text stringlengths 0 2.68k |
|---|
**Jerod Santo:** It's hard once you explain it. |
**Adam Stacoviak:** Yeah. |
**Jerod Santo:** It's easy in the headline to actually be unpopular, in the poll quote... But then when you actually explain it -- except for that one. That one should have been a dead ringer. |
**Ashley Jeffs:** I don't know... I feel like people get it and they just want to ruin my good time. They don't want to give me the win. |
**Jerod Santo:** Yeah. Fair. I think people just see my opinions and they tend to agree with them, so that's why mine was popular... |
**Adam Stacoviak:** Same. |
**Ashley Jeffs:** The thing is, I definitely have opinions that are unpopular, but I think if I shared them in that segment, I'd be very dissatisfied with just a normal "Oh, I don't like hexagonal architecture." Like, if I just came with something bland, then I would walk away from that thinking "Yes, maybe it was an u... |
**Jerod Santo:** You did. It was big, it resonated, it got remixed... It ultimately failed, but you know, go big... |
**Ashley Jeffs:** It benefited society as a whole, but not me. |
**Jerod Santo:** That's right. |
**Adam Stacoviak:** Did this help you at all with your acquisition? |
**Jerod Santo:** \[laughs\] His is unpop? |
**Adam Stacoviak:** That's right. |
**Ashley Jeffs:** Yeah, actually it was part of the materials when we were negotiating it. They wanted the rights to my unpopular opinions going forward; because they can't buy my prior unpopular opinions. |
**Jerod Santo:** \[08:14\] Right. Those are yours. But now if you say one today, we're going to have to get permission. |
**Ashley Jeffs:** So don't get any unpopular -- only deal exclusively unpopular opinions from me on the show, please. |
**Jerod Santo:** Fair. |
**Adam Stacoviak:** Well, thankfully we've had Alex on the show before, Alex Gallego. I believe that's how you say his last name. Gallego? |
**Ashley Jeffs:** I just call him Alex. |
**Adam Stacoviak:** Just call him Alex, yeah. Had him on the podcast on Founders Talk a while back, talked through his journey... I don't know if you listened to that or not, Ash, but if you haven't, and you're a fan of Alex - you may be a fan of Alex. That was a pretty good journey to Redpanda. |
**Ashley Jeffs:** Yeah, it's very counter to mine, I think, because mine's very fumbly and clumsy, and I just kind of accidentally ended up where I am... Whereas his was very up and coming. |
**Adam Stacoviak:** You'd be surprised. |
**Ashley Jeffs:** Well, the Redpanda part, the kind of like beginnings of Vectorized, and stuff like that - that's very... You can tell he knew where he was going. Like, he had a goal where he was going to go for it. Whereas for me, I didn't really know what I wanted to do anyway. I was just doing stuff that I kind of ... |
**Jerod Santo:** That is different. |
**Ashley Jeffs:** Which tends to be a different experience. |
**Jerod Santo:** Both your names start with A... |
**Ashley Jeffs:** Yup. We've actually discussed that at length... |
**Jerod Santo:** Oh, really? I don't know how long you can actually discuss that particular bit... |
**Ashley Jeffs:** It comes up a lot. Anytime we're interacting with a third person, they usually go "Hang on a minute. Your names... Have you realized?" And then we have to go into it again. |
**Jerod Santo:** Yeah. I'm so predictable. Well, let's go way back before we get to Redpanda and Alex, because there's a lot more to your stumbling around before that, right? Maybe let's talk data streaming. I mean, Benthos is how you got there. Kafka, data streaming... I don't know. How'd you get into this whole mess? |
**Ashley Jeffs:** So before I started Benthos as a project, I was working at a company that was just doing stream processing. It wasn't really a word that they would have used at the time. They would have just called it -- I don't know, like event streaming, or something. Just data engineering. It wasn't really a big t... |
I was quite junior-y. I wasn't like a full junior engineer. I'd had a job before that, but I was still quite -- I was green. And that kind of -- I did consider myself to be just a junior learning the ropes. But then I eventually got to the point where I was kind of looking around thinking like "All the services here ki... |
\[11:58\] So things like delivery guarantees were kind of a weird concept. Like, we had the terms at least once, and things, but we didn't really have many services similar to Kafka. I suppose RabbitMQ was another popular one at the time, where you actually have proper acknowledgements. |
And the idea of the acknowledgement being at the source of data was kind of battling constantly against the idea of like having a high throughput, low latency streaming services. So we had like ZeroMQ all over the place, and everything had to be super-high throughput. But we also considered ourselves to be like a stron... |
**Jerod Santo:** Tradeoffs, right? Aren't there tradeoffs? |
**Ashley Jeffs:** Yeah, exactly. You can't just throw things over the wire without a response, some sort of acknowledgement, without having a compromise in delivery guarantees. But the idea of that actually being a formal understood thing... I still think people get confused about it all the time, in all sorts of place... |
And then about 2015, I had enough traumas from deploying all those services that I kind of figured it would be an interesting exercise to try and solve those things in a way where you've got a piece of software that will do all the -- I would have just called it plumbing tasks in stream processing land, because a strea... |
But what we were doing was just plumbing. You're just reading data from somewhere, changing it a bit, and then moving it along, passing it somewhere else... And trying to do that in an efficient way. But the main difficulty, in my opinion, wasn't getting good throughput and the sort of quality of the data coming throug... |
So I was kind of experimenting with that stuff as an open source project. I was also learning Go at the time, because I was using C++ up until then... And I was just messing around. I didn't have like a goal. I was just seeing if I could do it. |
**Adam Stacoviak:** That's awesome. |
**Ashley Jeffs:** Made it open source. It didn't really do much. For about two years I was working on it and it didn't really -- it read from Kafka or ZeroMQ, and vice versa. I had a few brokering patterns for mixing things up... And then you could have like a disk buffer in there as well if you wanted it... Which I di... |
**Jerod Santo:** Why did you do that then? |
**Ashley Jeffs:** So I had this idea of -- because we had this service at where I worked where they had a disk buffer, and it was basically just a bridge. We called it a bridge. And it would just bridge different protocols with a disk buffer. And the idea is the disk buffer was the delivery guarantee part. That's how y... |
\[16:10\] So it was a bit of a weird one... But you could do backfills and stuff; it could prevent back pressure up to an extent. Obviously not forever. But it did have some sort of value. And what I wanted to do was replace it, and then once I'd replaced it, you could just get rid of the buffer, because it was an opti... |
**Jerod Santo:** So it's all just like show and tell. |
**Ashley Jeffs:** Yeah, exactly. It's just run it and see if you explode. See if somebody calls you in a week's time saying "Where's my data?!" |
**Jerod Santo:** It's kind of like pulling the tablecloth out and all the stuff still stays on the table... |
**Ashley Jeffs:** Yeah, exactly. |
**Jerod Santo:** Very impressive when it works, but when you're bad at it, you just make a mess. |
**Ashley Jeffs:** Yeah. So I kind of just -- I put the responsibility on other people. So I'd whisper to the ops team, "Hey, you could just delete this bit of config", and then I'd walk away, and then it's not my fault. I just gave them the mechanism for doing it. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.