Increase the GZip DeflaterOutputStream buffer size to increase performance (#36818)

scwhittle · web-flow · commit b843f4af3eb3 · 2025-11-18T10:53:33.000+01:00
diff --git a/sdks/java/core/src/main/java/org/apache/beam/sdk/io/Compression.java b/sdks/java/core/src/main/java/org/apache/beam/sdk/io/Compression.java
@@ -99,7 +99,9 @@ public ReadableByteChannel readDecompressed(ReadableByteChannel channel) throws
 
     @Override
     public WritableByteChannel writeCompressed(WritableByteChannel channel) throws IOException {
-      return Channels.newChannel(new GZIPOutputStream(Channels.newOutputStream(channel), true));
+      // Increase the default deflate output stream buffer size from 512 to 4096 for performance.
+      return Channels.newChannel(
+          new GZIPOutputStream(Channels.newOutputStream(channel), 4096, true));
     }
   },
 

Original file line number	Diff line number	Diff line change
`@@ -99,7 +99,9 @@ public ReadableByteChannel readDecompressed(ReadableByteChannel channel) throws`
`99`	`99`
`100`	`100`	`@Override`
`101`	`101`	`public WritableByteChannel writeCompressed(WritableByteChannel channel) throws IOException {`
`102`		`- return Channels.newChannel(new GZIPOutputStream(Channels.newOutputStream(channel), true));`
	`102`	`+ // Increase the default deflate output stream buffer size from 512 to 4096 for performance.`
	`103`	`+ return Channels.newChannel(`
	`104`	`+ new GZIPOutputStream(Channels.newOutputStream(channel), 4096, true));`
`103`	`105`	`}`
`104`	`106`	`},`
`105`	`107`