On Real-time Data Applications. Q&A with Elena Kolevska
” When I started Redis was used mostly for caching, but now you can do so much more, including things like vector similarity search, which would have been unimaginable a few years ago.“
Q1. You have presented at RedisDay SF a session titled “The Wild Ride into the New World of Redis Development”. Can you tell us in a nutshell what is it about?
A1: “The wild ride into the new world of Redis Development” was a guided tour of the different capabilities that Redis Stack offers through an example of a bike store. I wanted to simulate an environment that’s familiar to most developers out there (an e-commerce shop); a developer would be presented with a task like this and in their head they would start thinking right away
– “How am I going to model my data? Can I make it easily searchable? How am I going to keep statistics and do reports? Is there anything else that can improve the shopper’s experience?”
In my talk I tried to answer some of these questions.
Q2. What are the key challenges in building modern real-time data applications?
A2. Probably the first one that comes up on the list is latency. Users nowadays don’t want to wait more than a second or two for a page, or an app to load. So you have to look at your data layer first of all, because that’s where it all begins. All your other infrastructure tuning is for nothing if your database takes hundreds of milliseconds.
Another one is high availability. Components fail, and you need to make sure your system is prepared to handle hardware and software failure gracefully while staying online.
We can mention many more things here, like model flexibility, scalability and even developer experience, but any of these topics could easily be an interview in itself.
Q3. How Redis Stack can help developers rapidly build applications?
A3. Redis Stack gives you new ways to model your data, you can now use graphs where they make sense, time series, json documents and probabilistic data structures. Furthermore, it eliminates the need to create your own home-grown secondary indices or introducing a new technology in your stack specifically for that purpose. You can now index and search your data directly in Redis Stack, with the speed Redis is famous for.
Q4: How is Redis Enterprise different from Redis?
A4: Redis Enterprise is all about scalability, high availability and ease of operation. It is indeed built on Redis open source, and it uses its best features on the data layer, but the control layer has been completely reimagined to make your Redis deployment enterprise ready.
Q5: Can you please share your experiences of being a woman in tech and software development?
A5: Coming from an ex-socialist country, my experience in tech was much smoother, compared to the stories we hear from women around the world. A few people said, “are you REALLY going to choose that university”, but that was about it. Indeed, there have been some isolated situations here and there, but nothing that has hurt me or my career. I like to joke that, if anything, being a minority even kind of helped me because people notice you more when you give a talk or publish a blog post.
Q6: Anything else you wish to add?
A6: I’ve been a huge Redis geek for years and it’s so exciting to see how the product grows and develops. When I started Redis was used mostly for caching, but now you can do so much more, including things like vector similarity search, which would have been unimaginable a few years ago.
Elena Kolevska is a Software Engineer passionate about distributed system architecture, software design, mountain biking, electronic music and space exploration.
Sponsored by Redis.