Yes, WebViewer supports random access document loading and viewing. This is similar in concept to movie streaming… instead of having to wait for the entire document to download you can instantly see the first page. You can also jump to any page in the document and immediately see it without waiting for other pages.
➢ Also, does WebViewer need to perform any transformations server-side by processing the 100 pages before rendering?
WebViewer expects that files are optimized for fast viewing on the web. This means that files are linearized and pages are normalized so that we can guarantee fast and accurate display on any platform/device. As a normalization format we adopted Open XML Paper Spec (a.k.a. XPS or XOD). Besides being an international standard (http://www.ecma-international.org/memento/TC46.htm) an added benefit is that you can view & print XPS on any Windows machine without any software downloads/installs. XPS files could also be viewed & printed using .NET Framework (http://msdn.microsoft.com/en-us/library/system.windows.xps.packaging.xpsdocument.aspx) or using third party tools.
PDFNet SDK (http://www.pdftron.com/pdfnet/downloads.html) supports on the fly conversion (streaming via chunked transfer encoding) from PDF to XPS/XOD. So you can emitting XPS/XOD as soon as the first page is processed (no need to wait for entire document to process). This is show in WebViewerStreaming sample that comes as part of PDFNet SDK.
Even though streaming via chunked transfer encoding works very well and can be used to provide a more responsive webapp it is probably a good idea to cache converted files. This can decrease the load on your server (since WebViewer can read cached XOD/XPS straight from cloud storage without server involvement). Another benefit is that random access load which is not supported in the streaming mode.