The immediate idea would be to split the data on spaces, but that does not work entirely because your two last "fields" have embedded spaces:

SEG_NUM:1 of 1  DLV_ATT:0     
END_POINT:ESME   FINAL_STATE:DELIVERED   REG_DEL:1
[download]

So I would probably try to first process these two last fields with a regular expression, something like:

@endfields = /(SEG_NUM.+?)\s+(END_POINT.+?)/;
[download]

remove them from the string and then use something like split /\s+/ on the rest of the string, and finally to reassemble the array in the proper order.

Update: I did not originally noticed, but it appears that at least two other fields have embedded spaces:

DEST_IDNT:Syniverse A2P I_ERR:0.0       
PPS_ID: PPS_PROFILE:AO Submission - OA charged
[download]

So splitting on spaces becomes harder to use, at least for about the last half of the original string. Although I don't like the idea too much, perhaps a long regex with each field key is the only solution, at least for the eight fields or so.