Hello Timeplus Community!
As 2023 draws to a close, I want to send a quick note to thank you for your support over the past year. It’s been an amazing year for us at Timeplus. It’s so exciting to see the clear impact of our work in making better streaming analytics accessible to all developers. We might not be as big as Microsoft, but we are mighty in output. As the famous quote by Margaret Mead goes, “Never doubt that a small group of thoughtful committed individuals can change the world. In fact, it's the only thing that ever has." In that spirit, here’s a review of our Timeplus 2023 highlights:
Community
Our biggest news of the year was to open source our core engine Proton in September 2023.
This was an important step in our journey to serving the developer community, and feedback has been super positive. Developers love using Proton, as a single binary or as a Docker container, to easily connect with Apache Kafka or Redpanda clusters. Since October 2023, developers have downloaded Proton over 5,200 times. I am thrilled to see Proton deployments from containers as small as 2 CPU and 3GB memory, to larger servers such as 96 CPU and 700 GB memory. Proton really can be deployed anywhere!
In addition, we are thrilled to be contributing back to the ClickHouse community. We are huge admirers of the columnar database technology that the ClickHouse team has built. Our historical analytics leverages their amazing work. In September 2023, we started the process of committing over 7,300 lines of code to the ClickHouse repository. Our hope is that this contribution will add cutting-edge streaming functionality to the ClickHouse community.
Product
Our fundamental goal at Timeplus is to make really hard things easy for developers. This guiding principle determined a lot of our product focus in 2023. This was A LOT of work. It’s great to see developers are excited as we are about these improvements. I sincerely think this will change how developers everywhere use analytics.
Some of our accomplishments I am particularly proud of:
Building a lightweight single binary. Having a light footprint, developers can deploy streaming processing analytics anywhere: from the cloud to the edge. No JVM, no dependencies.
Building unified SQL streaming analytics. We are not the first team to build unified (batch + streaming) analytics. But, we’ve made some real innovations in creating a platform that is deployable anywhere (single binary) and easy to use (query historical + streaming from a single SQL pane of glass). We also provide streaming windowing, joins and ASOF joins, historical unification with correlation, backfill, and replay.
Building a highly performant engine. Performance matters, especially in industries where latency is expensive, like Financial Services. We’ve accomplished impressive figures in production: single-digit millisecond end-to-end latency, 10M+ EPS throughput, 10M+ high cardinality aggregations.
For a complete list of all product improvements, you can refer to our release notes.
Use Cases
Financial Services/FinTech
Timeplus is a natural fit for latency-sensitive scenarios, such as post-trading analysis. As we gained traction in serving FS clients, we’ve focused on rolling out additional purpose-built features and enhancements to serve their specific needs. Those enhancements in Timeplus Cloud and Proton include adding Open-High-Low-Close (OHLC) charts; adding WebSocket as a first-class data source; adding Complex Event Processing (CEP) with User-Defined Function (UDF); and adding ASOF JOIN to merge two or more data streams with the same ID but closest timestamp with low latency (5-70 milliseconds) high throughput (20K EPS) and minimum memory storage (~4GB memory).
Edge Processing
We’ve also seen a lot of customer interest in deploying Timeplus in edge processing scenarios. Given that Proton uniquely provides unified analytics in a small, single binary, that’s not surprising. Developers can set up lightweight streaming processing from edge devices or a micro-datacenter. Proton can be deployed as a single binary for x86 or ARM architecture, without JVM or other component dependencies. We started seeing our users choosing Proton to process local data and files with super-low latency, and consolidate the high-value insights in Timeplus Platform in the Cloud or BYOC environment. Key use cases include distributed cyber security detection and IoT/manufacturing monitoring.
Happy Holidays!
Starting Timeplus has been one of the most meaningful things in my life. Above all, I am so grateful for the hard work and achievements of our team. It feels like we are building something special. Looking forward to 2024, we will continue to build our team and have lots of exciting features in the pipeline (more details to come soon). But for now, I hope everyone can take some time off over the holidays and enjoy time with family and loved ones. We all have a lot to be grateful for.
Ting Wang
Founder and CEO, Timeplus