00:00.0
00:02.0
so 2017 we're making this big push into the Java market,
00:02.1
00:04.8
all the big wigs are like "Risitas, cmon, let's go
00:04.9
00:07.4
Java is big enterprise money"
00:07.5
00:09.0
But every Java app is like
00:09.1
00:10.9
"no way, I'm too fat for this container".
00:11.0
00:16.3
it just falls over dead, the OOM killer won't let it live. They're too fat for our container profiles.
00:17.7
00:21.6
all the container profiles were put together in 2012 for Drupal applications!
00:21.7
00:24.3
of course they don't fit!
00:29.2
00:31.0
we sold them anyway!
00:31.1
00:34.5
can you imagine an enterprise java app in a container with 76M of RAM?!
00:37.8
00:40.1
we tried tho!
00:41.2
00:47.4
these devs and our solutions team would spend days and days w the config files
00:47.5
00:51.3
trying to figure out how to make these apps start
00:51.4
00:56.2
and eventually they'd actually figure it out and get it to start
00:57.9
00:59.3
and the thing would immediately fall over
00:59.4
01:01.7
not a great look, as I'm sure you can imagine!
01:04.1
01:06.6
"of course we're enterprise ready!!"
01:06.7
01:08.7
ready for your money baby!
01:14.3
01:16.5
we had absolutely no java experience anywhere in the company
01:16.6
01:20.9
and this poor devrel guy who came from drupal trying to figure out gradle and maven
01:21.0
01:24.1
we showed him this tar pit
01:24.2
01:29.0
and he put on his swim trunks and went right on in
01:36.6
01:40.4
he showed up with a briefcase full of open source community credentials and left years later a hollowed out shell
01:41.1
01:44.1
so fast forward 5 years
01:44.2
01:46.9
and we've been trying to figure out a way to solve this problem ever since
01:47.0
01:49.6
in between we built this 1 knob that lets our customers tune allocations a bit more
01:50.0
01:53.8
We spent 2-3 years in product development trying to figure out how to sell it without confusing the hell out of our customers.
01:54.8
01:56.0
we never documented it anywhere.
01:56.1
01:57.9
but it works great
01:58.0
02:02.5
and so eventually the day comes that we decide to just sell some high-mem plans
02:09.9
02:10.9
so one of the support engineers
02:11.0
02:13.0
one of the support
02:14.6
02:17.0
one of the support engineers realizes that we rolled the sellable to our customers
02:17.1
02:21.0
but we didn't update the allocator to know that it could allocate more memory to these projects!
02:21.4
02:28.0
our customers are paying us more money for the same thing that they still can't see into to know what's going on.
02:29.0
02:29.9
people are paying for this?
02:30.0
02:32.6
absolutely! who doesn't want to pay more money for more RAM? lol
02:36.9
02:40.0
Everybody knows there are advantages and disadvantages
02:40.1
02:42.4
for customers of black boxes
02:50.0
02:55.0
but praise god, this one finally puts more dinero in our pocket
02:56.5
02:58.2
I'm curious what impact any of this has on your margins
02:58.3
03:03.5
well, customers paying more for the same thing means margins go up
03:03.7
03:04.7
It's brilliant.
03:05.5
03:09.0
as long as they don't find out.
03:09.1
03:11.5
my only regret is that I don't have more equity in this money machine
03:16.4
03:20.1
ask me about flex plans
03:23.9
03:26.5
or "high availability on the grid"
03:30.9
03:36.4
I tell you man, good times. it's the wild west out there
