-
|
maybe everyone wants to know this and make a choice |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 3 replies
-
|
We have done an benchmark to compare Alluxio, JindoFS and JuiceFS, the details is here: https://juicefs.com/blog/cn/posts/cache-system-of-hadoop-comparison/ Overall, they have similar performance when there is enough caching space (memory or SSD). But when the caching space is limited (<20% of dataset), JuiceFS is much faster than the other two. Because that the cached block is much smaller than the other two (4M < 128M), is better for random read (parquet and ORC) workload. |
Beta Was this translation helpful? Give feedback.
-
|
yes,I got it, Thanks for your reply for me. juicefs look great. Is the current open source all the code? What is the difference between version v0.9.1 and version JuiceFS 4.4.9 on the official website (https://juicefs.com/docs/zh/release.html)? from github relese can not get anything. |
Beta Was this translation helpful? Give feedback.
-
|
The open source community version has most of the source code in SaaS version 4.4.9, except the client that talk to our internal SaaS meta server. |
Beta Was this translation helpful? Give feedback.
-
|
The Java SDK used to big data workload will also be open sourced in community version. |
Beta Was this translation helpful? Give feedback.
We have done an benchmark to compare Alluxio, JindoFS and JuiceFS, the details is here: https://juicefs.com/blog/cn/posts/cache-system-of-hadoop-comparison/
Overall, they have similar performance when there is enough caching space (memory or SSD). But when the caching space is limited (<20% of dataset), JuiceFS is much faster than the other two. Because that the cached block is much smaller than the other two (4M < 128M), is better for random read (parquet and ORC) workload.