Hyper IO Configuration Item Reference

Configure the /opt/IO_Middleware/IO_Middleware_2.6.1_aarch64/conf/xxx.json file to set Hyper IO options, including the logging module, compression level, compression method, and middleware acceleration policy. /opt/IO_Middleware/ indicates the path to the Hyper IO installation package, and /xxx.json indicates the sample configuration file provided for individual applications. The configuration items are described in Table 1, (Optional) To use the file aggregation function, configure the global dataset group information., and (Optional) To enable out-of-order output for a single file, configure the dataset entity information..

**Table 1** Global dataset information (DATASET)
Item		Description	Type	Default Value
Global dataset general information (GENERAL)	LOG_FLOW	Prints FLOW module logs.	String ON: enable OFF: disable	OFF
	LOG_PROCESSOR	Prints PROCESSOR module logs.	String ON: enable OFF: disable	OFF
	LOG_BRIDGE	Prints BRIDGE module logs.	String ON: enable OFF: disable	OFF
	LOG_INTERCEPTOR	Prints logs of all INTERCEPTOR modules.	String ON: enable OFF: disable	OFF
	LOG_CONFIG	Prints CONFIG module logs.	String ON: enable OFF: disable	OFF
	TRANSCACHESIZE	Aggregates data into a single write operation when the total data volume reaches the cache size. The value range is [0, 128]. The unit is MB. The default value is 0, indicating that aggregation is not performed.	Integer	0
	COMPRESSMETHOD	Specifies the data compression method. The value range is [1, 2]. The default value is 1.	Integer 1: LZ4 compression algorithm 2: Bzip2 compression algorithm	1
	COMPRESSLEVEL	Specifies the data compression level. The value range is [0, 5]. The default value is 0, indicating that compression is not performed. A higher data compression level indicates a higher compression ratio.	Integer	0
	WITHOUTGROUP	Connects to Hyper IO without configuring group information.	String ON: enable OFF: disable	OFF
Dataset access pattern information (DATA_ACCESS_PATTERN)	FIXED	Indicates whether write operations of all entities remain unchanged between steps.	Boolean false true	false
	RANDOM	Indicates whether to randomize the particle order in particle applications.	Boolean false true	false
	WRITE_AGGREGATE_STEPS	Specifies the number of metadata steps to aggregate into a single write, mitigating frequent small I/O operations. The value range is [1, 64].	Integer	1
	WRITE_AGGREGATE_RANKS	Specifies the number of metadata ranks to aggregate into a single write. The value range is [1, 64].	Integer	1

(Optional) To use the file aggregation function, configure the global dataset group information.

**Table 2** Global dataset group information (GROUPS)
Name	Description	Type	Default Value
GROUPNAME	Mandatory. Name of the aggregated file in aggregation mode. The value must match the regular expression specified for FILEWHITELIST.	String of up to 64 characters	-
FILEWHITELIST	Optional. Regular expression to match files to be aggregated into the GROUPNAME file.	String of up to 64 characters	-

(Optional) To enable out-of-order output for a single file, configure the dataset entity information.

**Table 3** Dataset entity information (ENTITIES)
Name	Description	Type	Default Value
NAME	Mandatory. Name of the entity represented by a path.	String of up to 64 characters	-
SHUFFLE	Optional. Whether to randomize the particle order in particle applications.	Boolean false true The value is case-insensitive.	false

Parent topic: Usage Guide