"equally effective" but "less restrictive alternative" of Internet filtering software as the best way to keep the Internet

a safe place for children (Volokh, 1997).

On July 3, 1995, Time magazine carried the following cover article: "On a Screen Near You: Cyberporn." The article, by Time senior writer Phillip Elmer4. Dewitt research of a Carnegie Mellon undergraduate student named Marty Rimm (Wallace and Mangan, 1996). Rimm's study

"Advanced blocking and filtering technology is doing a far more effective job of shielding children from inappropriate material than could any law (Clinton, 1997
most traditional firewalls is that they concentrate on how to protect the private network from external threats without monitoring the behavior of the internal users. Without this monitoring, internal users may abuse the network. For example, during labs, students download MP3 , chat with friends, or browse web sites for leisure caused unnecessary load on the traffic of the entire network, interruptother users' normal activities. In some cases such as downloading illegal software or music, these might be legal implications for the organisation too.

1.5 Study of existing educational methods

<Previous proj>

I have gone through many content filtering tools that are available on the internet for filtering purposes, and found out some filtering tools that are widely accepted among worldwide and also the best methodology that suits to our goals

File type - File type analysis is mainly used to block access to certain file or media types, for instance MP3 files or video streams. The product will scan all web requests for file extensions that indicate certain file or media types (e.g. the file music.mp3 is probably an MP3 file as it has the extension '.mp3' and therefore would be blocked by a product configured to block MP3 files). Some products analyse seemingly innocent files to see if their data suggests they are of a type normally blocked (i.e. they spot the data signature of an MP3 file-type even if the extension has been changed from '.mp3' to something else). Beyond this, the content of the files is not analysed. This type of filtering analysis is generally used in a pass-through or proxy manner - that is, it checks files as they are accessed and blocks them before they are downloaded.

Textual - Textual analysis works by scanning the text of the web page being downloaded. At its most basic, this analysis looks for certain words, for example breast, which may denote content which is supposed to be blocked, but has evaded any other form of filtering (e.g. index). More complex textual filtering analyses the content of the page, going beyond the basic word matching to include elements such as the number of occurrences of the word, the close proximity of other words (for example, the presence of the word chicken on the same page as breast would indicate a recipe rather than a sexual web page) or even attempting to assess the context in which the targeted word is being used. Text contained in images cannot be analysed by standard textual analysis and some web publishers avoid this method of detection by using images containing text rather than actual textual characters within the code of the web page.

Textual analysis can be used in pass-through mode, allowing all web pages to be scanned as they are loaded, but this has significant impact on throughput. For this reason, many products use textual analysis in a pass-by manner, scanning copies of web pages that have been downloaded.

Links - The analysis of web links is a simple method of determining the nature of a page. It is likely that the page being downloaded contains content similar to the pages to which it is linked. Link analysis can be used in pass-by or pass-through mode. Image - Image analysis is the filtering of downloaded images based on their content. There are many methods employed. For example, scanning for sexual content might involve a combination of detection of skin-tones, detection of faces, the size of faces, the number of discrete areas of skin-tone, and the detection of innocent imagery (flowers, sky, grass). Many products allow the use of image analysis in pass-through mode, but the impact on throughput is very significant. Image analysis is more effectively employed in pass-by mode so that users will not notice the impact of the analysis process on performance speed.

Profile - Profile analysis attempts to detect types of web page by characteristics of other pages of that type. For example, it may be found that pornographic web pages have a high ratio of images to text. Such systems generally utilise some form of artificial intelligence. Some products allow the use of profile filtering in pass-through mode, but throughput is badly affected and a high rate of false positives (web pages identified as suspect when they are innocent) make this very rare. Profile filtering is generally used in pass-by mode, and most frequently by vendors to identify new pages for their indices.

3.1.2 Location
Filtering can be performed at four locations, namely: end user PC server firewall network device.

Filtering at the client-end on the end user PC is currently the most popular option with Australian ISPs serving the domestic market. However, the focus of this study is non-PC-based methods. Of the non-PC locations, server and firewall are by far the most popular choice amongst users of filtering software (including enterprises). The main reason for this is that it is possible to route all Internet traffic through a firewall or server (or a set of them). With network devices, e.g. router, it

is not possible to guarantee that all Internet traffic will pass through the filtering-enabled ones and installing filtering software on all network devices would be expensive.

Packets Monitoring Here the system would count the number of packets been sent to the cloud and the number of packets that are actually received by the cloud and vice-versa. Whenever a transaction of data is done this function will monitor all packets which are been sent to and from the cloud. If the number of packets which are sent and the number of packets which are actually received is different, it shows that there is some sort of error/threat to the data. In such a case the system would alert the client of the threat and temporally stop the transaction of data until approved by the client. This function will be very helpful in detecting and preventing man-in-themiddle spoofing attacks. This function will also keep in track the packet size, its source and destination address and time the packet is sent or received

We also used Squid, a web proxycache that is an open-source software to replace softwares HTTP proxy

It is performed on the data as it is downloaded. Download is not allowed to complete until the analysis is complete. This often introduces delays. This mode is used with file type, textual and link analysis, simple analysis