Глава 13. Masquerading and Firewalls

Содержание

13.1. Packet Filtering with iptables
13.2. Masquerading Basics
13.3. Firewalling Basics
13.4. SuSEfirewall2
13.5. For More Information

Whenever Linux is used in a network environment, you can use the kernel functions that allow the manipulation of network packets to maintain a separation between internal and external network areas. The Linux netfilter framework provides the means to establish an effective firewall that keeps different networks apart. With the help of iptables—a generic table structure for the definition of rule sets—precisely control the packets allowed to pass a network interface. Such a packet filter can be set up quite easily with the help of SuSEfirewall2 and the corresponding YaST module.

13.1. Packet Filtering with iptables

The components netfilter and iptables are responsible for the filtering and manipulation of network packets as well as for network address translation (NAT). The filtering criteria and any actions associated with them are stored in chains, which must be matched one after another by individual network packets as they arrive. The chains to match are stored in tables. The iptables command allows you to alter these tables and rule sets.

The Linux kernel maintains three tables, each for a particular category of functions of the packet filter:

filter

This table holds the bulk of the filter rules, because it implements the packet filtering mechanism in the stricter sense, which determines whether packets are let through (ACCEPT) or discarded (DROP), for example.

nat

This table defines any changes to the source and target addresses of packets. Using these functions also allows you to implement masquerading, which is a special case of NAT used to link a private network with the Internet.

mangle

The rules held in this table make it possible to manipulate values stored in IP headers (such as the type of service).

These tables contain several predefined chains to match packets:

PREROUTING

This chain is applied to incoming packets.

INPUT

This chain is applied to packets destined for the system's internal processes.

FORWARD

This chain is applied to packets that are only routed through the system.

OUTPUT

This chain is applied to packets originating from the system itself.

POSTROUTING

This chain is applied to all outgoing packets.

Рисунок 13.1, «iptables: A Packet's Possible Paths» illustrates the paths along which a network packet may travel on a given system. For the sake of simplicity, the figure lists tables as parts of chains, but in reality these chains are held within the tables themselves.

In the simplest of all possible cases, an incoming packet destined for the system itself arrives at the eth0 interface. The packet is first referred to the PREROUTING chain of the mangle table then to the PREROUTING chain of the nat table. The following step, concerning the routing of the packet, determines that the actual target of the packet is a process of the system itself. After passing the INPUT chains of the mangle and the filter table, the packet finally reaches its target, provided that the rules of the filter table are actually matched.

Рисунок 13.1. iptables: A Packet's Possible Paths

iptables: A Packet's Possible Paths

13.2. Masquerading Basics

Masquerading is the Linux-specific form of NAT (network address translation). It can be used to connect a small LAN (where hosts use IP addresses from the private range—see Раздел “Netmasks and Routing” (Глава 9, Basic Networking, ↑Содержание)) with the Internet (where official IP addresses are used). For the LAN hosts to be able to connect to the Internet, their private addresses are translated to an official one. This is done on the router, which acts as the gateway between the LAN and the Internet. The underlying principle is a simple one: The router has more than one network interface, typically a network card and a separate interface connecting with the Internet. While the latter links the router with the outside world, one or several others link it with the LAN hosts. With these hosts in the local network connected to the network card (such as eth0) of the router, they can send any packets not destined for the local network to their default gateway or router.

[Important]Using the Correct Network Mask

When configuring your network, make sure both the broadcast address and the netmask are the same for all local hosts. Failing to do so prevents packets from being routed properly.

As mentioned, whenever one of the LAN hosts sends a packet destined for an Internet address, it goes to the default router. However, the router must be configured before it can forward such packets. For security reasons, this is not enabled in a default installation. To enable it, set the variable IP_FORWARD in the file /etc/sysconfig/sysctl to IP_FORWARD=yes.

The target host of the connection can see your router, but knows nothing about the host in your internal network where the packets originated. This is why the technique is called masquerading. Because of the address translation, the router is the first destination of any reply packets. The router must identify these incoming packets and translate their target addresses, so packets can be forwarded to the correct host in the local network.

With the routing of inbound traffic depending on the masquerading table, there is no way to open a connection to an internal host from the outside. For such a connection, there would be no entry in the table. In addition, any connection already established has a status entry assigned to it in the table, so the entry cannot be used by another connection.

As a consequence of all this, you might experience some problems with a number of application protocols, such as ICQ, cucme, IRC (DCC, CTCP), and FTP (in PORT mode). Web browsers, the standard FTP program, and many other programs use the PASV mode. This passive mode is much less problematic as far as packet filtering and masquerading are concerned.

13.3. Firewalling Basics

Firewall is probably the term most widely used to describe a mechanism that provides and manages a link between networks while also controlling the data flow between them. Strictly speaking, the mechanism described in this section is called a packet filter. A packet filter regulates the data flow according to certain criteria, such as protocols, ports, and IP addresses. This allows you to block packets that, according to their addresses, are not supposed to reach your network. To allow public access to your Web server, for example, explicitly open the corresponding port. However, a packet filter does not scan the contents of packets with legitimate addresses, such as those directed to your Web server. For example, if incoming packets were intended to compromise a CGI program on your Web server, the packet filter would still let them through.

A more effective but more complex mechanism is the combination of several types of systems, such as a packet filter interacting with an application gateway or proxy. In this case, the packet filter rejects any packets destined for disabled ports. Only packets directed to the application gateway are accepted. This gateway or proxy pretends to be the actual client of the server. In a sense, such a proxy could be considered a masquerading host on the protocol level used by the application. One example for such a proxy is Squid, an HTTP and FTP proxy server. To use Squid, the browser must be configured to communicate via the proxy. Any HTTP pages or FTP files requested are served from the proxy cache and objects not found in the cache are fetched from the Internet by the proxy.

The following section focuses on the packet filter that comes with openSUSE. For further information about packet filtering and firewalling, read the Firewall HOWTO included in the howto package. If this package is installed, read the HOWTO with

less /usr/share/doc/howto/en/txt/Firewall-HOWTO.gz

13.4. SuSEfirewall2

SuSEfirewall2 is a script that reads the variables set in /etc/sysconfig/SuSEfirewall2 to generate a set of iptables rules. It defines three security zones, although only the first and the second one are considered in the following sample configuration:

External Zone

Given that there is no way to control what is happening on the external network, the host needs to be protected from it. In most cases, the external network is the Internet, but it could be another insecure network, such as a WLAN.

Internal Zone

This refers to the private network, in most cases the LAN. If the hosts on this network use IP addresses from the private range (see Раздел “Netmasks and Routing” (Глава 9, Basic Networking, ↑Содержание)), enable network address translation (NAT), so hosts on the internal network can access the external one. All ports are open in the internal zone. The main benefit of putting interfaces into the internal zone (rather than stopping the firewall) is that the firewall still runs, so when you add new interfaces, they will be put into the external zone by default. That way an interface is not accidentally «open» by default.

Demilitarized Zone (DMZ)

While hosts located in this zone can be reached both from the external and the internal network, they cannot access the internal network themselves. This setup can be used to put an additional line of defense in front of the internal network, because the DMZ systems are isolated from the internal network.

Any kind of network traffic not explicitly allowed by the filtering rule set is suppressed by iptables. Therefore, each of the interfaces with incoming traffic must be placed into one of the three zones. For each of the zones, define the services or protocols allowed. The rule set is only applied to packets originating from remote hosts. Locally generated packets are not captured by the firewall.

The configuration can be performed with YaST (see Раздел 13.4.1, «Configuring the Firewall with YaST»). It can also be made manually in the file /etc/sysconfig/SuSEfirewall2, which is well commented. Additionally, a number of example scenarios are available in /usr/share/doc/packages/SuSEfirewall2/EXAMPLES.

13.4.1. Configuring the Firewall with YaST

[Important]Automatic Firewall Configuration

After the installation, YaST automatically starts a firewall on all configured interfaces. If a server is configured and activated on the system, YaST can modify the automatically-generated firewall configuration with the options Open Ports on Selected Interface in Firewall or Open Ports on Firewall in the server configuration modules. Some server module dialogs include a Firewall Details button for activating additional services and ports. The YaST firewall configuration module can be used to activate, deactivate, or reconfigure the firewall.

The YaST dialogs for the graphical configuration can be accessed from the YaST Control Center. Select Security and Users+Firewall. The configuration is divided into seven sections that can be accessed directly from the tree structure on the left side.

Start-Up

Set the start-up behavior in this dialog. In a default installation, SuSEfirewall2 is started automatically. You can also start and stop the firewall here. To implement your new settings in a running firewall, use Save Settings and Restart Firewall Now.

Рисунок 13.2. The YaST Firewall Configuration

The YaST Firewall Configuration

Interfaces

All known network interfaces are listed here. To remove an interface from a zone, select the interface, press Change, and choose No Zone Assigned. To add an interface to a zone, select the interface, press Change and choose any of the available zones. You may also create a special interface with your own settings by using Custom.

Allowed Services

You need this option to offer services from your system to a zone from which it is protected. By default, the system is only protected from external zones. Explicitly allow the services that should be available to external hosts. After selecting the desired zone in Allowed Services for Selected Zone, activate the services from the list.

Masquerading

Masquerading hides your internal network from external networks (such as the Internet) while enabling hosts in the internal network to access the external network transparently. Requests from the external network to the internal one are blocked and requests from the internal network seem to be issued by the masquerading server when seen externally. If special services of an internal machine need to be available to the external network, add special redirect rules for the service.

Broadcast

In this dialog, configure the UDP ports that allow broadcasts. Add the required port numbers or services to the appropriate zone, separated by spaces. See also the file /etc/services.

The logging of broadcasts that are not accepted can be enabled here. This may be problematic, because Windows hosts use broadcasts to know about each other and so generate many packets that are not accepted.

IPsec Support

Configure whether the IPsec service should be available to the external network in this dialog. Configure which packets are trusted under Details.

There is another functionality under Details: IPsec packets are packed in an encrypted format, so they have to be decrypted and you can configure the way the firewall will handle the decrypted packets. If you select Internal Zone, the decrypted IPsec packets will be trusted as if they came from the Internal Zone - although they could possibly come from the external one. Choose Same Zone as Original Source Network to avoid this situation.

Logging Level

There are two rules for logging: accepted and not accepted packets. Packets that are not accepted are DROPPED or REJECTED. Select from Log All, Log Only Critical, or Do Not Log Any.

Custom Rules

Here, set special firewall rules that allow connections, matching specified criteria such as source network, protocol, destination port, and source port. Configure such rules for external, internal, and demilitarized zones.

When finished with the firewall configuration, exit this dialog with Next. A zone-oriented summary of your firewall configuration then opens. In it, check all settings. All services, ports, and protocols that have been allowed and all custom rules are listed in this summary. To modify the configuration, use Back. Press Finish to save your configuration.

13.4.2. Configuring Manually

The following paragraphs provide step-by-step instructions for a successful configuration. Each configuration item is marked as to whether it is relevant to firewalling or masquerading. Use port range (for example, 500:510) whenever appropriate. Aspects related to the DMZ (demilitarized zone) as mentioned in the configuration file are not covered here. They are applicable only to a more complex network infrastructure found in larger organizations (corporate networks), which require extensive configuration and in-depth knowledge about the subject.

First, use the YaST module Системные службы (Уровень запуска) to enable SuSEfirewall2 in your runlevel (3 or 5 most likely). It sets the symlinks for the SuSEfirewall2_* scripts in the /etc/init.d/rc?.d/ directories.

FW_DEV_EXT (firewall, masquerading)

The device linked to the Internet. For a modem connection, enter ppp0. For an ISDN link, use ippp0. DSL connections use dsl0. Specify auto to use the interface that corresponds to the default route.

FW_DEV_INT (firewall, masquerading)

The device linked to the internal, private network (such as eth0). Leave this blank if there is no internal network and the firewall protects only the host on which it runs.

FW_ROUTE (firewall, masquerading)

If you need the masquerading function, set this to yes. Your internal hosts will not be visible to the outside, because their private network addresses (e.g., 192.168.x.x) are ignored by Internet routers.

For a firewall without masquerading, set this to yes if you want to allow access to the internal network. Your internal hosts need to use officially registered IP addresses in this case. Normally, however, you should not allow access to your internal network from the outside.

FW_MASQUERADE (masquerading)

Set this to yes if you need the masquerading function. This provides a virtually direct connection to the Internet for the internal hosts. It is more secure to have a proxy server between the hosts of the internal network and the Internet. Masquerading is not needed for services that a proxy server provides.

FW_MASQ_NETS (masquerading)

Specify the hosts or networks to masquerade, leaving a space between the individual entries. For example:

FW_MASQ_NETS="192.168.0.0/24 192.168.10.1"
FW_PROTECT_FROM_INT (firewall)

Set this to yes to protect your firewall host from attacks originating in your internal network. Services are only available to the internal network if explicitly enabled. Also see FW_SERVICES_INT_TCP and FW_SERVICES_INT_UDP.

FW_SERVICES_EXT_TCP (firewall)

Enter the TCP ports that should be made available. Leave this blank for a normal workstation at home that should not offer any services.

FW_SERVICES_EXT_UDP (firewall)

Leave this blank unless you run a UDP service and want to make it available to the outside. The services that use UDP include include DNS servers, IPsec, TFTP, DHCP and others. In that case, enter the UDP ports to use.

FW_SERVICES_ACCEPT_EXT (firewall)

List services to allow from the Internet. This is a more generic form of the FW_SERVICES_EXT_TCP and FW_SERVICES_EXT_UDP settings, and more specific than FW_TRUSTED_NETS. The notation is a space-separated list of net,protocol[,dport][,sport], for example 0/0,tcp,22 or 0/0,tcp,22,,hitcount=3,blockseconds=60,recentname=ssh, which means: allow a maximum of three SSH connects per minute from one IP address.

FW_SERVICES_INT_TCP (firewall)

With this variable, define the services available for the internal network. The notation is the same as for FW_SERVICES_EXT_TCP, but the settings are applied to the internal network. The variable only needs to be set if FW_PROTECT_FROM_INT is set to yes.

FW_SERVICES_INT_UDP (firewall)

See FW_SERVICES_INT_TCP.

FW_SERVICES_ACCEPT_INT (firewall)

List services to allow from internal hosts. See FW_SERVICES_ACCEPT_EXT.

FW_SERVICES_ACCEPT_RELATED_* (firewall)

This is how the SuSEfirewall2 implementation considers packets RELATED by netfilter.

For example, to allow finer grained filtering of Samba broadcast packets, RELATED packets are not accepted unconditionally. Variables starting with FW_SERVICES_ACCEPT_RELATED_ allow restricting RELATED packets handling to certain networks, protocols and ports.

This means that adding connection tracking modules (conntrack modules) to FW_LOAD_MODULES does not automatically result in accepting the packets tagged by those modules. Additionally, you must set variables starting with FW_SERVICES_ACCEPT_RELATED_ to a suitable value.

FW_CUSTOMRULES (firewall)

Uncomment this variable to install custom rules. Find examples in /etc/sysconfig/scripts/SuSEfirewall2-custom.

After configuring the firewall, test your setup. The firewall rule sets are created by entering SuSEfirewall2 start as root. Then use telnet, for example, from an external host to see whether the connection is actually denied. After that, review /var/log/messages, where you should see something like this:

Mar 15 13:21:38 linux kernel: SFW2-INext-DROP-DEFLT IN=eth0 
OUT= MAC=00:80:c8:94:c3:e7:00:a0:c9:4d:27:56:08:00 SRC=192.168.10.0 
DST=192.168.10.1 LEN=60 TOS=0x10 PREC=0x00 TTL=64 ID=15330 DF PROTO=TCP 
SPT=48091 DPT=23 WINDOW=5840 RES=0x00 SYN URGP=0 
OPT (020405B40402080A061AFEBC0000000001030300)

Other packages to test your firewall setup are Nmap (portscanner) or OpenVAS (Open Vulnerability Assessment System). The documentation of Nmap is found at /usr/share/doc/packages/nmap after installing the package and the documentation of openVAS resides at http://www.openvas.org.

13.5. For More Information

The most up-to-date information and other documentation about the SuSEfirewall2 package is found in /usr/share/doc/packages/SuSEfirewall2. The home page of the netfilter and iptables project, http://www.netfilter.org, provides a large collection of documents in many languages.