Bluesky | Twitter | LinkedIn | YouTube | Instagram
This article is also available on YouTube!
Have you ever heard of Jean-Antoine Nollet? Back in the 18th century, Nollet carried out an experiment where he lined up 200 monks, each connected hand-to-hand with iron wires, forming a continuous chain over a mile (1.6 km) long. Once everything was set up, he connected a primitive electrical battery to the line, delivering a powerful electric shock to all of them simultaneously.
Now, Nollet wasn’t just zapping monks for kicks. His experiment had a serious purpose: to study the properties of electricity and see how far and how fast it could travel along a wire. This was groundbreaking at a time when sending a message 100 miles took nearly a day by horseback. Nollet’s work hinted at something revolutionary — t*he potential for electricity to be used for communication*.
Fast forward to the 19th century, and the telegraph brought this idea to life. Suddenly, messages that used to take days could travel in minutes. Samuel Morse and other inventors transformed Nollet’s findings into a world-changing technology. The telegraph became the 19th-century equivalent of the internet, connecting people in ways no one had imagined before.
As Tom Standage describes in The Victorian Internet, t*he telegraph was so fast it scared some people*. Critics even argued it was “too fast for the truth.” It sounds funny now, doesn’t it?
Today, we see the internet as almost instantaneous, but back then, the telegraph felt like a leap into hyperspeed.
That said, even with our modern tech, we sometimes still think the internet is slow. To Nollet, the speed we’ve reached would have been incomprehensible, but we know there are limits. For example, even if data could travel at the speed of light in a vacuum, it would still take about 56.7 milliseconds to get from London to Sydney. That’s just physics — it can’t get any faster.
But speed isn’t just about how fast data travels; it’s also about how quickly it gets processed.
With applications like real-time gaming, video streaming, and AI-powered services, every millisecond matters. That’s where Redis comes in.
Redis is an in-memory database designed for speed. Unlike traditional databases that rely on disks, Redis keeps everything in RAM, giving you access times measured in microseconds. This makes it ideal for real-time analytics, online gaming, and AI workloads where responsiveness is critical.
And guess what? Redis just got even faster with Redis 8.
Redis 8: Faster Than Ever
The latest milestone, Redis 8.0 M02, brings significant latency reductions across widely-used commands, such as up to a 36% reduction in latency for ZADD, 28% for SMEMBERS, and 10% for HGETALL compared to Redis 7.2.5. Over 70% of Redis users will experience noticeably faster responses with these improvements.
Redis, like the telegraph once was, is revolutionizing our expectations of speed. It ensures that not only does data reach its destination quickly, but that it’s immediately available for processing and analysis. In a world where even a 100-millisecond delay can impact user experience, Redis plays a crucial role in minimizing the lag.
Scaling Like Never Before
Redis 8 isn’t just faster, it’s more scalable too. It brings features that were previously only available in Redis Cloud and Redis Software, like horizontal and vertical scaling for the Redis Query Engine.
With horizontal scaling, you can handle much larger datasets by clustering databases, which boosts read and write throughput. Vertical scaling adds processing power, delivering up to 16x more throughput.
Benchmarking Redis 8: Breaking Records
To showcase its improvements, Redis partnered with Intel to test its performance with one billion 768-dimensional vector embeddings. The results? Redis handled up to 66,000 vector insertions per second with indexing for 95% precision and up to 160,000 insertions per second for lower precision indexing.
Even with high-precision queries, Redis delivered a median latency of 200 milliseconds for a 90% precision rate when searching the top 100 nearest neighbors. And by tweaking HNSW (Hierarchical Navigable Small World) parameters, you can fine-tune Redis to balance speed and accuracy for your specific use case.
See more of the benchmarks in the official Redis Blog.
Try Redis 8 Today
Redis 8.0 M02 is available now, and you can experience its speed and scalability for yourself. Whether you’re looking for better latency, scalable query engines, or support for billion-scale vector search workloads, Redis 8 is ready to deliver.
Start experimenting today by downloading an Alpine or Debian Docker image from Redis Docker Hub. See what Redis 8 can do for your real-time applications!
Top comments (0)