An early WS loanword in Akkadian is not to be excluded.
See also dūru ‘eternity; permanent position’ (AHw. 178, CAD D 197), dūriš ‘forever’ (AHw. 178), dāriš ‘forever’ (AHw. 163, CAD D 113), dāriš ‘for eternity’ (AHw. 164, CAD D 114), dārānu ‘eternaľ/’eternal sacrifices’ (AYw. 163, CAD D 109), dārīˀu ‘(ever)lasting, eternaľ (AHw. 164, CAD D 115), dārīˀa ‘eternally’ (CAD D 112), dārītu (AHw. 164, CAD D 114), dārūtu (AHw. 164, CAD D 118), dārūtaš (AHw. 1550, CAD D 118). This derivation, unquestionable in terms of the underlying consonantal root and the semantic development, is not without problems as far as derivational morphology is concerned, as the expected prototype *dawr- (Marrassini 1971: 48, Fronzaroli 1965: 143, 148) would yield dūru in Akkadian, actually well attested with the same meaning (see above). No ready solution for this problem is at hand. Shall one postulate a bivocalic pattern *dawar- contracted into *dār- in the prehistory of Akkadian? Derivation from the by-form *dahr- is not to be ruled out completely, but does not seem likely.
According to a broad concensus (AHw 164, CAD D 115, Streck 2000: 88), an early WS borrowing comparable to Hebrew dōr and its cognates listed here.
No etymology can be suggested. If the Sargonic spelling (preserved in OB copy) id-ke2-as2-su-nu-ma is taken at face value, the underlying third guttural must be *ḥ (сf. Kouwenberg 2010, 573), but no suitable Semitic parallel could be detected.