TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Ask HN: Experience with StarRocks?

7 点作者 harovoy8 个月前
Hi hn community..<p>I want to know if any of you folks have tried StarRocks and how has been your experiences with it so far.<p>We started POC&#x27;s with it a few months back, and frankly saying, it has vastly exceeded our expectations. From deployment to usage, everything just works out of the box. The query performance has been super good too. We were even able to join 2 tables, one containing over 700M records and another over 25M records on small 16GB RAM, 10GB disk instance, all while selecting all, 500+ columns (some more than 100 of these are nested columns ARRAY&lt;STRUCT&gt;). Of course the query took a long time, but, that it even ran to completion without issues has been a surprise for us. When we gave it sufficient resources, it gave us better performance&#x2F;price ratio than even BigQuery(using slots to save costs).<p>How has yours experience been? Were you caught in any issues that were serious?

1 comment

winwang8 个月前
Not a user of StarRocks, but I&#x27;m curious about your experience with BigQuery and its price-to-performance (as well as its alternatives).<p>I&#x27;m building a product with the gist of &quot;BigQuery but cheaper&quot; (~30%-50% expected cost reduction) by utilizing GPU acceleration (think Nvidia RAPIDS), but I&#x27;m still in the market validation&#x2F;talking to prospects phase.<p>How much better was StarRocks in terms of cost efficiency? Does it scale to your needs?