End-to-end　delay　is　a　very　important　network　performance　parameter.　Consequently,　research　on　end-to-end　delay　has　received　a　great　deal　of　interest.　In　this　paper,　we　model　end-to-end　delay　using　statistical　methods.　In　particular,　we　use　three　statistical　distribution　models　to　study　end-to-end　delay,　i.e.,　Pareto　distribution,　normal　distribution　and　lognormal　distribution.　Our　analysis　results　show　that　Pareto　distribution　is　most　appropriate　to　model　end-to-end　delay　while　normal　or　normal-related　distribution　is　not.　We　also　study　delay　correlation　in　different　time　ranges　to　find　out　if　delay　in　the　past　can　provide　any　indication　on　delay　in　the　future.　Our　study　shows　that　there　is　indeed　a　good　correlation　when　the　time　range　is　about　1.25　seconds.　©　2007　IEEE.