WEBVTT 00:00.000 --> 00:13.040 Was out of the scenario, and the companies were working parallel, but business change, of 00:13.040 --> 00:21.080 course, needs to start trying to make money, and as they move on, we finally release another 00:21.080 --> 00:23.600 real important. 00:23.600 --> 00:29.560 Me personally, I never saw my sequel for all versions, this is something in history, I couldn't 00:29.560 --> 00:35.880 even find in binaries or whatever, and as we can see, like really common things we 00:35.880 --> 00:42.640 do nowadays like unions, select star, something union, select something, it's not present 00:42.640 --> 00:50.320 by that time, so technology was evolving, and convey to remember, high degree was also evolving, 00:50.320 --> 00:57.440 so by 2000 and something we already started seeing the first hyper trading multi-core 00:57.440 --> 01:04.800 CPUs, right, so what we had there by that time as someone was saying my ison, my ison 01:04.800 --> 01:10.560 had is not a transactional engine, so if you're right that you do only that a base, you 01:10.560 --> 01:17.160 need to block, right, so anyone can write, it's table-level block, you can reading parallel, 01:17.160 --> 01:21.600 but as you see when you have more course, you're wasting hardware reserves, because you 01:21.600 --> 01:27.600 cannot scale, right, and that's when I know the B come to play, I already don't continue 01:27.600 --> 01:33.320 to evolve, and then we can see B3s, R3s, sub-credits and prepared statements, so it 01:33.320 --> 01:39.720 starts to get more familiar to what we see nowadays, although I didn't participate of this scenario 01:39.720 --> 01:48.320 yet, like my experience is started with my sequel 5 version, so as we move along, on 2005 01:48.320 --> 01:55.920 we see 5 old version, and Oracle buys I know base company, right, so if we get that 01:55.920 --> 02:02.720 marvel, what if, imagine if my sequel B, because my sequel tried to buy I know base, how 02:02.720 --> 02:07.920 different things would be, where we can't know how predict by that time they were trying 02:07.920 --> 02:19.280 to make money with it, and in terms of ecosystem and the word facts, we have the foundation 02:19.280 --> 02:28.360 of percono, whoever heard about percono here, okay, good month, so it was founded by 02:28.360 --> 02:34.040 Vadining Peter Zitesv, basically like percono provides a fork, completely open source as 02:34.040 --> 02:41.400 well, for my sequel, post, reason, and MongoDB, MongoDB is not open source, we need to 02:41.400 --> 02:46.920 inherit the license, right, so we cannot change, and things that we're going in my opinion 02:46.920 --> 02:54.120 to change, how we see cloud computing is the release of the first AWS becomes public, right, 02:54.120 --> 03:02.120 so in a few bottles, you can deploy a virtual machinesi to container and start working, 03:02.120 --> 03:07.480 so you can see how things were starting to get difficult for how my sequel was designed by 03:07.480 --> 03:13.000 that time, like we had my eyes, we had credit cashing, that was a good solution by that time, 03:13.000 --> 03:19.320 because it was a bottleneck, but it had a single CPU, so things were working fine, 03:19.320 --> 03:26.680 had multiple threads, but now you have several CPU cars trying to process and the old style 03:26.680 --> 03:33.800 of CPU of credit question was a point of contention, right, so they started deprecating things, 03:37.400 --> 03:46.520 and on 2008, another move, some by persistent acquires my sequel AB, and this is like where I 03:46.520 --> 03:55.640 start working with my sequel, so I had the opportunity to work with five one, my sequel, 03:55.800 --> 04:00.680 and we have some really cool release, like event schedule, row-based replication, 04:00.680 --> 04:06.600 things that were going to be improved, by the way, row-based replication is something 04:07.320 --> 04:12.040 it is the default nowadays, there is a good chance, I think it's already deprecated statement 04:12.040 --> 04:20.440 and will probably be removed in newer versions, and the point of that causes like the most 04:20.440 --> 04:30.280 hot discussions nowadays, it is 2010, or we will say 2009 where I saw in the Fred's presentation, 04:30.280 --> 04:37.160 but from I see the deal was closed on 2010, this is my reference, I can be wrong, but this is what 04:37.160 --> 04:44.680 posts with guys will say, but you guys are like, you have a company behind it, we are democratic, 04:44.680 --> 04:49.320 we are open-source, we are like whatever, and that's what causes a lot of food discussion. 04:50.920 --> 05:02.120 For the good and for the bed, now we have the modern ship Oracle, and at the same time, Oracle acquired 05:02.840 --> 05:10.520 some ecosystems, I'm not sure in the same day, but in the same year for sure, Montre Lise de Maria de Lise, 05:10.520 --> 05:16.360 right, there was a lot of concern about how Oracle would deal with my sequel, if they would stop 05:16.360 --> 05:23.080 providing features, like the project would end up dying because of obsolescence, or become like 05:24.120 --> 05:31.400 not open-source anymore. This is the first version that it is released on the water cruise 05:31.400 --> 05:39.640 hood, so 5, 5, it comes with them, it is a version that I work a lot, and I think through Montre Lise 05:39.640 --> 05:49.160 ago we got a customer trying to get out from there, so yeah, and as I said, it's really difficult, 05:49.160 --> 05:56.200 because I think that apart from the dump, we will have nothing that you can utilize, right, 05:56.200 --> 06:03.000 like it's really different. One more thing for those who work with Maria de B or planning to go 06:03.080 --> 06:08.680 or get out, Maria de B changed a lot, I think there was Maria de B days on Saturday, 06:09.400 --> 06:15.160 it continues to be a great database, but we can say for sure that it's not my sequel anymore, 06:15.160 --> 06:22.600 you cannot simply copy and paste the data, it's kind of like, I dare to say like moving from 06:22.600 --> 06:27.640 Oracle to post with right, apart from the table structures and everything you need to analyze, 06:28.120 --> 06:36.840 case by case, what you are going to do. On 2013, 5, 6 is released, I like this version as well, 06:37.560 --> 06:45.160 not much as I like 5, 7, but they are adding to the performance schema, performance schema 06:45.160 --> 06:51.400 provides some metrics and some views about what's going on inside my sequel. It has not a lot of 06:52.360 --> 06:58.840 metrics that you could see, 5, 7, exploded this universe when you go through it, you have like thousands, 06:58.840 --> 07:04.200 you know, to things of multiple monitor there with performance schema, but it was a good version, 07:04.200 --> 07:12.360 like I work it, I like it, it was quite stable, and another important fact here, again, 07:12.840 --> 07:23.400 we already had like a common place for cloud computing, Google was giving its try, Amazon as well, 07:23.400 --> 07:29.000 and then Kubernetes became a thing. At first we've known state full sets, one of the days like 07:29.000 --> 07:34.440 people are trying, we had I forgot the guide that did the presentation that they are running, 07:34.440 --> 07:42.040 Ubuntu, Canonico, they are running data bases on Kubernetes, it's also a very hot topic, 07:42.360 --> 07:46.600 if you ask something in years they will love it, others will say they will never do this, 07:49.160 --> 07:54.680 I also could repeat what I met as a device from Kubernetes. 07:55.480 --> 08:03.960 Yeah, well, so yeah, like I would say if you want to run in Kubernetes, my opinion is yeah, 08:03.960 --> 08:08.840 do it for a key way if you want to be very developed or something to test and everything, 08:08.920 --> 08:15.560 for a production maybe, like you don't know if you have 1,000, 2,000 databases, maybe you can think of it, 08:15.560 --> 08:23.080 but don't put like a 3 terabyte database on Kubernetes, like SSD process and everything, 08:23.480 --> 08:30.040 auto scale, like if you expand your auto scaling from 2 to 5, like it's 15 terabytes running 08:30.040 --> 08:40.600 or closer to build new nodes, so I don't like it. One more thing, five seven is released with 08:40.600 --> 08:46.280 JSON data type, I think it is cool because when I started putting all the facts together, 08:47.000 --> 08:53.080 postgres was also working on their JSON structure and the thing is by that time when you look at 08:53.080 --> 08:58.920 the Beijing's ranking, MongoDB was climbing really fast, so people were saying okay we need our 08:59.000 --> 09:05.400 JSON or document start solution because we have some, you know, someone trying to catch our market 09:05.400 --> 09:12.040 and you see like how things were moving that direction. At same time proxy SQL is released, 09:12.040 --> 09:21.000 a proxy SQL from Rene canal, basically proxy for a who knows proxy SQL, okay, 09:21.640 --> 09:29.800 it's a layer 7 proxy, so it is intelligent in terms of it can, it speaks the my SQL language, 09:29.800 --> 09:37.160 so it understands the query, it can rewrite, it can route, if I'm not mistaken my SQL router is 09:37.160 --> 09:48.680 layer 4, like 3, okay so yeah proxy SQL has its benefits as well, another open search tool that's 09:48.760 --> 09:59.560 part of the ecosystem. On 2016, by that time like group replication was released, 09:59.560 --> 10:06.760 it was going to be the future of final DB cluster. I had the chance to test a group replication 10:06.760 --> 10:14.200 from my SQL 57 and it was really not a mature product, it had several performance issues, 10:15.160 --> 10:22.440 it was really easy to make it crash and we see by that time what I see why they released, 10:23.080 --> 10:29.080 Galera replication Galera cluster was something that was really on the moment, it was really stable, 10:29.640 --> 10:36.280 so I think they were trying to get this piece of market with the cluster, but even nowadays Galera 10:36.360 --> 10:43.240 cluster remains, I strong solution per column makes the fork for it, with per column extra DB cluster, 10:43.240 --> 10:50.120 MariaDB as well, and even if you go, at least when I spoke with the Oracle guys in Brazil, 10:50.120 --> 10:56.440 they don't mention a lot, the group publication 57 because I think it's the ugly baby they had 10:56.440 --> 11:00.760 and they don't want to mention, like things got better on my SQL 8. 11:01.720 --> 11:10.680 As we move, as I said, group publications released and then it comes my SQL 8, right? 11:11.960 --> 11:18.120 It is a version that I like as well because I think they brought really good improvements, for example, 11:19.320 --> 11:29.240 you have now a complete redesign of the hydrologg, you can scale it, you can change the size 11:29.320 --> 11:35.000 online, right? So you don't need to restart that base, prior to that, you will need to do it, 11:37.160 --> 11:41.960 you have atomic DDLs, I think it's really important because if you're running, I don't know, 11:41.960 --> 11:48.440 you're automation, automated deploys and you're doing create drops, adding index, 11:48.440 --> 11:52.680 you don't know if everything worked, right? Because like it's not atomic, 11:53.640 --> 11:58.760 by this, what do you mean? You can do a begin, run all your sets of DDLs and commit, 11:58.760 --> 12:04.600 or you can do all or nothing, like it gives you a market visibility, dictionary became transactional, 12:04.600 --> 12:10.920 it's waiting side, they I know the beat, right? And also they brought the instant algorithm, 12:11.480 --> 12:16.760 by instant you can do some DDLs where like you only change the metadata, 12:17.480 --> 12:21.960 so it's not necessary like to rebuild the table or something like that, 12:23.640 --> 12:31.560 it was a bit problematic at first, but then it's got better, right? Another big thing that came, 12:32.200 --> 12:39.160 Shattipiti was released on 2022. Probably what you see most of falls then, like as this presentation 12:39.160 --> 12:45.080 has some AI, something everything you see at the touch of AI, we are all talking about 12:45.960 --> 12:52.600 of it, and if either lost a few billions last week, we do it through deep seek, 12:53.320 --> 13:01.880 but still it's a hot stop, it's coming, and this is also influencing Oracle, right? The same 13:01.880 --> 13:11.880 next year, they release hit wave on their public, on their Oracle Cloud, hit wave has some 13:11.960 --> 13:17.080 genuine AI features on that, I personally never tested, I need to take a look, 13:18.280 --> 13:25.240 in terms of concurrence, possibly 16 was released, so also possibly bringing some vector search, 13:25.240 --> 13:38.200 trying to bring some AI as well, right? And one more thing, like on 2004, we get eight 13:38.280 --> 13:45.960 four release, right? As Simon was saying, just to go a bit more deeper on that front, I don't remember, 13:45.960 --> 13:53.880 when in year Oracle changed the release model, so now we have two types of release, stable release, 13:53.880 --> 14:01.480 and innovation release as he was saying, so what what's happening with my six o' eight is you 14:01.480 --> 14:06.840 were working happily, the next version you had a completely new hydrologue with new parameters, 14:06.840 --> 14:11.480 and like how your automation broke now, so we started complaining in Oracle, 14:12.280 --> 14:17.800 listened to the community, changed it, so my six o' eight and eight four are the current 14:19.800 --> 14:25.640 stable release, this means that only bug fixes are coming, you have a more predictable behavior, 14:25.640 --> 14:34.440 and like my 12, nine, and nine one, I think that the moment nine two are innovation release, 14:35.400 --> 14:40.520 deprecated features can get out of the code without one new features can come, right? 14:41.400 --> 14:47.640 I think it is good for you to anticipate to test, of course, what in production, unless you have a very 14:49.560 --> 14:56.840 fasting that can adapt really quickly and evolve, I would say to remain on stable releases and 14:56.840 --> 15:03.400 wait for the next one to come. I think it was a miss of the opportunity because from what I see, 15:03.400 --> 15:09.000 my sequel version 10 will come as innovation, and I think 10 is around number, they should release 15:09.000 --> 15:18.840 as a stable, but I don't know what marketing we're thinking on that, and then here we are at 15:18.840 --> 15:28.040 1,000, 2,000, 25, talking about January and my sequel, personally like I like where we are heading, 15:29.080 --> 15:37.480 but what about the future? So I got the, some testimonials from people that are 15:37.480 --> 15:45.160 influenced in the community, so I got one bed, one okay, one optimistic, we all know like Peter likes 15:45.160 --> 15:51.800 to rent a bed, so he has concerns on how the ecosystem of my sequel is going, because Oracle is putting 15:51.800 --> 15:58.360 a lot of effort in the, in their cloud, right? The cool features are going to see I first, 15:59.160 --> 16:05.480 as I mentioned hypergraph exists, but you need to, you know, to enable the feature by yourself, 16:05.480 --> 16:12.840 so Peter is a bit concerned with that MariaDB brought, Mike Kalan did some tests, 16:13.400 --> 16:21.560 he's quite happy with a vector, it's not vector pudding, it's not sure it's vector ranging, 16:21.560 --> 16:27.000 but from MariaDB is really fast, so this is the testimony from Peter, right? 16:29.480 --> 16:36.360 I also had the opportunity to get from Sinisa, he's a developer of Oracle, he has 26 years of 16:36.360 --> 16:46.200 Oracle, he saw everything in life, he's the employee number two of my sequel B, and he likes it 16:46.200 --> 16:53.400 nowadays, because of a plot of farms, like hit wave, so one, one side people don't like it, 16:53.800 --> 16:59.080 but if you really are not attached to license, you just need your product to run and make it profitable, 16:59.080 --> 17:04.280 hit wave, rendez-be, you have like several things going on in the ecosystem, right? 17:05.080 --> 17:12.840 Unless from Monciola Fred, which is here, also good things that existed in the past, 17:12.840 --> 17:18.680 like I think it's the community feeling, I think falls in, because conference is quite over for now, 17:20.120 --> 17:25.480 preconalive doesn't exist anymore, so one of his best memories was from preconalive Santa Clara, 17:25.480 --> 17:30.120 for me, that I came later, for me is falls down, I like to come here for the Belgian days, 17:30.840 --> 17:36.040 share the two days of expensive, like sessions, then community dinner, I think it's really nice, 17:37.080 --> 17:46.120 and well he's not an NDA, so he can't say anything, but I like the positivity, so I hope that good 17:46.120 --> 17:53.800 things are coming, it's just a matter of time, so we wait and see what I would like to see, 17:54.520 --> 17:58.360 maybe in my sequel, it's some performance improvements, like, 17:58.360 --> 18:03.960 postgres was elected, the database of the year in the last three years, consecutively, 18:04.840 --> 18:10.280 my sequel was elected back in the past, so we have a huge base of our community, 18:10.840 --> 18:16.600 but I think that we can, we must catch up with some performance improvements, 18:16.680 --> 18:25.160 we can compare it to polls, you can, okay, cool, so you hear about it, 18:26.280 --> 18:31.480 so I think that's it, I think it was more relaxing session, rather than we will continue with 18:31.480 --> 18:36.200 the technical sessions now later, do you guys have any questions, 18:38.920 --> 18:44.680 or if not, I would like to say thank you for being the 30 years addition, it's really special to me, 18:45.320 --> 18:48.520 I love it, and I hope to see you next time, 18:56.120 --> 19:05.720 thank you, you can add a little bit of my strategy, yes, yes, yes, 19:14.680 --> 19:15.160 thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you 19:44.680 --> 20:14.680 thank you, thank you, thank you, thank you, thank you, thank you, thank you very well thank you, thank you, please 20:44.680 --> 20:57.680 That's all I've got to do, that's all I've got to do, that's all I've got to do, that's all I've got to do, that's all I've got to do, that's all I've got to do.