Deep web or “hidden” web refers to the part of the world wide web which generally cannot be found through a normal search engine. The deep web largely consists of specialized databases and websites, which are only generated dynamically by requests from specific databases. The size of the deep web is not clearly determined, but it is many times larger than the part which is visible to search engines, known as the visible web or surface web.
It is difficult to obtain accurate data on the deep web, for example, the study (Bergmann 2001) conducted by BrightPlanet  The following properties were isolated in this study:
The University of California, Berkeley published the following values in 2003 on the scope of the Internet:
According to Sherman & Price (2001), one can distinguish five types of deep web:
The opaque web designates websites that could be indexed, but are not for the time being due to technical efficiency or cost-effectiveness. Since search engines do not take into account all levels of directories and sub-pages of a website, relevant documents may not be reflected in lower hierarchical levels. Particularly affected are websites without hyperlinks or navigation system, as well as unlinked websites.
The private web includes websites that could be indexed, but are not indexed due to access restrictions by the respective webmaster. These may be internal websites, password-protected data or access to only specific IP addresses.
The proprietary web refers to sites which are accessible after confirming usage conditions or by entering a password and thus cannot be indexed. These sites are usually available only after identification.
The invisible web includes sites not indexed for strategic or commercial reasons. From a technical perspective, indexing could be done without a problem.
Websites of the truly invisible web are not indexed for technical reasons. They may be documents that cannot be displayed directly in a browser, or file formats that cannot be detected because of their complexity (mostly graphic formats), or non-standard formats (for example, Flash).
The deep web contains a lot more additional data as opposed to the surface web. An integration of these results can be beneficial to users because potentially applicable results could be obtained. However, an efficient implementation of such a search engine for both surface and deep web is difficult and getting a selection of appropriate sources for a search query could be problematic. In addition to scientific and legal data there are, however, many opaque sites in the deep web. In addition to a huge black market, there are also many websites of cybercriminals, political extremists (neo-nazis, revolutionaries), etc. Therefore, the deep web should be used with caution despite the large supply of helpful documents and data.
Search engine optimization specialists strive to achieve favorable search results for users. Therefore, well linked, contextually relevant websites should achieve a correspondingly high Google index ranking. The exact procedure for crawling and ranking of the deep web is still unknown, however, SEO specialists are developing strategies with which to efficiently make documents of the deep web accessible to search engine users.