Network Security. Different aspects of Network Security.

Editor's Notes

• #5: Text on the right is an example of a ‘whois’ query. It is not a good idea to name the administrative contact. News/web sites are useful for learning about different subsidiaries, staff names or positions, new merges (potentially with less security). Dumpster diving can sometimes produce internal documentation – use a shredder.
• #6: After the cracker knows something about the company, often the second stage would be to learn the network and computer configurations. War Driving: Listening with a high-powered receiver for wireless LAN signals. Tools indicate the power level, encryption type, and protocol details. War Dialing: Dials numbers within a range looking for a modem to answer. Network Mapping: Polls computers for which services they support Vulnerability Scanning Tools: Polls computers to learn services, service versions, configurations
• #7: Network Mapping = Footprinting, same as on previous page. Traffic Analysis: Does a lot of traffic go between Point A and Point B, or Point C? Is it encrypted? This might be a concern if you are the military.
• #8: Once a cracker knows the configuration of the network, it is possible to launch an attack to get in. The dog is ‘sniffing’ the login and password identification. These attacks will be defined on further slides. Note that they are of two varieties: attacks to the network, and attacks to the system.
• #9: Denial of service (DOS): Prevent service. E.g. flood a network with traffic so legitimate traffic can’t get through Spoofing: cracker alters the ‘from’ address in the packet header to look like a trusted entity Packet replay: common method of gaining unauthorized access – e.g. sniffer observes a remote logon, repeats it Message Modification: Bill changes Joe’s original message, which was intended for Ann.
• #10: 10.1.1.1 (2/3) are IP addresses The red computer here is pretending to be 10.1.1.1, and forward confidential information to 10.1.1.1.
• #11: This example shows that people can fool your generated programmed SQL statement by inserting unexpected logins and passwords. This may be done by adding conditions, additional SQL statements, or by accessing the OS command line. Always sanitize your input.
• #12: Calculation = <number of possible characters> to the <password length> power Result is maximum number of guesses needed to find the right password. This is taken from NIST, and assumes many computers are used in parallel to crack a password. Think criminal effort potentially using bots.
• #13: Once the cracker has entered, they can expand their access and hide their break-in. A RootKit hides itself in the OS. For example, when you list processes, the malware is not listed. The RootKit may delete specific logs, or open a backdoor, to enable the attacker to enter easily. A Trojan Horse is software that is useful, but hides its malware intentions. For example, a game may be passed all around the internet, but may include spyware or adware (or other malware) within it. Bots are computers that have been taken over, and are now being used by the attacker for whatever purpose they would like.
• #14: Because these networks span the world, it is very difficult for law enforcement to backtrack, find and prosecute the attacker. Multiple layers and countries can help to hide the attacker. In fact, laws vary in different countries, and law enforcement often has different priorities.
• #15: The terms ‘bot’ and ‘zombie’ are apparently interchangeable. A ‘botnet’ is a large number of bots (or zombies) used for DDOS attacks.
• #16: 2 = Distributed Denial of Service
• #17: 1
• #19: Defense in depth is like layers of an onion – to get in you must go through multiple defenses. Think of the effectiveness of multiple layers of defense years ago with the castle shown. Then consider the defenses shown for a computer on the right.
• #20: A bastion host is just a computer, server or system that is locked down against intruders. It is configured to have maximized security (strict firewall rules, well-patched) and minimized potential avenues of attack (minimal applications).
• #21: What is the easiest way to get into this network? It may not be through the firewalls. It may be through the dial-up access, CDs or DVD drives, or WLANs. Also notice that a good network will be divided into sections. The De-Militarized Zone here is for public access. The Private Network is for internal access, and requires going through 2 firewalls, each with filtering.
• #22: (From CISM) The Packet Filter may scan for source or destination IP addresses (computer IDs) and port addresses (service IDs).
• #23: A Packet Filter firewall looks at the incoming packets. Some of them may be requests for connections, or responses to our connections. Normally PCs only initiate connections, such as web or email. Therefore, web and email requests we would expect to travel in the other direction (from PC to Internet). Most of these requests are illegal. Most likely a cracker is attempting to break into a server, or a PC which is willing to act as one. Other attacks include uses of invalid IP addresses, such as an IP address representing the internal network (pretending to originate from the inside of the network). In this case, the only packets that should make it through are replies to our web requests and email requests to a mail server.
• #24: Here the red is the packet header being inspected, and the green is the part of the packet which is not inspected. When an A is displayed in the firewall, this means that the firewall has state information about each connection and can detect more anomalies. For example, connection-oriented protocols require you to connect before sending data. If data is received before the connection is established, then obviously the data is bad. In the Stateful Inspection, the state of Disconnected, Connected is maintained. In some cases, many states are possible.
• #25: Here the firewalls create separate connections with the two endpoints, thus maintaining extensive state information about each. Notice that the amount of the packet inspected (red) is a larger portion of the packet than with previous firewalls. Obviously, the best firewall would inspect all of the packet. However, the more it inspects, the more processing power the filtering requires. Thus, very good firewalls handle smaller packet volumes.
• #26: A screened host means a firewall with a border router that screens obvious attacks, such as network mapping. Multi-homed means that it has multiple zones to filter for. In this case there are 3 zones: Internet, DMZ and internal network. Notice the color scheme: Black/Brown: network security servers Green: Public services Yellow/orange: More security Red: Most secure – confidential information
• #27: ‘Rules’ means the settings on your defenses; what will the firewall allow past, what will cause the intrusion detection system (IDS) to react, etc. Rules are going to depend on the capabilities of your equipment and the goals and/or risk appetite of the organization, as reflected in policy.
• #28: This shows 3 services provided by a university, as well as the sensitivity and roles that normally access this information. We would not want to put the public web pages in the same network zone with Grades, for example.
• #29: The Path of Logical Access shows where requests enter and are processed. Two paths of logical access are shown, via brown arrows through WLAN and to server, and red arrows through laptop and server. Visitors from the internet must get through a firewall, then either the logical access controls (LAC) in the database servers in the demilitarized zone (DMZ), or through a second firewall and the LAC in the internal network’s servers. Entering via the wireless LAN bypasses all that (except for the internal LAC), as does using a disk or flash drive. The latter (wireless/portable media access) shows that this organization depends on physical controls and internal access control mechanisms (including employee trustworthiness) to prevent unauthorized use by those means. This leaves the private network server and the printer vulnerable.
• #30: Here the WLAN and dial-up interface must go through a firewall before accessing the private network – good idea!!!
• #31: Here are 4 services. The Required Contols are not fully specified yet, but will be towards the end of the presentation. Currently only security services are shown.
• #32: Notice the color coding and the zones.: Green: Public Yellow: Some confidentiality concern. Often located on a separate server from other colors and other functions. Internal rating. Orange: More security required. Private. Red: Most security required. Confidential. Sometimes two processes with different roles may have separate servers or VMwares, to minimize break-into both servers, if break-in into one occurs.
• #33: The difference between and IDS and IPS is that the IDS reports on something but does not filter it. The IPS filters and prevents attacks. An IDS may react to an attack by sending disconnect packets for a connection. While IPS definitely sounds better, the implementation may be difficult. Not all things that look like attacks are attacks – therefore, optimizing an IDS/IPS is necessary to get rid of false positives and false negatives – or normal events looking like attacks and vice versa. A HIDS is always on one computer, scanning that one computer. The NIDS monitors traffic in a network.
• #34: Anti-virus software is an example of Signature-based Software. Above you can see that for the graph, on Wed, we had some unusual traffic that needs looking into.
• #35: A Honeypot or Honeynet has no useful purpose other than to catch attackers. It may be used as a form of an IDS. While it sounds fun and interesting, they need to be maintained and monitored: if an attacker does gain entry, they now can attack from within the network.
• #36: The tools in parenthesis provide the features specified.
• #37: Symmetric encryption: each participant uses the same (shared secret) key. In the equation, P=Plaintext, E=Encryption, D=Decryption NIST = National Institute for Standards and Technology, an American department of recommendations.
• #38: Asymmetric encryption: each user has a public key and a private key. They are not easily mathematically related; that is, having the public key will not enable someone to calculate the private key. However a message encrypted with one can be decrypted with the other. The private key can also be used as a digital signature (next slide). This encryption technique can be used to send encrypted information or to authenticate a packet as originating from the sender, as shown above in the top and bottom examples, respectively. Public key encryption is a wonderful technique. However, it is processor-intensive, and not useful for longterm data communications sessions. Therefore, it is often used to provide a Secret key between two endpoints, and then the Secret key is used thereafter.
• #39: A VPN creates an encrypted point-to-point path between two computers. Here the line in red is encrypted. Often it uses Public Key Encryption to communicate a Secret Key, then uses Secret Key encryption to encrypt the session data.
• #40: Hashes implement Integrity. A message is hashed and the hash (H) is sent along with the message. When received, the message is hashed again and the two hashes are compared. Small changes to a message will result in large changes to the hash, so if the message was altered this method will detect it, although it won’t identify what those changes were. In the first case (MAC), the Hash is calculated using an associated secret key (K). In the second case (One-way hash), a standard-calculated hash is encrypted (E) using a secret key (K) Note that the message itself is not encrypted – it only gets a sophisticated checksum. MD = Message Digest SHA = Secure Hash Algorithm
• #41: A Digital Signature is used for authentication, integrity, and non-repudiation. It serves the same purpose as signing a contract with ink – but digitally. The private key is used in creating a hash of the message, which provides both integrity and nonrepudiation.
• #42: 3rd party authentication is used for authentication and non-repudiation. Steps 1-3 establish the Digital Certificate (DC). Steps 4-7 send a message which is verified using the Digital Certificate CA=Certificate Authority RA=Registration Authority
• #43: Someone dials up and wants to access our network. Should we trust him/her? No! Let’s call back to the location he/she lives and allow them access only from there. RADIUS and TACACS are well-known NAS products. Network Access Server implements an authentication security service.
• #44: HTML is stateless. That is, information about the connection and data transactions have to be held by the endpoint computers. This can be exploited by a skilled hacker. Cookies and client-side scripts are two examples. In some cases, servers do not retain state but instead send information in a request which can be manipulated by the client before being returned. This is another form of attack.
• #45: Penetration testing can test from outside the network to determine what vulnerabilities remain.
• #46: Here the different security services have been translated into specific tools for the red services.
• #48: 3 is correct. 1: It is actually called the Path of Logical Access
• #49: 2 – Application-level firewall
• #50: 3 – Digital Signature
• #51: 3 – Signature-based
• #52: 3 - Messag
• #53: Bastion host would have other requirements: up-to-date patches, applications turned off. A dual-homed firewall requires access to two networks. A screened host refers to a firewall with an external router screening it.
• #56: The Firewall will let certain locations and services enter and leave the network
• #58: Notice the color coding and the zones.: Green: Public Yellow: Some confidentiality concern. Often located on a separate server from other colors and other functions. Internal rating. Orange: More security required. Private. Red: Most security required. Confidential. Sometimes two processes with different roles may have separate servers or VMwares, to minimize break-into both servers, if break-in into one occurs.