SEGY is a data format developed by the Society of Exploration Geophysicists (SEG) for storing geophysical data. DAOS-SEGY aims at implementing the SEGY data format natively on top of the DAOS storage model in order to accelerate data access and to provide in-storage computing capabilities to process data in place.
DAOS changes to be considered:
- sorting based on dkey
- range query support
The project is still in its early days and the following mapping is under consideration:
Object SEGROOT
# Root object storing global metadata and references to 2nd level objects (shot, offset & cmp)
D-key "File Headers"
A-key "sort-key" sorting type CMP, CDP, etc
A-key "grd-orgn" Grid origin (Latitude, Longitude)
A-key "Binary File Header", binary file header
A-key "Textual File Header", textual file header
A-key "1", 1st extended textual file header
...
A-key "n", nth extended textual file header
D-key "Sorting Types"
A-key "oid", objid of shot-gather object
A-key "oid", objid of cmp-gather object
...
A-key "oid", objid of offset-gather object
-------------------------------------------------------------------------------------------------------------------------
Object shotGather
# Object storing all shots
D-key "shot_id_1", first shot
A-key "hdr", shot metadata
A-key "oid", array of trace oids[trace_oid1, trace_oid2,...] belonging to this shot.
D-key "shot_id_2", second shot
A-key "hdr", shot metadata
A-key "oid", array of trace oids[trace_oid1, trace_oid2,...] belonging to this shot.
...
D-key "shot_id_k", kth shot
A-key "hdr", shot metadata
A-key "oid", array of trace oids[trace_oid1, trace_oid2,...] belonging to this shot.
-------------------------------------------------------------------------------------------------------------------------
Object cmpGather (inline vs crossline?)
# Additional index object providing ability to lookup trace in CMP order
D-key "cmp_1", first cmp value
A-key "cmp", CMP value
A-key "ntraces", number of traces having same CMP value.
/* cmp value for these traces is equal to cmp value of this akey */
A-key "oid", array of trace oids[trace_oid1, trace_oid2,...]
D-key "cmp_2", second cmp value
A-key "cmp", CMP value
A-key "ntraces", number of traces having same CMP value.
/* cmp value for these traces is equal to cmp value of this akey */
A-key "oid", array of trace oids[trace_oid1, trace_oid2,...]
...
D-key "cmp_k", kth cmp value
A-key "cmp", CMP value
A-key "ntraces", number of traces having same CMP value.
/* cmp value for these traces is equal to cmp value of this akey */
A-key "oid", array of trace oids[trace_oid1, trace_oid2,...]
-------------------------------------------------------------------------------------------------------------------------
Object offsetGather
# Additional index object providing ability to lookup trace in offset order
D-key "offset_1", first offset value
A-key "offset", offset value
A-key "ntraces", number of traces having same OFFSET value.
/* offset value for these traces is equal to offset value of this akey */
A-key "oid", array of trace oids[trace_oid1, trace_oid2,...]
D-key "offset_2", second offset value
A-key "offset", offset value
A-key "ntraces", number of traces having same OFFSET value.
/* offset value for these traces is equal to offset value of this akey */
A-key "oid", array of trace oids[trace_oid1, trace_oid2,...]
...
D-key "offset_k", kth offset value
A-key "offset", offset value
A-key "ntraces", number of traces having same OFFSET value.
/* offset value for these traces is equal to offset value of this akey */
A-key "oid", array of trace oids[trace_oid1, trace_oid2,...]
-----------------------------------------------------------------------------------------------------------------------------
Object Trace
#object storing all trace related info.
D-key "hdr", trace metadata
D-key "data[0:199]"
A-key "data", array storing data of first 200 sample.
...
D-key "data[tmax-200:tmax]"
A-key "data", array storing data of last 200 sample.