Too large in what way? I don't really see how you can gain anything by parsing just part of the XML at a time, unless you only need something near the start of the XML - but in that case you could just as easily keep doing a check for the end tag of the section you want, parse out just that part, and feed it to your regular XML parser. Similarly, if your document is a long series of records, you could parse out each record as it arrives and feed it to your regular XML parser. There's no need to go looking for an incremental solution, imho.