[DFDL-WG] Fwd: Daffodil Infoset File Size Comparison between xml, exi (schema un-aware) and exisa (schema aware) Parsing

[Mike Beckerle]

I thought this might be of interest to the DFDL WG.

---------- Forwarded message ---------
From: Thompson, Dave <dthompson at ...>
Date: Wed, Oct 19, 2022 at 8:41 AM
Subject: Daffodil Infoset File Size Comparison between xml, exi (schema
un-aware) and exisa (schema aware) Parsing
To: users at ...

Just a heads up to daffodil users for the upcoming Daffodil v3.4.0 release,
which will include EXI support.

The attached spreadsheet shows the comparison of Daffodil infoset file size
when parsing to xml, exi (schema un-aware) and exisa (schema aware) for
some of the Daffodil supported schemas.

The comparison shows a significant decrease in infoset size between xml and
exi (schema un-aware) infoset files and even a greater reduction with exisa
(schema aware) infoset files. In most cases of the exisa (schema aware)
infoset files, the infoset file is only slightly larger than the original
source file.

*Source Data* *Daffodil Infoset File Sizes (bytes)*
*Data Format*
*Source File (bytes)* *XML*
*EXI **(Schema Un-aware)*
*EXISA **(Schema aware)*
1 bmp 4,264,316 8,529,476 8,528,926 4,264,305
21,600,054 43,200,952 43,200,406 21,600,044
2 cef 4,180 17,880 2,664 2,462
3 gif 950,734 2,353,949 1,916,689 966,667
4 iCal 5,370 47,345 4,606 4,019
5 Jpeg 1,030,473 2,076,278 2,062,048 1,030,817
15,477 47,023 32,181 15,821
6 Jpeg2000 55,399,819 124,654,407 112,527,466 56,438,548
16,298 41,260 33,989 16,590
7 nato-stanag-5516
(FOUO, not public) 1,952 123,818 5,208 2,366
62,464 3,954,922 122,760 71,682
8 Pcap 10,200 35,361 19,363 10,097
104,858,829 225,950,064 208,409,250 104,806,410
9 png 102,400 206,207 204,809 102,340
10 shp 25,048 255,124 42,122 22,505
2,487,024 14,953,554 3,734,330 3,031,262
11 vmf
(FOUO, not public) 1,141 39,744 4,840 1,403
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/html
Size: 11642 bytes
Desc: not available
URL: <https://lists.ogf.org/pipermail/dfdl-wg/attachments/20221103/faffe67a/attachment.txt>