When making ethers with an alcohol and an alkyl halide, the alcohol must usually be deprotonated for the reaction to proceed smoothly - that is, a base is added. Compare with the classic methylation of phenol, where PhOH + KOH + MeI -> PhOMe + KI + H2O, first the base reacts with phenol, forming water and potassium phenolate (PhOH + KOH -> PhOK + H2O). Potassium phenolate is much more reactive towards methyl iodide than phenol itself, and tus the reaction goes fast and smooth (PhOK + MeI -> PhOMe + KI).
If you are using an aliphatic alcohol instead of a phenol, KOH is usually not basic enough to completely convert the alcohol to the alkoxide (alcoholate) and a stronger base must be used, such as NaH, which irreversibly forms the alkoxide (ROH + NaH -> RONa + H2).
There are ways around the use of a stronger base, which either involves the use of a polar aprotic solvent (like your DMSO above) or alternatively the use of a PTC like TBAB/Aliquat 336 or other. The latter is much preferable, due to environmental, economic and practical viewpoints. If the reaction is performed with microwave heating, a solvent is not even necessary.
An example which combines the above advantages can be found at
http://www.cyf-kr.edu.pl/~pcbogdal/alcohol/
If you don't want to do it in a microwave, check out the references at the above page.