Percona’s version of MySQL Conference this year was awesome, and there were some great keynotes there, I’ll high-light two of them.
One was called “Future Perfect: The Road Ahead for MySQL” and had a vendor panel of “Industry leaders from HP, Amazon Web Services, McAfee, Clustrix, and Akiban discuss the future of MySQL”.
It went like this:
BaronTheModerator: “so, what is the future for MySQL?”
HP Cloud: “Clouds!”
Amazon Web Services: “Clouds and DynamoDB/RDS!”
McAffee: “More Security!”
Clustrix: “More Scaling!”
Akiban: “More document stores!”
BaronTheModerator, “what do your customers ask for?”
HP Cloud: “Clouds!”
Amazon Web Services: “Clouds and DynamoDB/RDS!”
…
After an hour of that I already knew a lot about road ahead for MySQL, much more than after pragmatic “What Comes Next”, which covered boring things like optimizing for future hardware, compression, bla bla.
Probably the most significant part of keynotes was Brian Aker’s disclosure of his impact on MySQL ecosystem in his “The New MySQL Cloud Ecosystem”. Actually, people should know that Brian actually coined the term “MySQL Ecosystem”. Brian remembered his days as MySQL’s Director of Architecture (05:16 in the video):
So, what I think we’ve now been entering for last couple years really is actually the cloud era. And we have glued some pieces of things… It’s funny, we’re talking about… you know… I’m looking at MySQL 5.5 and look at 5.6 and I’m like yep, I remember, thats the roadmaps I wrote, ha, glad to know, I’m curious what they gonna do after 5.6. Didn’t leave many.. eh.. Cliff Notes to what goes behi… after that.
This reveals that MySQL Community and Oracle customer base are actually locked in into a stagnant software product with no future. I think you should be worried. Heh.
domas mituzas 
Err sure… and all the forks die with it I presume?
I’m pretty sure percona live is proof enough that the fatches can survive without oracle/mysql.
I’d wager Oracle’s future for MySQL involves making a larger divide between enterprise and ‘free’/entry level. The plugins are just the tip of the iceberg IMHO.
We all know that future belongs to NoSQL and BigData in Clould.
So unless MySQL integrates NoSQL access to BigData, it is doomed.
Re-post:
My guess is Oracle will continue to polarize enterprise MySQL from entry-level/’free’. The plugins are the tip of the iceberg.
NoSQL has it’s share of maturing to do, and survival of the fittest is still weeding out the weakest – once that happens…
Yeah, that performance stuff is definitely pretty boring. Nothing to see there, move on. Speaking of which we are moving to the cloud where high performance doesn’t matter any way. If it did, nobody would use EBS. :)
I think MySQL will continue to live in it’s cloud form, but the associated DBA’s will either have to work for the big cloud providers or the few who are too big for someone else’s cloud but they will diverge the forks they already have further and further. All the little shops that use MySQL will move it to those cloud providers for one reason or another. MySQL’s proprietary and specialized descendents will live a long time, though.
To clarify, I’m not saying this is necessarily bad, just the only way I really see it panning out.
Hold on…
(1) What is clear from the panel is that they all are drinking and prophesying their own cool-aid and value-adds. This sounds just advertisement and marketing.
(2) When Robert Hodges (Continuent) says performance is to be glossed over… similar to the panel, it shows that he is prophesying replication and cloud are the future. Its interesting to note that version 5.6 is going to make large portions of Tungsten obsolete (or not needed).
(3) Large portions of Percona (larger log files, checksums, flushing related stability) are also going to be made obsolete by v5.6. However this is the same thing that happened in v5.5 and Percona still differentiated. In some ways Percona is acting like the “voice of customers” that MySQL product management is not hearing, but MySQL engineering is (thanks to Percona).
(4) Some lead that MariaDB has (due to better query plans and optimization) will be reduced again by v5.6.
SUMMARY:
Oracle’s announcement of the 5.6 features in the same week as PerconaLive conference, while being totally absent from the conference… and announcing a separate MySQL conference in San Francisco shows that it is worried about MySQL forks and fragmentation… but also that it is serious about advancements in MySQL and InnoDB. So why is it doomed again ?
Note that the panel was a paid sponsorship opportunity. For my part, I found it a real challenge (and a fun and interesting one) to have several very different participants with non-overlapping agendas, and try to weave it into a conversation that the audience would enjoy and actually benefit from, while making sure the sponsors got what they paid for.
Baron, I fully understand that :-) hence my lighthearted recital of the talk! It is just a bit funny how non-overlapping they were.
Yes. I was actually responding to TREBOR SEGDOH’s assertion that it’s advertising and marketing. Indeed, it is!
Who is this TREBOR SEGDOH guy anyway? He’s using my name backwards!
one could say that you are doing exactly the same ;-)