Role of CS-MARS in Your Network

CS-MARS as an STM Solution

This section explains the advantages CS-MARS provides beyond interaction with network and security devices. As discussed in Chapter 1, CS-MARS is a powerful solution that provides the necessary features of STM.

Reasons for an STM

In Chapter 1, you learned the benefits of moving from SIM to STM technology. Now let's look at a few of the specific advantages of an STM and take that first step to a proactive security framework. The following are some true examples of how STM gives you the advantage. The company names are fictitious, but the details of the stories are true.

Day-Zero Attacks, Viruses, and Worms

Think about Sasser, one of the more recent, notorious, fast-moving day-zero worms, before it had its name. What it did and what was affected were not apparent until the damage was already done.

You probably didn't even know it existed because there was not a known vulnerability for it to exploit. Network infrastructures were pummeled by the malicious code overnight, and 24 hours later, the worm was given a name. Now picture a system that informed you that a host on your network was behaving abnormally and opening hundreds of connections to other hosts. That system then allowed you to shut down the port that the offending host was connected to, giving you the switch name and specific interface. Because this is not normal behavior for a host on your network, you pushed the "red button" and stopped it. 24 hours later, an AV or HIPS signature was released, and news about the new worm spread. Now it's time to update those hosts with patches, and you're not scrambling around trying to do damage control.

CS-MARS stopped Sasser exactly as described. It recognized the attack using its anomaly algorithms and provided security responders with a timely and accurate mitigation suggestion, saving customers hundreds of thousands of dollars in tangible recovery costs.

Monitoring and Enforcing Security Policy

Widgets, Inc., just approved an addendum to its existing security policy and mapped out a new VLAN for its wireless access network. According to policy, users must authenticate to a Cisco VPN concentrator and tunnel into the LAN to use wireless. If a user does not have a Cisco VPN client, all it gets is a default gateway to the Internet. The policy calls for periodic audits of user access to the wireless segment. Security personnel are overwhelmed as it is with day-to-day responsibilities; however, because they have CS-MARS installed on their network, they just have reports automatically generated according to policy and sent via e-mail to the auditors.

Since the security department has configured the CS-MARS for the automated report, it has not had to spend any time on the new addendum. This has freed up the security engineering resources to work on security-related problems.

Insight, Integration, and Control of Your Network

At approximately 3:00 p.m., ABC University's help desk started to get trouble calls about failed access to servers on its network. The help desk immediately notified the network engineer on staff; she was at her desk eating lunch. When she received the call, she attempted to access the servers in question and confirmed that they could not be reached. She then attempted to access other devices from her machine and had no issues. She opened her web browser and decided to log in to her new CS-MARS appliance. She noticed a spike in port 53 activity and immediately began investigating. She discovered that her secondary DNS server on the private university network was going haywire. She made the decision to use the CS-MARS to shut it down. When the DNS server was offline, access to the other servers in the same segment was restored.

It took her 3 minutes to get it under control and 5 minutes to find out what was going on. She used CS-MARS to discover that several crafty students had hacked into the DNS server and placed a homemade Doom relay application using port 53 to circumvent the firewall policy that prevented Internet Doom contests. The students were identified and reprimanded.

The network engineer finished her lunch without getting up from her desk, and the university saved thousands of dollars in post-attack research and other potential recovery activities.

Auditing Controls

Investment firm Y is a public company and, therefore, must conform to SOX auditing. The third-party auditing firm requested access logs from its Microsoft servers, COBIT DS 9.4—Configuration Control, and Successful Object Modification logs from their Oracle database servers. The auditing company requested that the reports be prepared for review on the third day of the audit. The day the auditors arrived on the premises they were greeted, introduced to the staff, and handed the reports in a nice three-ring binder. At the end of the audit, the auditors commented on how well prepared the investment firm was and praised them on their efficiency.

It took the security department one hour to run, print, and bind the reports. This left them free to focus on security-related activities.

Monitoring Access Control

Bob is an employee of Company X and finished his project a little early on Friday. He had time on his hands before the end of the day. Bob decided to click his My Networks icon on his desktop and look around. He ran across an HR file server and clicked the icon.

Eric, the IT admin for HR, had a long day yesterday. The director for HR was having issues with getting access to the file server, and Eric just couldn't get permissions to work. To save time and embarrassment, he just set permissions to All.

Bob ran across a file marked "John D. offer letter." John D. is Bob's new teammate, and it was too irresistible not to open the file. He opened the document and, to his amazement, discovered that John earns $10,000 more per year than him and negotiated an extra week of paid vacation.

This is a good example of an intangible cost of bad security. Bob was upset and might have acted inappropriately based on what he saw, not to mention that he could have searched for other, more interesting data on the HR server. Because he was granted access to the server, there might be no audit trail, and Bob becomes a potential problem. With CS-MARS deployed appropriately, a single report showing access to the HR database would have been a great forensics tool for threat responders to use to recognize unauthorized access.

Using CS-MARS to Justify Security Investment

Eric was a one-man security shop for his company. He was again tasked with helping his ISO/IT director create the security budget for next fiscal year. Eric had his challenges last year asking for money to purchase IPS. He was given a small portion of his requested budget and told that, from a financial perspective and with the lack of security breeches on the network, they could not justify purchasing IPS. With a little investigation and help from a security analysis firm, he was advised to look at STM. Eric convinced his ISO to use the budget money to purchase CS-MARS and a server so he could use free-ware Snort on the demilitarized zone (DMZ) where clients access their extranet. With the CS-MARS in place, Eric used the existing network to send NetFlow data to the CS-MARS and correlate with the basic Snort event data from the extranet. He was able to identify numerous security threats to the network and customers' information in a matter of only one week.

Coincidently, Eric now manages three IPS sensors, two CS-MARS 100s, and a CS-MARS Global Controller. It's still the same budget year, and the company's sales media now coin a catchy security phrase identifying strong security as a reason people should do business with the company.

The STM Deployment

Chapter 1 defined the requirements for an ideal STM as a reporting and mitigation system that reduced the time and increased the accuracy of threat mitigation, threat containment, and threat reporting.

STM can be deployed many ways in your network; the choice is determined by the requirements of your organization. For example, if your organization has several remote locations and lower-bandwidth access portals, you might choose to locate a CS-MARS box in a remote location, to reduce the amount of traffic on that slow link and ensure that the network link is available for your company's revenue-based tasks.

How you deploy the CS-MARS in your network is critical to the success of achieving your organization's goals. With the CS-MARS product, there are two types of deployment scenarios: global and standalone.

To fully understand these deployment scenarios, you must first be familiar with the CS-MARS product line. At the time of this writing, Cisco offers two types of products in the CS-MARS portfolio:

• Global Controller —A master unit that allows for global management of one or more Local Controllers.
• Local Controller —A single appliance, ranging from a CS-MARS M20 to CS-MARS M200

Table 2-1 explains the Cisco offerings for the CS-MARS product family.

Table 2-1. CS-MARS Product Portfolio

A global deployment simply means that one or more Local Controllers are reporting to the CS-MARS Global Controller. In this deployment, Local Controllers report summarized event and session data to the Global Controller in both text and graphical format over an HTTPS session. Additionally, all operations in the Local Controller now become globally manageable. A Global Controller does not do global correlation—that is, the data from each Local Controller is not correlated. You need a global CS-MARS deployment for several reasons:

• To conserve WAN bandwidth
• To log data security
• To facilitate distributed processing of event data
• To facilitate distributed management and reporting
• For high availability and to archive log retention

In a standalone deployment, all event-reporting devices send their respective log data to a single CS-MARS device. All capabilities discussed in this text are the function of the Local Controller, unless specifically indicated otherwise. This deployment is the most common for small to medium-size businesses. These are some reasons for deploying a single Local Controller:

• Cost
• Isolated (non-WAN) or local network with Internet or VPN
• Minimal number of reporting devices