Why Should I Care About Network Security?
A company's network is like any other corporate asset: it has value, it is directly related to the success and revenue of that company, and, as such, it must be protected. One of the primary concerns of network administrators is the security of their network. Security attacks can range from malicious attacks to theft of information to simple misuse of company resources. Estimated losses attributed directly to network intrusions totaled more than $15 billion for 2001.
According to the FBI, the number of network attacks doubled from 2000 to 2001. They are expected to increase another 100 to 150 percent in 2002. It is believed that less than 50 percent of intrusions are actually reported. The majority of unauthorized access and resource misuse continues to come from internal sources. In addition, attacks from external sources continue to grow in number as less sophisticated hackers gain access to information and power tools designed for hacking. The figure above demonstrates this change.
What Are the Problems to Solve?
Security must be an inherent part of every network design based on the principles of protecting from the outside (perimeter security) and controlling the inside (internal security). In other words, keep the outsiders out, and keep the insiders honest. You should think of the network performing the dual roles of "gatekeeper" (perimeter) and a "hall monitor" (internal).
Balancing Trust and Security
Security and trust are opposing concepts. Trust is necessary for applications to run, but open access can expose a network to attacks or misuse. On the other hand, a very restrictive security policy might limit exposure but also reduce productivity. When security is a primary design consideration, you can determine a trust boundary on a per-user basis and strike the proper balance.
The first part of any security design is determining who is on the network. Without some knowledge of who the users are, you would have to make the network policies generic so they would likely be too open or too restrictive. Identity can include
User identity based on password, smart card, fingerprint, etc.
Device identity (e.g. IP phone) based on Internet Protocol (IP) or Media Access Control (MAC) address.
Application identity based on IP address or Transmission Control Protocol/User Datagram Protocol (TCP/UDP) port number.
Identity is tightly linked with authentication. After you establish identity, you can apply, monitor, and enforce the proper policy for that user, device, or application.
Perimeter security refers to controlling access to critical network applications, data, and services so that only legitimate users and network information can pass through the network. You typically control access with access control lists (ACLs) enabled on edge routers and switches, as well as with dedicated firewall appliances. A firewall is a device that permits only authorized traffic to pass (according to a predefined security policy). Other tools, such as virus scanners, content filters, and intrusion detection systems (IDSs), also help control traffic.
As networks grow in size and complexity, the requirement for centralized policy management grows as well. Regardless of the existence of sophisticated tools, companies must employ a sound policy with clear guidelines for enforcement. Generally, middle- to large-size companies appoint a chief security officer whose job is to develop and enforce corporate security policies.
Much of the information passing through a network is confidential. Whether it is business-specific (engineering or financial) or personnel (human resources correspondence) information, it must be protected from eavesdropping. You can implement encryption and data privacy schemes in Layer 2 (Layer 2 Tunnel Protocol (L2TP)) or Layer 3 (IP Security (IPSec) for encryption), Multiprotocol Label Switching (MPLS) for data privacy). This type of protection is especially important when implementing virtual private networks (VPNs).
Enabling security measures in a network is not enough. Network administrators must regularly test and monitor the state of security solutions. Using a combination of network vulnerability scanners and IDSs, the network administrator can monitor and respond to security threats in real time.
Top 13 Security Vulnerabilities
Inadequate router access control.
Unsecured and unmonitored remote access points, providing easy access to corporate networks.
Information leakage revealing operatingsystem and application information.
Hosts running unnecessary services.
Weak, easily guessed, and reused passwords.
User or test accounts with excessive privileges.
Misconfigured Internet servers, especially for anonymous FTP.
Software that is outdated, vulnerable, or left in default configurations.
Lack of accepted and well-promulgated security policies, procedures, guidelines, and minimum baseline standards.
Excessive trust domains in UNIX and NT environments, giving hackers unauthorized access to sensitive systems.
Unauthenticated services such as the X Window System.
Inadequate logging, monitoring, and detection capabilities.