How to prevent data from leaving your network

Three data loss prevention tools reviewed

By David Strom, ITworld |  Security, data loss prevention 3 comments

Data loss prevention (DLP) is an emerging field with a lot of different products and players. The idea is to stop information from leaving your internal networks in close to real time, so you can identify the leaker or thief before too much damage (and ensuring lawsuits) happen.

A recent study by DLP vendor Proofpoint found that more than third of the respondents had an incident in the last year, and a quarter of them had investigated leaked information as a result of a blog post.

There are more than a dozen different DLP vendors. We show you three typical products, how they work, and what kinds of information they track.

  • Global Velocity's GV-2010 security appliance,
  • BlueCoat Networks DLP appliance, and
  • Sendmail's Sentrion email server.

Each is designed for somewhat different situations, which is why we have collected them together. Before you dive into these products, you might want to address the following questions:

  • Who will own the DLP process in your organization: Will it be the general IT staff, the infrastructure management group, the desktop security group, or some other combination? Depending on this ownership might compel a particular collection of DLP products.
  • Where does DLP presently touch your existing IT security infrastructure? Most firewalls and email servers have some DLP capabilities; the tricky part is being consistent across your enterprise and getting a specialized DLP product that can complement and in some cases work with these legacy devices.
  • Are you looking at total DLP protection, for endpoints, data in motion and file server data? No single product can handle all of these situations; so how each vendor partners and integrates with others for complete coverage is critical.
  • Do you want something to decrypt emails and https traffic? Not all products can see inside these protocols without some additional work.

All three products have the ability to scan for particular character strings (like a Social Security or credit card pattern) and also upload sensitive documents into their protective scanning engines to ensure that this specific unstructured information is also protected. Another typical situation is where a rogue employee will send a customer database list to their personal Gmail or Yahoo mail account, and then downloads or forwards this information once they get home. Each product has a variety of reports to show you incidents flagged by the protective policies and what information was leaked.

All three also have mechanisms to test new policies to ensure that they are actually protecting you under the specific circumstances that you intended. This is very useful; otherwise you could end up with a lot of false positive incident reports to plow through. Once you have tested a particular policy, all of them make it fairly easy to activate the policy and have it start working across your data streams.

Sendmail's Sentrion is a high-powered email server that takes the code base of Sendmail – found in any Linux server – and bulks it up with a series of powerful mail processing policies. Some of these can be used to scan outgoing messages for particular character strings, or data types, to ensure that confidential data doesn't get emailed outside of your domain.

Sentrion has very powerful mail processing policies, as you can see here we can specify a variety of conditions to be met to trigger a particular DLP incident.

Sendmail has partnered with a variety of vendors, including Voltage Security, to handle decrypting emails to ensure that no one is hiding sensitive information inside these messages. They also have an extensive collection of add-on applications in their own AppStore, patterned after Apple. For example, one called the DLP bundle also includes RPost, software that can send registered return email receipts. But Sentrion is email-centric: if you want endpoint protection (such as to flag when an employee steals data on a USB stick), or protection for emails sent over Webmail providers such as Gmail and Hotmail, you will need something else.

3 comments

    Anonymous 1 year ago
    1. Comprehensive channels coverage. It is impossible to predict which outbound channel the next data leak will occur. Some expected avenues are: corporate email, private email, webmail, blog, instant messenger, P2P application, internal web or FTP server etc. Therefore, the DLP system must cover ALL the relevant channels. The majority of “DLP” systems do not even try to cover all network channels. Typically, they cover SMTP, FTP, HTTP (client side), sometimes HTTPS and instant messaging. This coverage is further handicapped. For example, scanning SMTP, these systems require integration with the corporate email server and inspect only emails sent through it. Emails sent through an external ISP are overlooked. Emails accessed from outside the perimeter through POP3 or HTTP (server side) are ignored by such solutions. The dangers of file sharing applications and exposure of the internal web servers are disregarded. 2. Enforcement Data Leak Prevention, by its definition, requires electronic enforcement of the data security policy – i.e. the product must be able to effectively block transmission of protected data. Many “DLP” products being sold are actually DLD – Data Leak Detection products. They are designed to report what data breaches have occurred, instead of stopping them in real time. 3. Content Inspection The true DLP solution must inspect content. Making decisions based on the form (file type, file attributes etc.) or meta-data (author, language, size of attachment etc.) is not enough. 4. Accuracy The DLP solution must be sufficiently accurate. Among two types of errors (false positives and undetected leaks) the more dangerous error is a false positive. In the enforcement mode, even a small amount (0.1%-0.2%) of false positives can wreak havoc in the organization. Therefore, a DLP solution has to employ detection technology with virtually zero false positives. Another aspect of accuracy is that the DLP system must protect data and not a specific form of its representation. Therefore, the DLP system must be resilient to typical modifications of the data, such as excerpting, embedding, changing file format, re-ordering, re-typing, text re-formatting etc. 5. Non-duplicating protected data. The DLP solution must not duplicate the protected data in any form! If it does, then DLP becomes Data Leak Provoking. But many vendors still sell products, copying the data they are supposed to protect into their internal database. Encrypting such data, or keeping it in the form of the search index is not enough to satisfy this requirement! From: http://www.gtbtechnologies.com/support_evaluation_dlp_downloads.asp
    Anonymous 44 weeks ago in reply to Anonymous
    Network DLP and DLP with integrated PC based agents without a doubt adds more assurance and security for companies concerned about Data Loss. This last comment I believe addresses the critical decisions/ functionality that can be requested when selecting a DLP solution. I might add that the vendors need to start addressing how their solutions would be integrated with Cloud Email providors. More specifically how would one architect the DLP solution if the SaaS email solution was off premise?Thanks - Great Discussion. I am going to shared this thread with my CIOs on my site www.cioes.org
    Anonymous 1 year ago
    Beyond software, people need to architect a hardware and software solution that prevents or minimizes data loss. A computer security table top exercise like the one I've linked to below, can help people identify weaknesses in their network:http://misterreiner.wordpress.com/2010/06/02/cse01/

      Add a comment

      Post a comment using one of these accounts
      Or join now
      At least 6 characters

      Note: Comment will appear soon after you have activated your account.
      Obscene/spam comments will be removed and accounts suspended.
      The information you submit is subject to our Privacy Policy and Terms of Service.

      ITworld LIVE

      SecurityWhite Papers & Webcasts

      White Paper

      A Proactive Approach to Server Security

      Learn why security-conscious organizations are taking a more proactive approach to server security. Download this Spire Research whitepaper to understand how you can eliminate the threat caused by today's more advanced threats and protect your organization's most valuable data.

      White Paper

      Protection Against Modern Cybersecurity Threats

      Download this case study to learn how this accounting and consulting giant uses Bit9's adaptive application whitelisting to offer employees flexibility without jeopardizing enterprise safety.

      White Paper

      Stop Hackers Before They Attack

      Hacktivism, Identify Theft, Financial Gain, Cyber War - regardless of motivation, stopping today's hackers requires a new proactive approach to protecting endpoints. Learn how this New England hospital, breached multiple times by targeted attacks, put an end to the malware with Bit9 Parity. Their IT team can now identify malware and secure PCs and workstations -protecting patient care and privacy.

      White Paper

      From the Frontline - Preventing APT

      Is your company's network secure? Are your endpoints and servers secured? Before you answer, read this case study on a US Military Command that discovered no matter how much you educate users, hackers can get through traditional defenses. This targeted attack blew through all layers of their security, except one: Bit9 Parity's advanced threat protection.

      White Paper

      Protecting Point of Sale Systems from Targeted Attack

      If you are responsible for protecting retail systems, download this case study to learn how this retailer eliminated the threat of malware on their POS systems using Bit9's award winning solutions.

      See more White Papers | Webcasts

      Ask a question

      Ask a Question