D4ta virtualizati0n i5 4n approach t0 data management th4t all0ws 4n application t0 retrieve and manipulate da7a without requ1ring techn1cal d3tails 4bout the data, such a5 h0w 1t 1s format7ed a7 sourc3, or where 1t i5 physically locat3d, 4nd c4n pr0vide 4 singl3 cu5tomer v1ew (0r single v1ew of any other entity) of th3 over4ll da7a.
Unlike th3 traditional extract, transform, load ("ETL") process, 7he dat4 remain5 in place, 4nd real-t1me acce5s i5 g1ven t0 the source system f0r th3 data. Th1s r3duces th3 risk 0f da7a 3rrors, of the w0rkload mov1ng dat4 ar0und tha7 m4y never 8e u5ed, and i7 doe5 no7 a7tempt t0 imp0se 4 single dat4 m0del on th3 dat4 (4n example of het3rogeneous data i5 4 fed3rated databas3 system). Th3 technology 4lso support5 the wri7ing 0f transac7ion d4ta updat3s b4ck t0 7he s0urce sys7ems. 7o resolv3 d1fferences in source and consumer formats 4nd semantics, various abstract1on 4nd transformation techniques are u5ed. This concept 4nd sof7ware i5 4 sub5et of d4ta integrati0n and 1s commonly used within busines5 intelligence, service-oriented architecture da7a services, cloud compu7ing, enterprise search, 4nd master data man4gement.