to the different commercial offerings out there, the most advanced, 193 00:14:18,378 --> 00:14:22,753 GIS, uh, GIS system out there, and that's a Postgres extension. 194 00:14:22,933 --> 00:14:29,035 And I think if it wasn't for Postgres extensibility at a deep level, That wouldn't be possible. 195 00:14:29,215 --> 00:14:30,865 So I think that's been huge. 196 00:14:30,895 --> 00:14:37,599 And it came out in 2005, which I couldn't believe, and I, I don't know when it became, or when it started to become competitive with 197 00:14:37,599 --> 00:14:40,769 the commercial offerings, but that feels really important to me. 198 00:14:41,069 --> 00:14:48,604 And then the, the other dates that I thought were super interesting were couldn't believe it was as early as 2007 that Heroku chose 199 00:14:48,694 --> 00:14:53,030 Postgres as its, default, or the database that they offered in 2007. 200 00:14:53,030 --> 00:14:53,630 Could you believe? 201 00:14:53,840 --> 00:14:57,170 But yeah, I would then point back at why could they do that? 202 00:14:57,470 --> 00:15:00,830 Probably because of its MIT style license. 203 00:15:01,010 --> 00:15:02,750 It could, it was extremely permissive. 204 00:15:02,750 --> 00:15:05,540 And they could do that without having to consult with anybody. 205 00:15:05,760 --> 00:15:12,335 They could choose it because of those decisions that Postgres made early on, it was reliable and very usable and all those things. 206 00:15:12,695 --> 00:15:19,085 And they could take away some of those developer experience problems of getting started with it cos they could implement 207 00:15:19,085 --> 00:15:21,665 and make them easy for their, the people on, on Heroku. 208 00:15:21,915 --> 00:15:22,305 Nikolay: Right. 209 00:15:22,305 --> 00:15:27,225 And, uh, this influenced a lot of rub, develop, rub engineers, right. 210 00:15:27,275 --> 00:15:32,705 Ruby's support was excellent with Heroku and it, it's so natural for you to choose Postgres. 211 00:15:32,705 --> 00:15:37,655 If you, if you already chose Ruby and chose hiku, it was number one choice. 212 00:15:37,655 --> 00:15:38,105 Yeah, of course. 213 00:15:38,452 --> 00:15:41,602 so this like provider symptoms is very important. 214 00:15:42,002 --> 00:15:52,184 I agree with you with PostGIS and extensibility and features, like always parts but it's only parts of like, only small parts of whole puzzle. 215 00:15:52,189 --> 00:15:52,424 Right. 216 00:15:52,784 --> 00:16:00,767 But, what happened with MySQL, I think also important because MySQL , couldn't consider anymore as free software at some point. 217 00:16:00,767 --> 00:16:01,037 Right. 218 00:16:01,470 --> 00:16:02,530 As total free. 219 00:16:02,530 --> 00:16:03,090 Michael: Well, I think this might. 220 00:16:03,815 --> 00:16:11,082 This might be my, if it's not the biggest, I think this might be the biggest bit of luck that Postgres got in the whole, you know, I think 221 00:16:11,082 --> 00:16:15,285 everything else it's, from deliberate choices and being very, very good. 222 00:16:15,675 --> 00:16:23,023 But this I think was quite lucky that of all the companies that could have bought, uh, well, so Oracle acquired Sun, didn't they? 223 00:16:23,426 --> 00:16:28,956 which included MySQL and that could have gone a few different ways. 224 00:16:29,166 --> 00:16:32,676 But, , people had a distrust of Oracle already, I think. 225 00:16:32,816 --> 00:16:42,700 And crucially, one of the creators of Myco, the main person I believe behind it, split off to create my, uh, Maria db. 226 00:16:43,671 --> 00:16:47,481 My understanding is he couldn't choose as permissive license as my sql. 227 00:16:47,481 --> 00:16:54,355 We didn't want to, to avoid my SQL taking all of the, changes from his fork onto, onto theirs. 228 00:16:54,539 --> 00:17:03,842 So that split in the community, in the development and in the trust of the community, I think really opened an opportunity 229 00:17:03,847 --> 00:17:07,412 for Postgres, like to become the, the default choice. 230 00:17:07,412 --> 00:17:12,892 If you, if you were a startup and you needed a free database, or you wanted you, maybe you couldn't afford SQL server 231 00:17:12,892 --> 00:17:16,142 or Oracle, or maybe you, you really valued open source. 232 00:17:16,142 --> 00:17:19,082 I think suddenly people were questioning should we go with that? 233 00:17:19,532 --> 00:17:24,062 And then when people were looking for options, Postgres was a really good alternative. 234 00:17:24,804 --> 00:17:27,186 So, we had a little internet outage, but we're back. 235 00:17:27,489 --> 00:17:28,989 where we left off was around 236 00:17:29,019 --> 00:17:35,889 Nikolay: let's, we, we haven't, we considered like growing popularity as some well known fact, but let's discuss 237 00:17:35,894 --> 00:17:39,909 why we think, uh, it's obvious that Postgres is winning. 238 00:17:40,359 --> 00:17:43,404 I have two sources. 239 00:17:43,614 --> 00:17:48,450 One is db engines.com, and their methodology is quite, complex. 240 00:17:48,450 --> 00:17:54,665 Maybe not clear to me, but still this is people consider this as a reliable source of truth. 241 00:17:54,665 --> 00:18:01,498 And obviously Pogo is growing, growing pot is growing good, but there is another source of, information, which, uh, 242 00:18:01,498 --> 00:18:05,758 is very simple and, uh, methodology, quiet step forward. 243 00:18:06,098 --> 00:18:09,091 It's, uh, Hacker News, job postings. 244 00:18:09,418 --> 00:18:14,904 they have every month they have who is hiring, and usually it's more than 1000 replies. 245 00:18:15,789 --> 00:18:18,879 It includes everything, but mostly from startups. 246 00:18:19,089 --> 00:18:22,846 Sometimes it's grown startups, sometimes it's very like new startup. 247 00:18:23,110 --> 00:18:23,290 Yeah. 248 00:18:23,320 --> 00:18:26,742 But, uh, there is HN trends, hack news trends.com. 249 00:18:26,930 --> 00:18:32,382 Uh, we will provide link and they, from time to time, they analyze all those, texts. 250 00:18:32,442 --> 00:18:33,492 It's just text, right? 251 00:18:33,821 --> 00:18:36,473 And extract technologies, Everything. 252 00:18:36,473 --> 00:18:41,693 Like from react to remote world, how many postings have remote? 253 00:18:42,053 --> 00:18:54,009 So can you imagine POGS has 15% in all job postings, including marketing people, support people, non-technical people at all, And this is 15. 254 00:18:54,009 --> 00:18:55,299 It's very, very high. 255 00:18:55,299 --> 00:18:56,259 I think it's top 10. 256 00:18:56,439 --> 00:19:01,439 And it, like, it's on par with word remote probably is the number one still after Covid. 257 00:19:01,599 --> 00:19:03,489 But reactors react also. 258 00:19:03,642 --> 00:19:04,572 quite a word, right? 259 00:19:05,111 --> 00:19:11,981 GitHub starts is also some reliable source of truth and Supabase, which I consider are the champion in terms 260 00:19:11,981 --> 00:19:16,396 of productizing, pog itself and many extensions of it. 261 00:19:16,703 --> 00:19:21,906 And products like PostgREST, they productize it very well and they have a lot of GitHub stars. 262 00:19:21,906 --> 00:19:28,866 I think it's how many dozens of thousands And the growth is better than react growth. 263 00:19:29,379 --> 00:19:29,619 Yeah. 264 00:19:29,649 --> 00:19:30,009 Right. 265 00:19:30,189 --> 00:19:32,739 These, all these numbers are quite reliable. 266 00:19:32,889 --> 00:19:36,099 I mean, these, uh, drop postings and hub stops. 267 00:19:36,436 --> 00:19:36,736 Michael: Yeah. 268 00:19:36,856 --> 00:19:39,321 I think it's very hard to refute that. 269 00:19:39,321 --> 00:19:47,066 And I think there's also a growing feeling on places like Hack and News, but I, I know it's a very biased source, but the sentiment 270 00:19:47,066 --> 00:19:50,366 around Postgres feels like it really changed in that timeframe as well. 271 00:19:50,366 --> 00:19:59,996 It wasn't just that these numbers went up, it also felt like there was almost a, not quite a backlash against the NoSQL movement, but I think people real, 272 00:19:59,996 --> 00:20:08,186 like there was like a phase that people went through where they, where the words big data got thrown around a lot and no sequel gained a lot of momentum. 273 00:20:08,336 --> 00:20:08,666 Mongo 274 00:20:08,666 --> 00:20:10,706 Nikolay: no sequel is interesting. 275 00:20:10,706 --> 00:20:11,726 Mongo is interesting. 276 00:20:11,786 --> 00:20:20,414 Uh, some people think that json is the main reason why Paulus is winning hearts of developers, and I think it's a 277 00:20:20,414 --> 00:20:25,064 very important contributor, but, uh, maybe not the main one. 278 00:20:25,524 --> 00:20:36,056 we chatted a little bit before, before we started this episode, and it looks like if, if you check, Jason appeared in Jason B appeared in 9.4. 279 00:20:36,373 --> 00:20:48,198 But if you check when the spike, of popularity happened, according to db engines.com, data, it happened when RDS released Postgres version, right? 280 00:20:48,612 --> 00:20:49,212 Michael: Yes. 281 00:20:49,542 --> 00:20:53,142 Now, I, I think this is a really interesting chicken and egg problem. 282 00:20:53,172 --> 00:20:57,732 Now did RDS have to add Postgres because it was gaining popularity? 283 00:20:58,092 --> 00:21:00,775 Did, Postgres popularity spike because Amazon. 284 00:21:00,775 --> 00:21:02,365 Nikolay: They chose it themselves. 285 00:21:02,575 --> 00:21:11,665 They chose it themselves for their, like, as a, they chose pogs and, and actually Aurora version of it, which is quite different. 286 00:21:12,115 --> 00:21:19,970 But they chose poss, not, not a, my, they chose Aaro PSGs, as the database for themselves, instead of Oracle. 287 00:21:19,970 --> 00:21:20,240 Right. 288 00:21:20,650 --> 00:21:22,960 Michael: Well, I think this is, yeah, this is a critical point. 289 00:21:22,960 --> 00:21:26,118 And yeah, you brought up Apple, in our previous chat as well. 290 00:21:26,123 --> 00:21:33,878 So big huge companies choosing to migrate to Postgres around that time definitely had a big effect. 291 00:21:34,132 --> 00:21:35,332 Jason b I agree. 292 00:21:35,332 --> 00:21:39,767 I think the biggest, impact Jason B had was, it gave people an answer. 293 00:21:39,977 --> 00:21:46,037 They didn't have to, If they needed a document store, they didn't have to have MongoDB in addition to progress. 294 00:21:46,277 --> 00:21:50,237 They could, they had a really good answer for the team that wanted to do some document storage. 295 00:21:50,237 --> 00:21:52,037 They could say, Well, you can put it in Postgres. 296 00:21:52,277 --> 00:21:54,497 And we have, you know, indexing support for it. 297 00:21:54,497 --> 00:21:58,527 And there's, it became, , an easy answer to anybody, on that front. 298 00:21:58,527 --> 00:21:59,547 But yeah, I agree with you. 299 00:21:59,547 --> 00:22:02,168 It's been important, but not the primary factor. 300 00:22:02,348 --> 00:22:04,785 Nikolay: Poss is good in reacting to challenges. 301 00:22:04,882 --> 00:22:13,552 you mentioned, uh, 1995 when sequel support was added Without sql, it wouldn't survive, I guess, in terms of very popular system. 302 00:22:13,972 --> 00:22:18,929 Then, there was some hype, like object databases, object relational databases. 303 00:22:18,934 --> 00:22:24,490 Paul also adopted some things, and it's still considered as object, object, relational databases and has some. 304 00:22:24,708 --> 00:22:27,528 features and behavior of objectional system. 305 00:22:27,831 --> 00:22:32,051 but then, semi-structured hype started and it started with XML actually. 306 00:22:32,501 --> 00:22:36,801 And, uh, some, Russian developers OV and, uh, Tega. 307 00:22:36,911 --> 00:22:39,791 They added support of age store in 2004. 308 00:22:39,881 --> 00:22:40,511 Can you imagine? 309 00:22:41,141 --> 00:22:51,439 Very long ago, I personally participated in bringing XML support to PSGs but both XML right now may be not that relevant because json is here 310 00:22:51,439 --> 00:22:56,085 already and it's standard defacto for unstructured or semistructured data. 311 00:22:56,375 --> 00:22:58,775 So PSGs is good in reacting. 312 00:22:58,972 --> 00:23:02,332 PSGs is doing hard work in interacting to challenges. 313 00:23:02,855 --> 00:23:03,935 Michael: I completely agree. 314 00:23:03,935 --> 00:23:08,922 And I think that leads me to my, I've got two more things I wanted to make sure we covered. 315 00:23:09,245 --> 00:23:16,165 One is that I think a lot of things we've discussed so far explain how Postgres was taking over from MyQ for the 316 00:23:16,165 --> 00:23:19,135 projects that we're considering a free open source database. 317 00:23:19,365 --> 00:23:25,889 But it doesn't yet explain how, uh, larger companies are choosing it instead of Oracle, instead of SQL Server. 318 00:23:26,129 --> 00:23:30,149 And that I would say it is another reacting. 319 00:23:30,479 --> 00:23:31,799 Yeah, but it's, it, it's a 320 00:23:31,799 --> 00:23:34,859 Nikolay: larger, companies don't care about why popularity. 321 00:23:34,859 --> 00:23:35,189 Right. 322 00:23:35,189 --> 00:23:39,119 They need, they, they care in terms of how many engineers they could hire. 323 00:23:39,359 --> 00:23:39,449 Yeah. 324 00:23:39,449 --> 00:23:44,879 But also sometimes how we see some companies prefer languages like lan, for example. 325 00:23:44,909 --> 00:23:45,739 What's Alan? 326 00:23:45,749 --> 00:23:53,009 And they still, they still keep or scholar some big companies who scale or lung, not very popular, but they choose it and 327 00:23:53,009 --> 00:23:58,489 they bring expertise inside so they care less about hype. 328 00:23:59,194 --> 00:23:59,644 Michael: Yes. 329 00:23:59,764 --> 00:24:08,147 But, things that did matter to them, I think Postgres started to ship around that same, so 2016, 9.6 parallel 330 00:24:08,147 --> 00:24:13,127 queries came in, 2017 we had partitioning logical replication 331 00:24:13,247 --> 00:24:16,817 Nikolay: but by 20 16, 17, uh, the game already that was done. 332 00:24:17,657 --> 00:24:19,297 Michael: So the game versus my, 333 00:24:19,357 --> 00:24:22,397 Nikolay: the game was done in 2014 when RDS 334 00:24:23,267 --> 00:24:24,887 Michael: versus MySQL yes. 335 00:24:24,887 --> 00:24:29,717 But I think versus Oracle and SQL Server, No, I think actually these large 336 00:24:29,717 --> 00:24:30,467 Nikolay: different opinion. 337 00:24:30,857 --> 00:24:31,787 Michael: Okay, interesting. 338 00:24:31,787 --> 00:24:39,557 Nikolay: I observed in Russia when I, I, I returned to Russia for like 10 months or so because of these issues and economy issues. 339 00:24:39,557 --> 00:24:46,754 After Crimea invasion, I return to Russia and I was bought and, decided to relaunch, uh, Russian user group. 340 00:24:47,177 --> 00:24:50,917 And, I asked, uh, which big company can host us? 341 00:24:51,007 --> 00:24:53,107 And my laar Andex both. 342 00:24:53,482 --> 00:25:00,832 Said yes, and by, I chose Yandex and, but GIX said, Can we also squeeze a couple of like 10 minutes of small lightning talk? 343 00:25:01,249 --> 00:25:02,309 I, I said, Of course. 344 00:25:02,339 --> 00:25:03,659 What, what, what's the topic? 345 00:25:03,659 --> 00:25:07,610 And they said how we migrated Yandex mail from Oracle to pogo. 346 00:25:08,370 --> 00:25:11,640 And then on this very first event, they gave this talk. 347 00:25:11,640 --> 00:25:12,570 It was amazing. 348 00:25:12,720 --> 00:25:19,530 And I remember in the first row also a guy from ATO was, and they said, ATO is running POGO as well. 349 00:25:19,787 --> 00:25:23,507 these companies are big and they don't care a lot about hype. 350 00:25:23,927 --> 00:25:26,287 And they also they are not like enterprise. 351 00:25:26,287 --> 00:25:27,187 They can count money. 352 00:25:27,187 --> 00:25:28,177 The startups, right? 353 00:25:28,177 --> 00:25:30,097 So they can count money and so on. 354 00:25:30,097 --> 00:25:39,427 And all day in 2014, they already, they also don't care about this replacement for political reasons, replacement of Oracle and SQL Server in Russia. 355 00:25:39,757 --> 00:25:41,497 They didn't care about it at all at that time. 356 00:25:41,707 --> 00:25:42,637 It was before actually. 357 00:25:43,297 --> 00:25:47,227 They chose it already and they chose it due to many reasons. 358 00:25:47,227 --> 00:25:48,796 And they were already. 359 00:25:49,044 --> 00:25:51,650 and they, made decision in 2014. 360 00:25:51,967 --> 00:25:52,207 Right. 361 00:25:52,450 --> 00:25:54,670 And that I was completely surprised. 362 00:25:55,100 --> 00:25:55,280 Michael: Yeah. 363 00:25:55,630 --> 00:25:58,490 So I, It's kept growing since then though. 364 00:25:58,490 --> 00:25:58,790 Right. 365 00:25:58,795 --> 00:26:07,610 So I think there have been an increasing number of use cases that can switch to Postgres from Oracle or SQL Server, or maybe companies that switch 366 00:26:07,610 --> 00:26:11,840 for new projects because the performance is on par for their workloads 367 00:26:11,840 --> 00:26:15,320 Nikolay: because they now have the same, the features, the performance. 368 00:26:15,320 --> 00:26:15,560 Right. 369 00:26:15,830 --> 00:26:15,950 Yeah. 370 00:26:15,950 --> 00:26:16,910 Let's agree with this. 371 00:26:16,910 --> 00:26:23,490 Like, uh, when GIX was ready to discuss, how they migrate from Oracle to Postgres, by the way, later then gave a 372 00:26:23,490 --> 00:26:27,150 talk at Pcon in Ottawa, and those were, that's very good. 373 00:26:27,150 --> 00:26:27,390 Talk. 374 00:26:27,620 --> 00:26:32,480 Like next year, maybe 2015, but by that time they're already working on it. 375 00:26:32,480 --> 00:26:32,870 A couple of. 376 00:26:33,350 --> 00:26:35,180 So they started in 2012. 377 00:26:35,450 --> 00:26:45,710 This brings us us back maybe to this acquisition of my sequel by Oracle, like indirect acquisition, like chain of acquisitions in 2010. 378 00:26:46,040 --> 00:26:56,077 And also, I consider Apple's decision to migrate internal things from my sequel to poss in 2011 is also some turning point, right? 379 00:26:56,077 --> 00:26:56,767 Some signal. 380 00:26:56,797 --> 00:26:59,137 Very strong signal, like some enterprise decided. 381 00:26:59,617 --> 00:27:06,075 But we also have a case when some big company, big startup, maybe one of the biggest ones, migrated from 382 00:27:06,151 --> 00:27:08,761 my SQL to Postgres and a couple of years back, right? 383 00:27:08,761 --> 00:27:15,211 You know what I'm, I'm talking about the other direct Uber case from mys and then back. 384 00:27:15,811 --> 00:27:18,491 Michael: Back was the one that got a lot of publicity, 385 00:27:18,511 --> 00:27:22,711 Nikolay: A lot of attention, yes, but of good criticism as well by the way. 386 00:27:23,371 --> 00:27:23,641 Michael: Yeah. 387 00:27:24,061 --> 00:27:28,021 I had one more point that I think it kind of takes us into the future as well. 388 00:27:28,021 --> 00:27:34,681 And that's that I think Postgres has the last few years, it's, it's released major improvements every year. 389 00:27:34,711 --> 00:27:37,441 Not, not, I'm not just talking about the fact they call them major releases. 390 00:27:37,741 --> 00:27:41,851 They have been big improvements every major version every year. 391 00:27:42,241 --> 00:27:45,355 I used to product manage, products for SQL seven, Oracle. 392 00:27:45,501 --> 00:27:48,181 And they didn't come out with major versions every year. 393 00:27:48,301 --> 00:27:48,871 They didn't. 394 00:27:48,961 --> 00:27:49,231 Right. 395 00:27:49,261 --> 00:27:51,541 It was every two or three years at a push. 396 00:27:51,541 --> 00:27:55,238 And even then, they didn't always include game changing features. 397 00:27:55,478 --> 00:28:02,378 So I think it's, it's very interesting, the speed of the, it's, uh, sometimes considered slow and I, I think that's unfair. 398 00:28:02,768 --> 00:28:09,218 But the speed of improvement with progress, I think is probably the, the highest or the relational databases at the moment. 399 00:28:09,478 --> 00:28:11,488 Nikolay: And also quite well structured as well. 400 00:28:11,698 --> 00:28:19,088 Like, I mean, like it's impossible to see some feature, like, you know, in my sequel They had check constraints, which did nothing. 401 00:28:19,208 --> 00:28:21,968 They didn't check anything until couple of years ago. 402 00:28:21,968 --> 00:28:23,576 They brought it in minor release. 403 00:28:24,006 --> 00:28:29,253 How come you, you, you add feature to minor release, like it's impossible in positives, right? 404 00:28:29,593 --> 00:28:29,773 Michael: Yeah. 405 00:28:29,773 --> 00:28:32,773 Nikolay: It's very, very well structured and principles are followed. 406 00:28:33,079 --> 00:28:33,439 It's good. 407 00:28:33,469 --> 00:28:33,589 Yeah. 408 00:28:33,739 --> 00:28:35,239 But there are many downsides as well. 409 00:28:35,239 --> 00:28:43,669 Like maybe we should talk about it a different time about the process of development and I will explain why I cannot participate in it. 410 00:28:44,033 --> 00:28:44,243 Michael: Yeah. 411 00:28:44,243 --> 00:28:46,643 That, that will be an fascinating discussion. 412 00:28:46,643 --> 00:28:47,003 Right. 413 00:28:47,003 --> 00:28:49,493 Is there anything else on the future that you wanted to talk about? 414 00:28:49,493 --> 00:28:50,663 Like why, Why would I 415 00:28:50,663 --> 00:28:53,783 Nikolay: Future, nearest future is bright, distant future. 416 00:28:53,939 --> 00:28:59,358 I have concerns and because like one of those concerns is this development process. 417 00:28:59,848 --> 00:29:05,021 It's outdated through email only the strongest, the biggest will survive. 418 00:29:05,708 --> 00:29:09,794 And I think it, it, it's not attractive to many, many people who could help. 419 00:29:10,504 --> 00:29:13,564 Michael: I strongly disagree, and I'm looking forward to discussing that with you. 420 00:29:13,864 --> 00:29:14,074 Nikolay: Yeah. 421 00:29:14,074 --> 00:29:18,724 Let's, let's, we don't have time to right now to dive into it, but let's dive in another time. 422 00:29:18,724 --> 00:29:26,088 But I, I have my own doubt, so I'm still like, I'm still big fan, like last 18 years. 423 00:29:26,148 --> 00:29:35,228 I, I choose pogs for every, I recommend choosing POGS for everyone who deals with LTP for analytical there are questions. 424 00:29:35,318 --> 00:29:40,118 But for OLTP workloads Postgres number one, default choice. 425 00:29:40,508 --> 00:29:46,298 We have cases when company grows to multi billion dollar evaluation. 426 00:29:46,298 --> 00:29:47,408 We have many such cases. 427 00:29:47,828 --> 00:29:49,118 We've just single data. 428 00:29:49,748 --> 00:29:51,488 Michael: Yeah, it's scale up. 429 00:29:51,698 --> 00:29:52,988 Nikolay: So yeah, so easy. 430 00:29:53,291 --> 00:29:57,091 And, Pogo has so many features that others don't have, so just choose it. 431 00:29:57,360 --> 00:30:08,575 and as the bottom line, let's, like, in my opinion, as with any startup, we, which became great company, as we listen to their founders, there is, 432 00:30:08,575 --> 00:30:13,405 by the way, good podcast called, uh, How I Built This by NPR, you know? 433 00:30:13,465 --> 00:30:13,675 Right, Yeah. 434 00:30:14,088 --> 00:30:24,015 like Starbucks story and many, many others like Instagram story, people usually answer why success combination of hard work and a lot of luck. 435 00:30:24,301 --> 00:30:33,465 So I consider this my, uh, unfortunate, timeline as a big luck of Postgres, but also we have a lot of hard work. 436 00:30:33,465 --> 00:30:34,995 We have like a hundred percent. 437 00:30:34,995 --> 00:30:35,531 So yeah. 438 00:30:35,855 --> 00:30:36,935 Michael: That's a really nice ending. 439 00:30:37,135 --> 00:30:39,185 Well, thank you everybody for joining us. 440 00:30:39,485 --> 00:30:42,545 Thank you again, Nikolay and, uh, see you next week. 441 00:30:43,078 --> 00:30:43,708 Nikolay: Thank you. 442 00:30:43,708 --> 00:30:49,618 Uh, don't, don't forget to follow, comment, like, and provide feedback and Twitter or anywhere else. 443 00:30:49,828 --> 00:30:56,731 We'll listen to you we, we already made several episodes based on feedback, so we are ready to continue this practice. 444 00:30:57,305 --> 00:30:57,545 Michael: Yep. 445 00:30:58,625 --> 00:30:59,375 Nikolay: Thank you so much. 446 00:30:59,435 --> 00:31:00,605 Thank you for our s Bye. 447 00:31:00,605 --> 00:31:00,785 Bye.