The Business Forum

"It is impossible for ideas to compete in the marketplace if no forum for
  their presentation is provided or available."           Thomas Mann, 1896


Web Content Filtering in the Corporate Network Perimeter

Contributed by Panda Software (USA), Inc.

 

 

1 Introduction

The Internet is the source of information most widely used by companies worldwide. Two in three employees use it as a work tool. However, the Internet also offers a wide range of non work related content, such as leisure services, online shopping and a large list of etceteras, which employees with Internet access could also access during work hours. For this reason, it is vital for companies to be able to control Internet content that their employees can access, and ensure beneficial use of this technology and avoid loss of productivity.

1.1 Objective

Discover the different techniques for preventing the users in a corporate network form accessing Internet content that is not relevant to the company and how GateDefender Performa handles undesirable Internet content and the filtering function of GateDefender Performa.

1.2 Audience

The content of this document is aimed at:

• IT users.

• Network administrators.

• Security experts.

• CTOs (Chief Technology Officer).

• CIOs (Chief Information Officer).

2 Why is web filtering needed?

Internet provides a wide range of content related to all topics. A large part of it is necessary to develop business activity. For this reason, companies in the twenty first century need the information available on the Internet to guarantee good results. However, the universal nature of this content allows employees with Internet access to make personal use of company resources, accessing content that is not related to their work, and thereby degrading the company’s profitability.

2.1 Economic impact

Misuse of the Internet has a negative influence on companies’ results by:

2.1.1 Lower staff productivity

Companies lose productivity when their employees waste time visiting non work related web content , not only content related to their hobbies but also personal matters, such as online banking, employment offers, chats, etc. The loss to the company can be measures in the cost of the hours employees spend visiting these types of websites.

2.1.2 Increasing use of network resources

All the information transferred between the Internet and the corporate network uses costly network resources, such as bandwidth. If employees use these resources to access content that does not benefit the company, the traffic of information that is important to the company will be affected, delaying the processing of this information and the work of colleagues

2.1.3 Increasing security risks

The weakest part of a security system is the human component. Employees downloading potentially dangerous programs or files from the Internet could infect the network with threats like spyware, capable of extracting confidential information and sending it to third-parties.

2.1.4 Increasing legal risks

Accessing web pages with illegal, offensive sexual or violent content or downloading these types of content can damage corporate image, which should always try to guarantee a dignified work environment for its employees. The damage to the corporate image is worsened if this content is forwarded from a corporate workstation.

2.2 Quantative studies

Companies need to be able to access their employees’ Internet access. Between 30 and 40 percent of Internet use is non-work related. At least 60 percent of employees use the Internet at work for personal reasons (chats, auctions, online shopping, etc). What’s more, data shows that 70 percent of accesses to web pages with pornographic content are visited during working hours, forcing companies to consider non-work related Internet use as a responsibility that they must assume.

The damage that access to irrelevant web content can cause the companies is varied.

There is financial damage, as it reduced employee productivity generating huge costs calculated as working hours.

It could also cause legal damage, as the company could be related to illegal Internet content, such as fraud, piracy, pedophilia, etc.

The technical damage must also be considered, as unproductive web browsing could use costly resources like bandwidth and generally degrades local network performance.

3 Preventing access to undesirable Internet content

There are many different ways of preventing employees from access irrelevant content in the Internet. However, not all are truly effective.

3.1 Possible alternatives

Given the structure of the Internet, the only way of preventing access to undesirable content is to increase employee awareness.

3.1.1 Corporate policy

Many companies establish rules in their policy that prohibit employees from accessing certain types of pages. In large companies the problem is difficult to control, as the large number of employees makes prevents companies from ensuring that all employees adhere to the corporate policies.

3.1.2 Monitoring

Attempting to control employee access through supervision is an expensive option, as it requires staff dedicated to analyzing Internet access. In companies with a large staff or large volume of Internet traffic it is impossible to implement this option.

3.1.3 Restrict access through a proxy server:

There are network components known as proxies, which can be hardware or software devices and prevent certain computers in the corporate network from accessing certain Internet IP addresses. However, this solution is difficult to maintain, as the IP address of pages with undesirable content change every day.

What’s more, access is restricted based on IP addresses and not content. This restricts access to entire domains, in which not all information is irrelevant and could contain material that is very to useful for the company’s staff.

3.1.4 Intelligent web content filtering

This option involves the company restricting access to certain content and preventing this undesirable content from reaching employees. What’s more, it offers manual control over access to certain web pages and allows users to be excluded from the web filtering, making the only truly effective option available for implementing reliable content filtering.

4 The GateDefender Performa web filtering solution

Panda GateDefender Performa offers the most effective protection for preventing users of the corporate network from accessing undesirable web content, ‘intelligently’ analyzing employees’ requests for access to web pages and offering a rapid and effective response if these pages are considered undesirable for the company.

4.1 Automatic filtering

The corporate network administrator can access a series of categories and allow or deny access to them. These categories are divided into groups and access can be allowed or restricted for each group or each sub-category.

The 59 undesirable content categories in GateDefender Performa range from media to online auctions, through spyware, violent content, instant messaging, etc. It is obvious that it is not easy to add a web page to a certain category based on simple coincidences.

To sum up, the only thing the administrator needs to do to filter web content is select the content categories that must not be downloaded. By doing this, GateDefender Performa blocks access to web pages containing elements that fall under this category.

In order to include web pages in one category or another, an extremely advanced intelligent analysis method is used, and the results are included in a file.

4.1.1 Creating a category file

The page file organized into categories is generated remotely and automatically updated every 90 minutes without user intervention. This prevents GateDefender Performa from having to analyze every page viewed, which would delay access to the page, making web browsing extremely tedious.

In order to create category files, servers located in different geographical locations are constantly scanning the Internet for these types of contents. Over 2,600 million pages are scanned and added to the file, generating over 100,000 new entries a day. This results in the biggest database of URLs in the world, which grows every day and is automatically updated without degrading its filtering activity.

4.1.2 Intelligent scanning of web pages

Each web page found is thoroughly scanned using different methods.

• Text analysis

• Image analysis

Intelligent analysis of a web page through both of these methods allows pages to be accurately added to a category, as an in-depth scan of the web pages prevents classification errors.

Text analysis

A medical page or psychiatric page contains words related to violence as it describes mental health problems and could, therefore, easily be included in the violence category. However, this would not happen with the scan for generating the categories, as it does not only look for single words but analyzes the content in which the appear.

Some pages could easily be classified as undesirable could include texts in images that the text analysis would not pick up. However, the text analysis engine incorporates OCR (Optical Character Recognition) capable of analyzing text inserted in images as of it were written text on the page.

Image analysis

The images in a web page allow pages to be identified as online shopping, pages with sexual content, etc,

The image analysis engine can detect the content of each web page, distinguishing

• Faces

• Nudity

• Logos

• Objects

What’s more, its powerful texture analysis feature helps determine the nature of each image. This is better explained using an example:

When a face is detected in an image, the percentage of the image that corresponds to the face is calculated. If the image is much bigger than the face, it is probably a photograph of a full body. In this case the texture of the face is analyzed (the skin) and the percentage of the image with a similar texture is evaluated. If the percentage of the skin texture in the photograph is quite high, it is easy to determine that the image corresponds to a naked body.

The real analysis process is much more complex than described in the example in order to avoid mistakes, but it is without a doubt an artificially intelligent engine, capable of detecting the nature of every image analyzed.

Once a web page has been blocked by both engines, the page is classified based on the result of each analysis. As you can see the classification implemented by the automatic filtering by categories is extremely reliable and the most interesting aspect of this process is that GateDefender Performa does not have to carry it out, as this is done remotely and the results are entered in GateDefender Performa, resulting in optimum performance.

4.2 Manual filtering

A tool is also available to administrators to manually add or exclude the content of certain pages from filtering.

4.2.1 White list

The administrator can add trusted domains, URLs or IP addresses to the white list that network users will always be able to access, regardless of the category or whether they are prohibited or not. This is better explained using an example.

A company may not want employees to be able to access news sites to prevent them from wasting time. If the newspapers and magazine category is selected, this content will be filtered and users will not be able to access them. However, a publication could contain information that is important for the company and therefore, employees need to be able to access it. Simply add the domain of this publication to the white list and access will be permitted.

It is important that access to a page in the GateDefender Performa white list is never analyzed to check the category it belongs to. Access is allowed, optimizing the speed and taking the workload off GateDefender Performa.

4.2.2 Blacklist

The blacklist does just the opposite to the white list. The administrator can add trusted domains, URLs or IP addresses to the blacklist that network users will never be able to access, regardless of the category or whether these pages are in a prohibited category or not. This is better explained using an example.

It is understandable that a company does not want its employees to access rival companies’ websites that offer jobs in order to prevent staff from being lost to these companies. If the company activity is “normal”, the websites of these companies probably don’t belong to an undesirable web content category and GateDefender Performa will not block access. In this case, the domains of rival companies can be added to the blacklist. This will prevent corporate network users from accessing these pages.

It is important that GateDefender Performa denies access to a page in the blacklist without analyzing it to check the category it belongs to, optimizing the speed and taking the workload off GateDefender Performa.

4.2.3 Excluded from filtering (VIP list)

Administrators can access a list called the VIP list, where they add the address of computers that will be able to access all Internet content. These users will be excluded from web filtering.

4.3 Actions

When a computer connected to the corporate network tries to access undesirable web content, GateDefender Performa will block access to the page. The network administrator can define the alternative page that will be displayed. The content of this page can be edited by the administrator in the preferred language to inform network users that access to the page is restricted.

Whenever necessary, administrators can access real-time graphic information about the functioning of the web filtering system. What’s more, GateDefender Performa automatically generates detailed graphic reports on access blocked or allowed from corporate network workstations. The content of the reports and the fields to include can be easily defined by the administrator.

5 Main benefits of GateDefender Performa

The main advantages of the new version of GateDefender Performa are its efficiency, automatic updates, ease of use (‘plug-in and forget’) and high performance and scalability.

  • Complete protection: It scans the HTTP web browsing protocol through all the ports defined by the network administrator.
     

  • High performance, transparent to users: Panda GateDefender Performa takes over the workload of the traditional protection, optimizing use of the network resources. It guarantees excellent performance and the highest scanning capability of its class. It can scan up to 170 Mbps completely transparently to corporate network users and without degrading network performance.
     

  • Auto-updates: The updates system against new spam messages is programmed by default to be carried out completely automatically every hour and a half. This means that Panda GateDefender Performa will be the most up-to-date protection across the network.
     

  • Simple administration. Plug-in and forget: Panda GateDefender Performa is designed to be implemented simply in any network, with no need for redirecting network traffic. GateDefender Performa sends to the administrator in a proactive way all relevant information about the Web Filtering activity, in real time graphic reports. It is managed remotely and securely through a simple and intuitive web console.
     

  • Low cost of ownership: By preventing saturation of network resources and loss of productivity, Panda GateDefender Performa offers higher resource management capabilities. All of this, along with its 'plug-in and forget' operation, requiring minimum administrator intervention, result in low cost of ownership of Panda GateDefender Performa.
     

  • High scalability and load balancing: GateDefender Performa adapts perfectly to the needs of all small to large companies, adjusting scan capacity to the overall volume of network traffic. Its load balancing is completely automatic and native, allowing workload to be shared across multiple units. The result is increased scalability and improved antivirus performance for complete protection of your network perimeter.
     

  • Detailed graphic reports and customizable alerts: GateDefender Performa offers graphic statistics on the network traffic and system activity. It also offers comprehensive web filtering graphic reports, as well as customizable alerts and notifications.

Besides, the chance to send all important events through the standard log format Syslog, allows the administrator to analyze everything is happening in the GateDefender Performa units with a standard analysis system which can monitor many network devices at the same time, in order to control all network elements. Similarly, all events produced in GateDefender Performa units are sent through SNMP protocol in order to allow centralized monitoring of all devices installed in the company network.

Other options available: As well as the web filtering protection, GateDefender Performa also offer other types of protection

  • Anti-malware protection: This prevents unknown viruses, worms or any other malicious code from getting into the company and saves network resources and bandwidth, blocking potentially dangerous content before it enters the network.
     

  • Anti-spam protection: This prevents junk mail from reaching corporate network users, avoiding unnecessary use of corporate network resources and optimizing employee productivity, as they do not have to waste valuable time opening, reading and deleting spam messages.

6 Technical details

6.1 Operating system and interception software

The software incorporated in GateDefender Performa is based on the GNU/Linux operating system, reinforced and optimized to offer maximum security and high performance. The operating system used in GateDefender Performa only includes the services and processes it needs to function correctly.

Panda GateDefender Performa acts as a transparent bridge between the Internet and the corporate network. This means that the traffic passes through transparently and the device intercepts HTTP sessions.

6.2 Self-repair system:

Under extreme circumstances (excessive heat, etc.) damage could be caused to the Panda GateDefender Performa hardware, such as hard disk corruptions or physical errors in some partitions, etc.

To prevent system failures in these cases and ensure continuous perimeter antivirus services, Panda GateDefender Performa includes a system for controlling partitions or a self-repair system. When Panda GateDefender Performa starts, a control partitions starts up that determines which work partition will be used to run the anti-spam system.

If the self-repair detects that the work partition has failed for any reason, the next work partition established will be started up. It also incorporates a process for restoring the partition that has failed in order to guarantee that it works in subsequent startups. The partitions are kept updated with the latest versions of the software, in order to make recovery as little traumatic as possible.

There are other perimeter solutions that promise high-availability using SCSI RAID disks.

Panda Software believes that SCSI RAID disks are only necessary for storage servers that have to protect data and provide high-availability. However, SCSI RAID disks don’t offer any benefits in devices that are not geared to storing data, such as Panda GateDefender Performa or other appliances on the market. This is really just another marketing angle that does no more than add to the cost for the end-user and Panda Software would prefer to eliminate unnecessary costs. Panda GateDefender Performa offers high-availability by using tools like WatchDog and Master Boot Manager, as well as load balancing.

6.3 Watchdog: fault tolerance system

The motherboard of GateDefender Performa incorporates a system monitoring circuit to prevent the system from blocking or failing for long periods of time. The WatchDog system receives signals indicating that it is functioning correctly and if it does not receive this signal within a specific interval, WatchDog will completely reset the system from the motherboard, avoiding loss of network services for excessive periods of time.

Thanks to this fault tolerance system, Panda Antivirus GateDefender Performa can recover automatically from service failures either in the applications or the operating system.

As well as WatchDog hardware, Panda GateDefender Performa also incorporates WatchDog software, which periodically monitors the status of all the processes that are running. If it detects a process is not responding, WatchDog will try to recover it without needing to restart the system. If it cannot be recovered through the software, GateDefender Performa will be completely restarted from the WatchDog hardware.

Thanks to this fault tolerance software system Panda GateDefender Performa can recover automatically from service failures without needing to completely restart the system.

6.4 Load balancing

Load balancing allows the workload to be shared across multiple GateDefender Performa units, providing improved performance and higher availability. No additional hardware or software is needed to implement load balancing across multiple GateDefender Performa units and although it is highly advisable to use switches, these can also be connected through hubs.

One of the appliances will act as the master, with the rest acting as slaves. From the administration console, users can view all the appliances in the load balancing system and the mode each one is running in.

When several GateDefender Performa units are installed in parallel, they will automatically negotiate the role or mode of functioning of each one and whenever a new appliance is incorporated, the modes of functioning will be re-negotiated.

The master appliance implements a load balancing algorithm and redirects connections to the different slave appliances in order to balance the system workload. The master appliance will also scan connections and let clean traffic through.

Slave appliances will not let traffic through and will simply scan the connections redirected to them, returning clean traffic to the master appliance.

7 Summary and conclusions

The negative impact of browsing undesirable web content during working hours on companies is a proven fact. As is nearly always the case, adopting preventative measures is actually more profitable than having to correct an undesirable situation which may increase the risk of damaging the reputation of the organization.

Employees must therefore, be kept well informed when it comes to using Internet resources to do their job without causing drops in performance during working hours. Controlling Internet access manually in companies and using no other method is not an option.

The same can be said for banning using from accessing the Internet.

Although all members of the company should be involved in preventing the dangers of using the Internet for personal use someone - the network administrator - must coordinate these efforts. And this coordination task is much easier if the administrator has a tool like GateDefender Performa which allows the web content filtering policy to be installed, maintained and supervised with a minimum effort.

APPENDIX A. Glossary of terms

Proxy: Device for centralizing Internet queries from a network.

IP: Internet Protocol

IP address: Address of each network devices, compatible with the routing mode of the IP protocol (Internet Protocol).

URL: Stands for Uniform Resource Location. This is the location of a specific resource.


Visit the Authors Web Site

Website URL:

 http://www.pandasoftware.com

Your Name:
Company Name:
E-mail:

Inquiry Only - No Cost Or Obligation




 3D Animation : red star  Click Here for The Business Forum Library of White Papers   3D Animation : red star
 


Search Our Site

Search the ENTIRE Business Forum site. Search includes the Business
Forum Library, The Business Forum Journal and the Calendar Pages.


Disclaimer

The Business Forum, its Officers, partners, and all other
parties with which it deals, or is associated with, accept
absolutely no responsibility whatsoever, nor any liability,
for what is published on this web site.    Please refer to:

legal description


Home    Calendar    The Business Forum Journal     Features    Concept    History
Library     Formats    Guest Testimonials    Client Testimonials    Experts    Search
News Wire
     Join    Why Sponsor     Tell-A-Friend     Contact The Business Forum



The Business Forum

Beverly Hills, California United States of America

Email:  [email protected]

Graphics by DawsonDesign

Webmaster:  bruceclay.com
 


© Copyright The Business Forum Institute 1982 - 2012  All rights reserved.