Transcript ISACA
ISACA Presentation Network Security Fundamentals Prepared and Presented by Deloitte & Touche Strategic Security Services Introductions Your Instructor Introduce Yourself • Brief Background • Familiarity With Security Expectations • My Expectations As Your Instructor • The ONE Thing You Want To Learn Most From This Session? Session Overview Network Overview TCP/IP protocol Components of a secure network Firewall technology Encryption Overview VPNs Digital Certificates Kerberos Web security Overview SSL Redirectors and Load balancing CGI considerations Sample Files Session Overview Network Overview TCP/IP protocol Components of a secure network Firewall technology TCP/IP and Networking Application Anatomy of a TCP/IP Packet Presentation Session Transport Network Data Link Physical Source Destination MAC Address MAC Address Source IP Address Destination IP Address Source Port Number Destination Port Number Payload (Session Data) CSUM CSUM CSUM TCP/IP and Firewalls Different firewalls operate at different levels of the OSI model. Routers and some appliance based firewalls are capable of screening packets only at the Network and Transport layers of the OSI model. Proxy servers Protect networks at the application layer only. This is the highest level of protection but operates at the highest cost. This means that all functions are extremely resource intensive (requiring huge amounts of processing power, memory, and sometimes disk space). Stateful Inspection engines protect the network from the Data Link layer to the Application layer (generally focusing on the network and transport layers). Adaptive proxy is similar to Stateful Inspection but approaches the sliding security scale from the application layer. This allows you to apply higher level security to some types of transactions and lower levels of security to others. Protect The Network! But How? Zones of Containment Commonly referred to as Demilitarized Zones or DMZ’s What is the function of a DMZ? Separate the inside network from the publicly accessible network. Maintain internal network security if DMZ network gets compromised. Provide an access between sites with different levels of trust. Provide distinct layers of security (zones can only be crossed through firewalls) Why use zones Zones are utilized for the following reasons: • To restrict access between trusted and non-trusted networks • To provide Internet access to services without compromising the security of your internal network • To mitigate risk between distinct networks (i.e. Between development, test, and production networks) • To allow vendors or partners access to data on your network but not within your production environment • To segregate the different divisions within your own production network (i.e. HR, Development, Procurement, Finance) Why use zones Zones are a part of the concept of layered security. •The goal of a “layered” approach is to have the most sensitive network nested as far away as possible from the Internet. •Each network is separated by a firewall. It is strongly recommended that a heterogeneous firewall implementation be employed. •Access to more internal networks should be authenticated and/or encrypted according to the sensitivity of the data contained on the network. Zone Model This is an example of how a firewall would be employed to create distinct security zones or layers of security. Remember that this is only an example and should not be considered the only way to achieve high levels of security. Internet Intrusion Detection System Encrypted Authenticated Mail Relay HTTP CVP DNS IDS FireWall “The DMZs” VPN’s Protected Web Application Servers E-Commerce Application PKI DC Protected servers Address Translation Network Address Translation, commonly referred to as NAT, is a mechanism for reducing the need for globally available network addresses. Address Translation allows organizations without globally unique addresses to connect to the internet by translating those non-unique addresses into “public” addresses. Address translation is simply the process of exchanging an IP address for another. This can be done in a variety of ways but is most often handled by a firewall or router. The firewall or router handles the process of changing the IP address of the packets as they enter and leave the protected network space. This is done through the use of a dynamically generated table called a NAT table. The NAT table is simply a list of internal and external address to address or address to port mappings. How NAT works 207.250.227.19 (Legal) 10.10.10.0 (Illegal) 1 Mail server (email.ctl.com) fw.ctl.com Outside Inside 2 Internet Web server (www.ctl.com) Router Router Firewall’s Translation Table NAT Mode Outside (Legal) Inside (Illegal) Name Static 207.250.227.19 10.10.10.2 email.ctl.com Static 207.250.227.20 10.10.10.3 www.ctl.com Hide 207.250.227.21 10.10.11.0 ctl-net ctl-net (10.10.11.0) Use of NAT RFC 1918 has reserved a set of IP network addresses that can be used for internal networks. These are: 1 Class A Network Number: 10.0.0.0 16 Class B Network Numbers: 172.16-31.0.0 256 Class C Network Numbers: 192.168.0-255.0 Internal networks with RFC 1918 network numbers can reach all internal hosts and the Internet by implementing NAT Advantages of RFC 1918 • Additional Security — The addresses are not routeable on the Internet • Can Demand Quality Customer Service — If an ISP is not providing the quality of service desired, renumbering the network can be as few as (2 to 3 devices). What Is A Firewall? Firewall: a device or set of devices that is used to implement & enforce a written security policy regarding communication between protected and unsecure networks A Firewall acts as a control portal between a protected network and an unsecured network. Firewalls restrict the entrance and exit of traffic based on acceptability. Trusted Networks Untrusted Networks & Servers Firewall Router Intranet Server Segment Untrusted Users Public Accessible Servers & Networks DMZ Trusted Users What Is A Firewall? A Firewall cannot protect you from: malicious authorized users connections that don’t go through it all threats New ways to break through networks are continually developed. To combat this, Firewall vendors continually develops and distributes new methods of protection against unauthorized network access. A Firewall is only as effective as the security policy it supports and the networks it protects. Home Grown vs. Commercial Firewalls Is it possible to build your own firewall? The answer is yes. Is it cost effective and the right solution for you? That depends on your business model and the type of protection you are working towards. There are currently dozens of firewall products available as shareware, freeware, portions of an OS or source code. These solutions can even out perform off-the-shelf firewall products. The drawback to using a “Home Grown” solution is in the lack of extensibility and the need for highly specialized staff. Commercial firewall solutions may be expensive but they offer three major benefits over “Home Grown” firewalls: They are highly extensible They are supported by the Vendor There are a greater number of trained resources available to build and maintain them What makes the firewall safe Firewalls are safe simply because they are extremely specialized systems that are built specifically to restrict access. A firewall is not necessarily any more secure that your average server though. Do not assume that simply because an organization employs a firewall at the network border that they are safe. Just like any other network device, a firewall may be misconfigured. Types of firewalls that are especially susceptible to misconfigurations are appliance based firewalls (especially those that only allow HTTP based administration) and routers. In most cases however, firewalls are proof against standard Denial Of Service attacks. This is often built into the source code of the firewall product (whether it is an appliance, application based, or hybrid). Routers/Access Lists Description Commonly referred to as “Packet Filtering”, packets are examined at the network layer only. No information in the upper four layers is reviewed. Pro’s • Transparent to users • Low performance impact (varies with size of ACL and router) • Inexpensive Con’s • Difficult to configure, monitor, and maintain • Limited ability to manipulate data • Looks at only a small part of packet • Poor logging and alerting abilities • Vulnerable to various IP level attacks Proxy Servers Description Also known as an application gateway. Functions at the Application layer (7) of the OSI model. These are good for certain smaller environments without complex communications needs. Pro’s • Provide good security • Examine the upper layers of the packet Con’s • Usually not transparent to users • Performance isn’t as good as other solutions due to OSI layer implementation • Doesn’t examine information in lower layers of packet • Vulnerable to application and operating system bugs • Can not fully support all communications services • Break the client server model of communication Stateful Inspection Engines Description A firewall technology introduced by CheckPoint to overcome the failings of proxy servers and packet filters. Examines data from “all” layers of the OSI model. Maintains tables regarding state of communication sessions and application sessions. Pro’s • • • • Good security Good performance Transparent to users Maintains information on state of communications with data from multiple OSI layers Con’s • • Relies on Operating System to function. Expensive in both initial cost and specialized training requirements Adaptive Proxy Description Developed by Network Associates, Adaptive Proxy technology protects both in-bound and out-bound services while supporting high throughput rates by authenticating the first packet at the application layer and then passing all additional packets in the session at the network layer. Pro’s • • Faster than traditional proxy servers More secure than pure packet filtering technologies Con’s • Vulnerable to OS level security flaws Firewall Appliance Description A special purpose piece of equipment that provides firewall functionality without requiring a specific operating system such as Unix, Linux, or NT Pro’s • • • Faster than most Proxy servers Easy to configure Not vulnerable to OS specific vulnerabilities Con’s • • • Not as extensible as application based firewalls Typically have a fixed configuration Some standard features for application firewalls may be treated as add-ons by the appliance vendors Session Overview Encryption Overview VPNs Digital Certificates Kerberos The Need for VPNs London Office Mobile User Remote Office New York Office Private WAN Corporate Headquarters Building private wide-area networks to accommodate organizations is expensive and provides little flexibility Remote Access Using VPNs London Office Remote Office VPN New York Office Internet Mobile User Corporate Headquarters How Encryption Works Step 1: The original data (cleartext) is passed through an encryption algorithm that uses the secret key to uniquely scramble the data. Step 2: The result is called ciphertext. Step 3: The VPN receives the cipher text and uses a secret key to decrypt the text. VPN Types Firewall-to-Firewall Data is encrypted when it leaves Firewall #1 and crosses the Internet The data is authenticated and decrypted when it reaches Firewall #2. Firewall Module #2 Firewall Module #1 Payroll Not Encrypted PRIVATE Encrypted PUBLIC Not Encrypted PRIVATE Sales VPN Types Client-to-Firewall Firewall or Gateway With Encryption module Client with Encryption package installed Symmetric Encryption Shared Secret key is a secret decryption format needed to encrypt and decrypt data Primarily for faster encryption performance Keys must be kept secret and should be changed periodically Cleartext Message This is the original text before encryption Ciphertext DES | RC4 sdfklj98a475 $56jhgv9845 6vjnf84576F GHH78lfkghj 506#6lkjg4# $5;lkn;t7;lsk % Cleartext Message DES | RC4 This is the original text before encryption Asymmetric Encryption Pros Each node uses two mathematically related keys: a public key and a private key The private key is not derivable from the public key, hence the public key is freely distributed Allows: Computation of shared secrets over insecure channels (Diffie-Hellman) Digital signature (RSA) Public key encryption: each node publishes a public key. Anyone wishing to send an encrypted message to the node encrypts it using that key. Only the holder of the private key can decrypt the message Cons Up to 1000 times slower than symmetric cryptography Typically used to encrypt small amounts of data (e.g. shared keys) What Should Be Encrypted? Should an encryption method encrypt packet header or data? Tunneling-Mode vs. In-Place Encryption Tunneling-Mode Encrypts packet, then encapsulates packet within the encryption protocol header In-Place Encrypts the payload portion of the packet and leaves the header intact. Allows for greater performance than that provided by Manual IPSec, ISAKMP/Oakley (IKE) or SKIP encryption. Digital Signatures Digital Signature: a code that can be attached to an electronically transmitted message uniquely identifying the sender. Guarantees that the individual sending the message really is who they claim to be. Important for electronic commerce and is a key component of most authentication schemes. Digital signatures must be unforgeable. One-Way Hash Function Certificate Authority Certificate Authority (CA): a trusted third party from whom a public key can be obtained reliably, even via the Internet. CA certifies a public key by generating a certificate. The digital signature acts as proof of sender’s identity. Digital signature is created using a public encryption key scheme. Encryption and the Audit Sniffing: the process of stealing data from the network by setting your Network Interface Card into “promiscuous” mode. Sniffing allows you to steal data from the data stream by telling your NIC that all traffic on the local network should be reviewed. If this traffic is encrypted then you will need to acquire the appropriate “cracking” technology. If the Encryption Algorithm is strong enough it will not be worth the effort of cracking. Man in the Middle: If you perform a “man in the middle” test for Gateway to Gateway VPNs, make sure that you are able to sniff from the same point that the packets reach the Firewall. This means that you will plug into the network on the same segment as the Firewall’s external interface. Encryption and the Audit (Cont.) Misconceptions: A client may assume that simply because the have a VPN or use encryption that they are safe. This is not true. You must determine where encryption is used and how. For example a client may use SSL into their environment but have clear text transaction between their Web and Database servers. This may protect them from most external threats to transaction sniffing, but does nothing against internal threats. Kerberos Benefits Kerberos is an authentication system. Kerberos can be used to authenticate users or services (principals). A principal is defined by these components: Primary Name Instance Realm The Principal is used to identify users or services within distinct domains. For instance a users primary name would be their login (with a null instance), while a service would likely utilize service name and machine name (i.e. rlogin.machinename). Principals obtain tickets from kerberos servers. Each ticket contains identifying information for the principal as well as encryption information. Once a session is established all subsequent transactions can be encrypted. Kerberos applies a specific lifetime to each ticket. Once this has exceeded a new ticket must be requested from the kerberos server. Kerberos (continued) Disadvantages Kerberos was originally designed to authenticate end-users to a selected number of servers. The Authentication structure however was not designed with overall network security in mind. The greatest issue is in key storage. Most of the workstations that would utilize kerberos do not have a secure location for key storage. The initial ticket-granting dialog is initiated with a plain text key. This must be stored in a secure location. If this key is compromised, the ticket granting server can then be compromised by utilizing data contained within the key. Additionally there is an issue with multiuser workstations/servers. If a workstation/server supports multiple simultaneous users, then the cached key information for one user can be gained by another with relative ease. For a list of current and past risks with kerberos please refer to the “security bugware” web site. Session Overview Web Sites Internal & External Web security Overview SSL Redirectors and Load balancing CGI considerations Sample Files SSL Secure Socket Layer provides transport layer protection through the use of a variant of the TCP socket interface and encryption. SSL is usually bundled with an application (such as a web server) so that integration with the underlying protocol stack is eased. SSL was originally developed by Netscape as a part of their overall security package. The SSL protocol specification has since been utilized by other web server vendors and remains a standard. Additional information on IPSEC protocols may be located at the following sites: ftp://ftp.internic.net/rfc/rfc1825.txt ftp://ftp.internic.net/rfc/rfc1826.txt ftp://ftp.internic.net/rfc/rfc1827.txt - Security Architecture for the Internet Protocol - IP Authentication Header - IP Encapsulating Security Payload [ESP] Usage of SSL Secure Socket Layer meets the following security objectives: Protects transactions against attack on the Internet - SSL protects against many common network based attacks. While it cannot protect against DOS attacks it can protect against data manipulation and spoofing. Ensures security without prior arrangements between customers and vendors SSL provides a mechanism to verify identity for both customers and vendors. This allows transactions between entities that have no previously established relationships. Applies cryptographic protection selectively - SSL allows the implementation of cryptography in a selective manner. Only sessions that need the added layer of security invoke SSL through the selected protocol interface and socket. Protects the receiving host from attacks by incoming messages - While SSL cannot protect against actual attacks on the machines hosting or utilizing the service, it can protect the service (data stream) in use. Attacks on SSL Secure Socket Layer protects against most common attacks. However, there are still a number of attack methods that can work. Most of these have been addressed by more recent versions of the protocol, however new attacks are released regularly. Predictable Keys - Early versions of SSL generated keys based on a small amount of internal information that could be used to assist with key prediction. This data was similar to having access to the random seed used to generate encrypted passwords in a shadow file. Man in the Middle - SSL stops standard MIM attacks by validating the public key certificate before using its public key. This does not prevent the switching of valid certificates with subverted sites. The latest SSL implementation will extract the host name from the server’s certificate and compare it to hostname in the URL. Short checksum keys - Early versions of SSL utilized the same key for both encryption and computing a keyed hash on the data. This allowed a successfully cracked key to be used to forge data sent from that key. Newer versions use different keys for encryption and data integrity. Replay - While SSL originally protected against session replay it did not prevent the use of captured session data from being used to extend a session. It was therefore possible to hijack an SSL session and bump the valid client and extend the session with captured session data. SSL 3.0 incorporates sequence numbers to prevent this sort of attack. SSL - Additional Information For additional information on SSL please refer to the following: Network Security: Private Communication in a Public World - Kaufman, Perlman, and Speciner - Overview of mathematics and public key cryptography The Public Key Cryptography Standards (PKCS) - RSA Laboratories -Standards on how to implement public key cryptography facilities “On the Difficulty of Factoring” - Rivest - a set of estimates on the difficulty of cracking an RSA key SSL 3.0 Specification - Netsacpe Communications http://www.netscape.com/libr/ssl/ssl3/index.html - The official definition of SSL. Take note of the appendix on attacks. The Need for Load Balancing ? • Web and FTP servers may see a tremendous amount of requests in a short period of time. • Often, a single system cannot effectively handle the load. ? Internet ? Overloaded Server becomes bottleneck FTP Server ? How Load Balancing Works Load balancing distributes the requests among a group of mirrored servers. ! Requests Balanced Between Mirrored Servers ! target1.client.com Client requests to target.client.com target2.client.com Internet target3.client.com ! ! target4.client.com Logical Servers Server 1 Server 2 Firewall acts as the logical server Packets flow to the firewall Firewall distributes network traffic among its server group Server 3 Load Balancing Components Load Balancing Daemon Directs client packets to a server Load Balancing Algorithms Determine which physical server will fulfill request Problems with Load Balancing Problem: You wish to hide the true IP address of the physical server to which your HTTP redirect rule directs HTTP traffic. HTTP load balancing may rewrite the HTTP logical server’s name when you tie several logical server names to one IP address. The HTTP protocol has a feature that uses a server’s name in the HTTP request. HTTP load balancing rewrites the logical server name to the actual physical server it represents. Test: Attempt to browse to the client web server and watch for the redirect. If you get a response other than the stated Web server then the Load Balancing tool is misconfigured and is leaking information. How This Affects the Audit Since Load Balancing acts as a broker for requests to certain services, you will not be able to target all of the Internet accessible servers with an external audit. Requests Balanced Between Mirrored Servers target1.client.com Client requests to target.client.com Testing from across the Internet target2.client.com target3.client.com ! target4.client.com CGI Scripts The problem with CGI scripts is that each one presents yet another opportunity for exploitable bugs. CGI scripts can present security holes in two ways: They may intentionally or unintentionally leak information about the host system that will help hackers break in. Scripts that process remote user input, such as the contents of a form or a "searchable index" command, may be vulnerable to attacks in which the remote user tricks them into executing commands. CGI Scripts CGI scripts are potential security holes even though you run your server as “nobody.” A subverted CGI script running as “nobody” still has enough privileges to mail out the system password file, examine the network information maps, or launch a log-in session on a high numbered port (it just needs to execute a few commands in Perl to accomplish this). Even if your server runs in a chroot directory, a buggy CGI script can leak sufficient system information to compromise the host. Safe CGI Development Avoid giving out too much information about your site and server host. If you're coding in a compiled language like C, avoid making assumptions about the size of user input. Never pass unchecked remote user input to a shell command. JAVA Nothing in life is completely secure; Java is no exception. If you're using an up-to-date Web browser, you are usually safe against the known attacks. However, nobody is safe against attacks that haven't been discovered yet. Other Web “scripting” tools such as JavaScript, Visual Basic Script, or ActiveX face the same sorts of problems as Java. “Plug-in” mechanisms provide no security protection. If you install a plug-in, you're trusting that plug-in to be harmless. JAVA Attacks There are two classes of security problems: nuisances and security breaches A nuisance attack merely prevents you from getting your work done. For example, it may cause your computer to crash. Security breaches are more serious: your files could be deleted, your private data could be read, or a virus could infect your machine. Java, ActiveX and Java Script • Moving the Risk from Server Processing to Client Processing • Alternative to CGI Java • Programming Language • Platform Independent • Built in Security Manager – Will not allow Java to open IP Connections – Will not read/write to local disk – Applets downloaded controlled by the security manager The Real Java • • • • Applets can open IP connections Applets can execute binary code Applets can bypass security manager Denial Service Attacks common (Looping) ActiveX • OLE for the Internet • NO Security model • Certificate Trust model – Even bad people can have a Drivers License • Full control of the system • No Audit trail Java Script • Scripting language – Code in HTML files • Ability to control the browser • Security not designed within • Security by feature removal So what is unsafe with my browser? Reading your Private Files Making you do something that you really should not do What files are private? A few files provide a lot of information about you. These include: Cache files History file Bookmarks Configuration Important Files History File Since the default is 30 days to expire a link, typically you can see the last 30 days worth of web surfing by examining the history file. Bookmark File Bookmarks are a problem for the same reason the history file is a problem. It shows what sites you feel that are important. Cache The cache is your browser's way of making things faster. Every query is stored in cache. Typically every form submittal including accesses to pages requiring an ID and password will be there, unless a site has tagged an HTML document NOT to be cached. What can my Web browser make me do? You can be tricked into supplying user IDs and passwords, sending personal information like Social Security numbers and credit card information. If your browser supports HTML 3.0 extensions and Java, your files could be plucked from your hard drive. Your machine could be used to terrorize other resources behind your firewall and send critical information offsite. The Basic FireWall Design Security & Phase 1 Performance Internet Encrypted Authenticated VPN’s “The DMZ” Intrusion Detection Monitor Protected Web Application Servers IDM FireWall Mail - HTTP -FTP - DHCP - ACE - E-Commerce Relay DNS CVP PC Banking Network Address Translation Centralized Management Security Policies Router Control User Access Scalable Internal Resources Mainframe - SAP - People Soft - Oracle - Internal Mail - Etc. Layer 1 The Layered FireWall Design Enhanced Containment Internet Encrypted Authenticated “The DMZ” Intrusion Detection Monitor IDM FireWall VPN’s Protected Web Application Servers Mail - HTTP - DHCP - ACE - E-Commerce Relay DNS CVP PC Banking Centralized Management Security Policies Router Control User Access Scalable Layer 1 Batch Data Bases FTP - SQL- Access - Oracle - Etc. Layer 2 Intra-Wall Internal Resources Mainframe - SAP - People Soft - Oracle - Internal Mail - Etc. Layer 3 The Perimeter Design for Performance, Reliability, and Security Internet OUT-Bound Intrusion Detection Monitor Encrypted Authenticated IN-Bound Protected Web Application Servers “The DMZ” IDM FireWall IDM State Sync Centralized Management Bandwidth Management FW & Svr Load Balancing FW High Availability VRRP Scalable VPN’s Mail - HTTP - DHCP - ACE - E-Commerce Relay DNS VRRP IDM Layer 1 Batch Data Bases FTP - SQL- Access - Oracle - Etc. Layer 2 Intra-Wall IDM Firewall Reporting Engine Layer 3 Internal Resources Mainframe - SAP - People Soft - Oracle - Internal Mail - Etc. So What Can I Do Now? • On My Own – Install Firewalls & Intrawalls properly – Install Security Monitoring – Audit Security Logs – Acquire Scanning Tools & Lock Down Platforms – Perform Security Testing – Attend Security Training • From Security Experts – Security Verification Assistance • Not all third parties have the same expertise! – Tiger Team Analysis – Security Training • Firewalls • Attack Monitoring • Scanning Tools • Vulnerabilities Security Skills Improvement Recommendations • Entry Level Training – OS Administration Training – Scanning Tool Certification – Firewall Certification • Advanced Level Training – OS Security Lock Down Training – FireWall Lock Down – Compromise Response Initiatives Management Awareness • Security Awareness Seminar Security Issues & Initiatives Review Architecture & Design for Unprotected Access – Scan network for all available network routes & security short cuts – Redesign Security Architecture & Layer FireWall- Security Controls Scan all IP Devices for Vulnerabilities – Remote & Internal Testing Internet Scanner TM Lock Down all Security Holes – Operating Systems – Applications – Files & Data Bases Intranet Firewall Web Scanner Scanner Security – Networks & Access Points Scanner Develop & Document Security Baselines & Policies Resolve Security Vulnerabilities & Re-test all Targets Implement On-line Security Monitoring & Auditing Develop & Document Security Baselines & Policies Resolve Security Vulnerabilities & Re-test all Targets Implement On-line Security Monitoring & Auditing Adaptive Security Management Monitor The ability to monitor, detect, and respond to threat and vulnerability conditions. Respond 7/16/2015 Detect Intrusion Detection & Attack Recognition Corporate Network Active Attack Monitoring Active Attack Monitoring Internet Intranet Extranet Vulnerability Detection & Response Manufacturing Engineering DMZ • E-Mail • File Transfer • HTTP Comms Server Internet Route r Network Marketing Intranet Human Resources Corporate Network Internet Scanner Vulnerability Detection & Response Manufacturing Engineering DMZ • E-Mail • File Transfer • HTTP Comms Server Internet Route r Applications Marketing Intranet Human Resources Corporate Network Internet Scanner Vulnerability Detection & Response Manufacturing Engineering DMZ • E-Mail • File Transfer • HTTP Comms Server Route r Internet Systems Marketing Intranet Human Resources Corporate Network System Security Scanner Threat Detection & Response Manufacturing Engineering ALERT! Record Session, Send Message, Terminate Marketing Connection DMZ • E-Mail • File Transfer • HTTP Attack Recognitio n Route r Extranet Business Partner Intranet Human Resources RECONFIGURE Router or Firewall to Block IP Address Corporate Network Extranet External Threats Terminate Connection Manual Tests • http://www.target.com/msadc/Samples/SE LECTOR/showcode.asp?source=/msadc/S amples/../../../../../boot.ini • ftp://www.target.com - often displays the Web root as a set of files • IIS character buffering on open ports