Where great minds meet Honeycomb Connect
Home | My Cell | Help | Contact Us
Select Hive/Cell
Corporate Information Technology (CIT50)
March 18, 2018  
Corporate Information Technology (CIT50)
Corporate Information Technology
Executive Member Profiles
CIT50 Executive Reports
Solutions Exchange Gallery
About Honeycomb Connect
Contact Us
Corporate Information TechnologyCIT50 Executive Reports
Securing The Data Center, Part 2: Bursting Your Security Balloon
July 1, 2007
Robert Grapes, Cloakware
...back to list


Find Out MoreSuppose we compare your data center to a water balloon. Both have a relatively secure perimeter. Both contain valuable content (water is the raison d’etre of a water balloon). Both face significant threats from pointed attacks. And in both cases, Bad Things happen when the perimeter is breached.

So why the comparison? Consider the following: How much do you spend to protect your data center applications from outside attacks? How about from attacks launched inside your network security perimeter? How secure is your valuable data against the misuse of privileged access accounts? When was the last time you changed all of your database passwords or all of your server passwords? Often, the answers to these questions reveal that a typical data center is about as secure as a water balloon.

In this article we burst the bubble of some common yet risky misperceptions about data center security. The goal is to get you thinking about threats that you may not have considered before. Then we describe some proven strategies you can adopt to resist these threats and improve the security of your valuable data.

Perimeter security is only a first step

If you’re like many organizations, in the past decade you’ve focused your IT security efforts on strengthening the perimeter security of your data center. For instance, you may have installed an intrusion detection system (IDS). In our balloon analogy, this is like using thicker rubber: the barrier is stronger, but it remains the only defense against breaches. 

A security strategy based on perimeter security assumes a lot of faith in the strength of the barrier. While it may be possible to resist known threats, the trouble is attackers eventually find news ways around barriers. And when they do, you’ll want to have in place additional layers of defense to limit the scope and depth of the breach.

By itself, perimeter security is problematic for another reason: it’s increasingly hard to define exactly where the perimeter is. As your company expands through organic growth or acquisitions, so too does the makeup and complexity of your network. Partners and customers gain ever more access to your online services, blurring the line between your network and theirs. New applications and technologies deployed beside your legacy systems create new administrative silos that may span historic network boundaries. How do you resolve the conflicting challenge of enabling easy access to and availability of your corporate data, while ensuring that the data is secure? It’s tough to fill a balloon with water when the balloon itself is dissolving.

Unmanaged passwords are everywhere

The dissolving perimeter of the data center creates new security risks in previously simple network operations. Consider basic authentication, whereby devices and services in the data center establish their identities with each other. Over 90% of these operations rely on an ID and password. If you include router passwords, server passwords, application passwords, and database passwords, you quickly find that these “elevated privilege” accounts outnumber end-user accounts. Most organizations force their users to change their passwords often, but how often do you change the hard-coded “application to application (A2A)” passwords in your data center? If you’re like most organizations, it’s rarely or not all. However, a dissolving perimeter exposes these unmanaged A2A passwords as a growing risk. 

It gets worse. Consider all the administrative accounts in your data center. How many of your administrators share the same IDs and passwords in order to maintain your systems? How many developers know the passwords to your database systems? While we assume we can safely trust our insiders, legislation (such as SOX, PCI, HIPAA, MITS, FISMA, etc.) says that you can’t, and therefore insists that you eliminate default passwords, establish unique IDs and passwords for every system and connection, and update all these passwords frequently. If your data center includes routers, switches, storage area networks, UNIX servers, Windows servers, mainframes, databases, and other systems, then it’s likely you have inconsistent policies for password composition and updating, different tools to make the updates, and different resources responsible for the updates. Then there are audit and traceability requirements for all of these credentials. Now how confident are you that your systems comply with the relevant legislation?

Even if you have started to implement Kerberos or a PKI on your network, how much penetration of those techniques have you achieved in your data center? Will these techniques ever deliver 100% coverage for your authentication needs? Also, when considering the insider threat,  have you asked yourself how to protect the Kerberos and PKI keys during usage in memory, or for “auto-start” processing? Anyone who underestimates these risks is encouraged to review Department of Justice statistics on insider threats. It’s also wise to consider some high-profile security breaches in 2007, including the multi-billion-dollar theft of credit card information from a large retailer (a perimeter security breach), the successful hacking of DVD encryption keys (unprotected in memory), and the repeated successful hacking of security-related U.S. government networks (perimeter security breaches).

We’re asking lots of questions here and that’s the whole point of this article. A Threat/Risk Assessment (TRA) is a formal method for asking the questions needed to identify the threats that face your network, systems, and applications. The next step, risk mitigation, is all about applying the appropriate security techniques to thwart the identified threats. Do you conduct regular TRAs for your network and applications? Do you consider the insider threat in those assessments?

Let’s look again at the challenge of basic authentication in the context of a TRA. As part of your analysis, ask yourself, what security techniques do you require to:

  • stop tampering attacks on your unattended applications?
  • stop software copying and execution on other machines?
  • stop the sharing of IDs and passwords with your developers and administrators?
  • stop the lifting of decryption keys from memory or disk?

Fortunately, there are answers to all of the questions posed in this article, and there are solutions that deliver all of the necessary security techniques to these problems. Improving your password management capabilities in the data center and hardening your applications against threats that breach your perimeter defenses are two vital steps you can take to build layered defenses. The lesson today is that you need to do much more than strengthen the perimeter of your water balloon in order to protect its contents from external and internal threats.

Printer-friendly version
E-mail this to a friend
Comment on this story
...back to list
Member Site Tour
User Name:


Forgot password?
My Cell | Help | Contact Us Privacy Policy | Legal disclaimers
Copyright © Honeycomb Connect