1 .\" Copyright (C) 1998 Matthew Dillon. All rights reserved.
3 .\" Redistribution and use in source and binary forms, with or without
4 .\" modification, are permitted provided that the following conditions
6 .\" 1. Redistributions of source code must retain the above copyright
7 .\" notice, this list of conditions and the following disclaimer.
8 .\" 2. Redistributions in binary form must reproduce the above copyright
9 .\" notice, this list of conditions and the following disclaimer in the
10 .\" documentation and/or other materials provided with the distribution.
12 .\" THIS SOFTWARE IS PROVIDED BY AUTHOR AND CONTRIBUTORS ``AS IS'' AND
13 .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
14 .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
15 .\" ARE DISCLAIMED. IN NO EVENT SHALL AUTHOR OR CONTRIBUTORS BE LIABLE
16 .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
17 .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
18 .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
19 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
20 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
21 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
31 .Nd introduction to security under FreeBSD
33 Security is a function that begins and ends with the system administrator.
36 multi-user systems have some inherent security, the job of building and
37 maintaining additional security mechanisms to keep users
40 one of the single largest undertakings of the sysadmin.
42 only as secure as you make them, and security concerns are ever competing
43 with the human necessity for convenience.
46 in general, are capable of running a huge number of simultaneous processes
47 and many of these processes operate as servers \(em meaning that external
48 entities can connect and talk to them.
49 As yesterday's mini-computers and mainframes
50 become today's desktops, and as computers become networked and internetworked,
51 security becomes an ever bigger issue.
53 Security is best implemented through a layered onion approach.
55 what you want to do is to create as many layers of security as are convenient
56 and then carefully monitor the system for intrusions.
58 overbuild your security or you will interfere with the detection side, and
59 detection is one of the single most important aspects of any security
61 For example, it makes little sense to set the
66 on every system binary because while this may temporarily protect the
67 binaries, it prevents an attacker who has broken in from making an
68 easily detectable change that may result in your security mechanisms not
69 detecting the attacker at all.
71 System security also pertains to dealing with various forms of attacks,
72 including attacks that attempt to crash or otherwise make a system unusable
73 but do not attempt to break root.
74 Security concerns can be split up into
76 .Bl -enum -offset indent
78 Denial of Service attacks (DoS)
80 User account compromises
82 Root compromise through accessible servers
84 Root compromise via user accounts
89 A denial of service attack is an action that deprives the machine of needed
91 Typically, DoS attacks are brute-force mechanisms that attempt
92 to crash or otherwise make a machine unusable by overwhelming its servers or
94 Some DoS attacks try to take advantages of bugs in the
95 networking stack to crash a machine with a single packet.
97 only be fixed by applying a bug fix to the kernel.
98 Attacks on servers can
99 often be fixed by properly specifying options to limit the load the servers
100 incur on the system under adverse conditions.
101 Brute-force network attacks are harder to deal with.
102 A spoofed-packet attack, for example, is
103 nearly impossible to stop short of cutting your system off from the Internet.
104 It may not be able to take your machine down, but it can fill up Internet
107 A user account compromise is even more common than a DoS attack.
109 sysadmins still run standard
115 servers on their machines.
116 These servers, by default, do not operate over encrypted
118 The result is that if you have any moderate-sized user base,
119 one or more of your users logging into your system from a remote location
120 (which is the most common and convenient way to log in to a system)
121 will have his or her password sniffed.
122 The attentive system administrator will analyze
123 his remote access logs looking for suspicious source addresses
124 even for successful logins.
126 One must always assume that once an attacker has access to a user account,
127 the attacker can break root.
128 However, the reality is that in a well secured
129 and maintained system, access to a user account does not necessarily give the
130 attacker access to root.
131 The distinction is important because without access
132 to root the attacker cannot generally hide his tracks and may, at best, be
133 able to do nothing more than mess with the user's files or crash the machine.
134 User account compromises are very common because users tend not to take the
135 precautions that sysadmins take.
137 System administrators must keep in mind that there are potentially many ways
138 to break root on a machine.
139 The attacker may know the root password,
141 may find a bug in a root-run server and be able to break root over a network
142 connection to that server, or the attacker may know of a bug in an SUID-root
143 program that allows the attacker to break root once he has broken into a
145 If an attacker has found a way to break root on a machine,
146 the attacker may not have a need to install a backdoor.
147 Many of the root holes found and closed to date involve a considerable amount
148 of work by the attacker to clean up after himself, so most attackers do install
150 This gives you a convenient way to detect the attacker.
152 it impossible for an attacker to install a backdoor may actually be detrimental
153 to your security because it will not close off the hole the attacker used to
154 break in in the first place.
156 Security remedies should always be implemented with a multi-layered
158 approach and can be categorized as follows:
159 .Bl -enum -offset indent
161 Securing root and staff accounts
163 Securing root \(em root-run servers and SUID/SGID binaries
165 Securing user accounts
167 Securing the password file
169 Securing the kernel core, raw devices, and file systems
171 Quick detection of inappropriate changes made to the system
175 .Sh SECURING THE ROOT ACCOUNT AND SECURING STAFF ACCOUNTS
176 Do not bother securing staff accounts if you have not secured the root
178 Most systems have a password assigned to the root account.
180 first thing you do is assume that the password is
183 This does not mean that you should remove the password.
185 password is almost always necessary for console access to the machine.
186 What it does mean is that you should not make it possible to use the password
187 outside of the console or possibly even with a
190 For example, make sure that your PTYs are specified as being
195 so that direct root logins via
201 other login services such as
203 make sure that direct root logins are
204 disabled there as well.
205 Consider every access method \(em services such as
207 often fall through the cracks.
208 Direct root logins should only be allowed
209 via the system console.
211 Of course, as a sysadmin you have to be able to get to root, so we open up
213 But we make sure these holes require additional password
214 verification to operate.
215 One way to make root accessible is to add appropriate
216 staff accounts to the
220 The staff members placed in the
225 You should never give staff
228 access by putting them in the
230 group in their password entry.
231 Staff accounts should be placed in a
233 group, and then added to the
238 Only those staff members who actually need to have root access
239 should be placed in the
242 It is also possible, when using an
243 authentication method such as Kerberos, to use Kerberos's
245 file in the root account to allow a
247 to root without having to place anyone at all in the
251 may be the better solution since the
253 mechanism still allows an
254 intruder to break root if the intruder has gotten hold of your password
255 file and can break into a staff account.
259 is better than having nothing at all, it is not necessarily the safest
262 An indirect way to secure the root account is to secure your staff accounts
263 by using an alternative login access method and *'ing out the crypted password
264 for the staff accounts.
265 This way an intruder may be able to steal the password
266 file but will not be able to break into any staff accounts or root, even if
267 root has a crypted password associated with it (assuming, of course, that
268 you have limited root access to the console).
270 get into their staff accounts through a secure login mechanism such as
274 using a private/public
276 When you use something like Kerberos you generally must secure
277 the machines which run the Kerberos servers and your desktop workstation.
278 When you use a public/private key pair with SSH, you must generally secure
279 the machine you are logging in
281 (typically your workstation),
283 also add an additional layer of protection to the key pair by password
284 protecting the keypair when you create it with
287 to *-out the passwords for staff accounts also guarantees that staff members
288 can only log in through secure access methods that you have set up.
290 thus force all staff members to use secure, encrypted connections for
291 all their sessions which closes an important hole used by many intruders: that
292 of sniffing the network from an unrelated, less secure machine.
294 The more indirect security mechanisms also assume that you are logging in
295 from a more restrictive server to a less restrictive server.
297 if your main box is running all sorts of servers, your workstation should not
299 In order for your workstation to be reasonably secure
300 you should run as few servers as possible, up to and including no servers
301 at all, and you should run a password-protected screen blanker.
302 Of course, given physical access to
303 a workstation, an attacker can break any sort of security you put on it.
304 This is definitely a problem that you should consider but you should also
305 consider the fact that the vast majority of break-ins occur remotely, over
306 a network, from people who do not have physical access to your workstation or
309 Using something like Kerberos also gives you the ability to disable or
310 change the password for a staff account in one place and have it immediately
311 affect all the machines the staff member may have an account on.
313 member's account gets compromised, the ability to instantly change his
314 password on all machines should not be underrated.
315 With discrete passwords, changing a password on N machines can be a mess.
317 re-passwording restrictions with Kerberos: not only can a Kerberos ticket
318 be made to timeout after a while, but the Kerberos system can require that
319 the user choose a new password after a certain period of time
321 .Sh SECURING ROOT \(em ROOT-RUN SERVERS AND SUID/SGID BINARIES
322 The prudent sysadmin only runs the servers he needs to, no more, no less.
323 Be aware that third party servers are often the most bug-prone.
325 running an old version of
328 .Xr popper 8 Pq Pa ports/mail/popper
329 is like giving a universal root
330 ticket out to the entire world.
331 Never run a server that you have not checked
333 Many servers do not need to be run as root.
340 daemons can be run in special user
342 A sandbox is not perfect unless you go to a large amount of trouble, but the
343 onion approach to security still stands: if someone is able to break in
344 through a server running in a sandbox, they still have to break out of the
346 The more layers the attacker must break through, the lower the
347 likelihood of his success.
348 Root holes have historically been found in
349 virtually every server ever run as root, including basic system servers.
350 If you are running a machine through which people only log in via
357 then turn off those services!
360 now defaults to running
366 Another program which may be a candidate for running in a sandbox is
370 includes the arguments necessary to run
372 in a sandbox in a commented-out form.
373 Depending on whether you
374 are installing a new system or upgrading an existing system, the special
375 user accounts used by these sandboxes may not be installed.
377 sysadmin would research and implement sandboxes for servers whenever possible.
379 There are a number of other servers that typically do not run in sandboxes:
385 There are alternatives to
386 some of these, but installing them may require more work than you are willing
388 (the convenience factor strikes again).
389 You may have to run these
390 servers as root and rely on other mechanisms to detect break-ins that might
393 The other big potential root hole in a system are the SUID-root and SGID
394 binaries installed on the system.
395 Most of these binaries, such as
398 .Pa /bin , /sbin , /usr/bin ,
401 While nothing is 100% safe,
402 the system-default SUID and SGID binaries can be considered reasonably safe.
403 Still, root holes are occasionally found in these binaries.
405 was found in Xlib in 1998 that made
406 .Xr xterm 1 Pq Pa ports/x11/xterm
407 (which is typically SUID)
409 It is better to be safe than sorry and the prudent sysadmin will restrict SUID
410 binaries that only staff should run to a special group that only staff can
411 access, and get rid of
412 .Pq Dq Li "chmod 000"
413 any SUID binaries that nobody uses.
414 A server with no display generally does not need an
417 SGID binaries can be almost as dangerous.
418 If an intruder can break an SGID-kmem binary the
419 intruder might be able to read
421 and thus read the crypted password
422 file, potentially compromising any passworded account.
424 intruder who breaks group
426 can monitor keystrokes sent through PTYs,
427 including PTYs used by users who log in through secure methods.
431 group can write to almost any user's TTY.
433 is running a terminal
434 program or emulator with a keyboard-simulation feature, the intruder can
436 generate a data stream that causes the user's terminal to echo a command, which
437 is then run as that user.
438 .Sh SECURING USER ACCOUNTS
439 User accounts are usually the most difficult to secure.
441 draconian access restrictions on your staff and *-out their passwords, you
442 may not be able to do so with any general user accounts you might have.
444 you do have sufficient control then you may win out and be able to secure the
445 user accounts properly.
446 If not, you simply have to be more vigilant in your
447 monitoring of those accounts.
448 Use of SSH and Kerberos for user accounts is
449 more problematic due to the extra administration and technical support
450 required, but still a very good solution compared to a crypted password
452 .Sh SECURING THE PASSWORD FILE
453 The only sure fire way is to *-out as many passwords as you can and
454 use SSH or Kerberos for access to those accounts.
456 crypted password file
458 can only be read by root, it may
459 be possible for an intruder to obtain read access to that file even if the
460 attacker cannot obtain root-write access.
462 Your security scripts should always check for and report changes to
465 .Sx CHECKING FILE INTEGRITY
467 .Sh SECURING THE KERNEL CORE, RAW DEVICES, AND FILE SYSTEMS
468 If an attacker breaks root he can do just about anything, but there
469 are certain conveniences.
470 For example, most modern kernels have a packet sniffing device driver built in.
477 An intruder will commonly attempt to run a packet sniffer
478 on a compromised machine.
479 You do not need to give the intruder the
480 capability and most systems should not have the
484 But even if you turn off the
486 device, you still have
492 the intruder can still write to raw disk devices.
493 Also, there is another kernel feature called the module loader,
495 An enterprising intruder can use a KLD module to install
498 device or other sniffing device on a running kernel.
499 To avoid these problems you have to run
500 the kernel at a higher security level, at least level 1.
501 The security level can be set with a
507 set the security level to 1, write access to raw devices will be denied and
516 flag is set on critical startup binaries, directories, and
517 script files \(em everything that gets run
518 up to the point where the security level is set.
519 This might be overdoing it, and upgrading the system is much more
520 difficult when you operate at a higher security level.
521 You may compromise and
522 run the system at a higher security level but not set the
525 system file and directory under the sun.
526 Another possibility is to simply
532 It should be noted that being too draconian in
533 what you attempt to protect may prevent the all-important detection of an
536 The kernel runs with five different security levels.
537 Any super-user process can raise the level, but no process
539 The security levels are:
542 Permanently insecure mode \- always run the system in insecure mode.
543 This is the default initial value.
545 Insecure mode \- immutable and append-only flags may be turned off.
546 All devices may be read or written subject to their permissions.
548 Secure mode \- the system immutable and system append-only flags may not
550 disks for mounted file systems,
555 (if your platform has it) may not be opened for writing;
558 may not be loaded or unloaded.
560 Highly secure mode \- same as secure mode, plus disks may not be
561 opened for writing (except by
563 whether mounted or not.
564 This level precludes tampering with file systems by unmounting them,
565 but also inhibits running
567 while the system is multi-user.
569 In addition, kernel time changes are restricted to less than or equal to one
571 Attempts to change the time by more than this will log the message
572 .Dq Time adjustment clamped to +1 second .
574 Network secure mode \- same as highly secure mode, plus
575 IP packet filter rules (see
580 cannot be changed and
584 configuration cannot be adjusted.
587 The security level can be configured with variables documented in
589 .Sh CHECKING FILE INTEGRITY: BINARIES, CONFIG FILES, ETC
590 When it comes right down to it, you can only protect your core system
591 configuration and control files so much before the convenience factor
597 bit on most of the files in
601 is probably counterproductive because
602 while it may protect the files, it also closes a detection window.
604 last layer of your security onion is perhaps the most important \(em detection.
605 The rest of your security is pretty much useless (or, worse, presents you with
606 a false sense of safety) if you cannot detect potential incursions.
608 the job of the onion is to slow down the attacker rather than stop him
609 in order to give the detection layer a chance to catch him in
612 The best way to detect an incursion is to look for modified, missing, or
615 way to look for modified files is from another (often centralized)
616 limited-access system.
617 Writing your security scripts on the extra-secure limited-access system
618 makes them mostly invisible to potential attackers, and this is important.
619 In order to take maximum advantage you generally have to give the
620 limited-access box significant access to the other machines in the business,
621 usually either by doing a read-only NFS export of the other machines to the
622 limited-access box, or by setting up SSH keypairs to allow the limit-access
623 box to SSH to the other machines.
624 Except for its network traffic, NFS is
625 the least visible method \(em allowing you to monitor the file systems on each
626 client box virtually undetected.
628 limited-access server is connected to the client boxes through a switch,
629 the NFS method is often the better choice.
630 If your limited-access server
631 is connected to the client boxes through a hub or through several layers
632 of routing, the NFS method may be too insecure (network-wise) and using SSH
633 may be the better choice even with the audit-trail tracks that SSH lays.
635 Once you give a limit-access box at least read access to the client systems
636 it is supposed to monitor, you must write scripts to do the actual
638 Given an NFS mount, you can write scripts out of simple system
643 It is best to physically
645 the client-box files boxes at least once a
646 day, and to test control files such as those found in
651 When mismatches are found relative to the base MD5
652 information the limited-access machine knows is valid, it should scream at
653 a sysadmin to go check it out.
654 A good security script will also check for
655 inappropriate SUID binaries and for new or deleted files on system partitions
661 When using SSH rather than NFS, writing the security script is much more
663 You essentially have to
665 the scripts to the client box in order to run them, making them visible, and
666 for safety you also need to
668 the binaries (such as
670 that those scripts use.
673 daemon on the client box may already be compromised.
675 using SSH may be necessary when running over unsecure links, but it is also a
676 lot harder to deal with.
678 A good security script will also check for changes to user and staff members
679 access configuration files:
680 .Pa .rhosts , .shosts , .ssh/authorized_keys
681 and so forth, files that might fall outside the purview of the MD5 check.
683 If you have a huge amount of user disk space it may take too long to run
684 through every file on those partitions.
685 In this case, setting mount
686 flags to disallow SUID binaries on those partitions is a good
693 is what you want to look into.
694 I would scan them anyway at least once a
695 week, since the object of this layer is to detect a break-in whether or
696 not the break-in is effective.
701 is a relatively low-overhead feature of
702 the operating system which I recommend using as a post-break-in evaluation
704 It is especially useful in tracking down how an intruder has
705 actually broken into a system, assuming the file is still intact after
708 Finally, security scripts should process the log files and the logs themselves
709 should be generated in as secure a manner as possible \(em remote syslog can be
711 An intruder tries to cover his tracks, and log files are critical
712 to the sysadmin trying to track down the time and method of the initial
714 One way to keep a permanent record of the log files is to run
715 the system console to a serial port and collect the information on a
716 continuing basis through a secure machine monitoring the consoles.
718 A little paranoia never hurts.
719 As a rule, a sysadmin can add any number
720 of security features as long as they do not affect convenience, and
721 can add security features that do affect convenience with some added
723 Even more importantly, a security administrator should mix it up
724 a bit \(em if you use recommendations such as those given by this manual
725 page verbatim, you give away your methodologies to the prospective
726 attacker who also has access to this manual page.
727 .Sh SPECIAL SECTION ON DoS ATTACKS
728 This section covers Denial of Service attacks.
729 A DoS attack is typically a packet attack.
730 While there is not much you can do about modern spoofed
731 packet attacks that saturate your network, you can generally limit the damage
732 by ensuring that the attacks cannot take down your servers.
733 .Bl -enum -offset indent
735 Limiting server forks
737 Limiting springboard attacks (ICMP response attacks, ping broadcast, etc.)
742 A common DoS attack is against a forking server that attempts to cause the
743 server to eat processes, file descriptors, and memory until the machine
748 has several options to limit this sort of attack.
749 It should be noted that while it is possible to prevent a machine from going
750 down it is not generally possible to prevent a service from being disrupted
754 manual page carefully and pay specific attention
760 Note that spoofed-IP attacks will circumvent
765 so typically a combination of options must be used.
766 Some standalone servers have self-fork-limitation parameters.
771 .Fl OMaxDaemonChildren
772 option which tends to work much
773 better than trying to use
775 load limiting options due to the
778 .Va MaxDaemonChildren
779 parameter when you start
781 high enough to handle your expected load but not so high that the
782 computer cannot handle that number of
784 without falling on its face.
785 It is also prudent to run
790 .Pq Fl ODeliveryMode=queued
791 and to run the daemon
792 .Pq Dq Nm sendmail Fl bd
793 separate from the queue-runs
794 .Pq Dq Nm sendmail Fl q15m .
795 If you still want real-time delivery you can run the queue
796 at a much lower interval, such as
798 but be sure to specify a reasonable
799 .Va MaxDaemonChildren
802 to prevent cascade failures.
806 daemon can be attacked directly and it is strongly recommended that you use
809 option whenever possible, and the
813 You should also be fairly careful
814 with connect-back services such as tcpwrapper's reverse-identd, which can
815 be attacked directly.
816 You generally do not want to use the reverse-ident
817 feature of tcpwrappers for this reason.
819 It is a very good idea to protect internal services from external access
820 by firewalling them off at your border routers.
821 The idea here is to prevent
822 saturation attacks from outside your LAN, not so much to protect internal
823 services from network-based root compromise.
824 Always configure an exclusive
829 ports A, B, C, D, and M-Z
832 way you can firewall off all of your low ports except for certain specific
835 (if you are primary for a zone),
838 and other internet-accessible services.
839 If you try to configure the firewall the other
840 way \(em as an inclusive or permissive firewall, there is a good chance that you
843 a couple of services or that you will add a new internal
844 service and forget to update the firewall.
845 You can still open up the
846 high-numbered port range on the firewall to allow permissive-like operation
847 without compromising your low ports.
851 control the range of port numbers used for dynamic binding via the various
852 .Va net.inet.ip.portrange
854 .Pq Dq Li "sysctl net.inet.ip.portrange" ,
856 ease the complexity of your firewall's configuration.
857 I usually use a normal
858 first/last range of 4000 to 5000, and a hiport range of 49152 to 65535, then
859 block everything under 4000 off in my firewall
860 (except for certain specific
861 internet-accessible ports, of course).
863 Another common DoS attack is called a springboard attack \(em to attack a server
864 in a manner that causes the server to generate responses which then overload
865 the server, the local network, or some other machine.
866 The most common attack
867 of this nature is the ICMP PING BROADCAST attack.
868 The attacker spoofs ping
869 packets sent to your LAN's broadcast address with the source IP address set
870 to the actual machine they wish to attack.
871 If your border routers are not
872 configured to stomp on ping's to broadcast addresses, your LAN winds up
873 generating sufficient responses to the spoofed source address to saturate the
874 victim, especially when the attacker uses the same trick on several dozen
875 broadcast addresses over several dozen different networks at once.
876 Broadcast attacks of over a hundred and twenty megabits have been measured.
877 A second common springboard attack is against the ICMP error reporting system.
879 constructing packets that generate ICMP error responses, an attacker can
880 saturate a server's incoming network and cause the server to saturate its
881 outgoing network with ICMP responses.
882 This type of attack can also crash the
883 server by running it out of
885 especially if the server cannot drain the
886 ICMP responses it generates fast enough.
889 kernel has a new kernel
890 compile option called
892 which limits the effectiveness of these
894 The last major class of springboard attacks is related to
897 services such as the UDP echo service.
899 simply spoofs a UDP packet with the source address being server A's echo port,
900 and the destination address being server B's echo port, where server A and B
901 are both on your LAN.
902 The two servers then bounce this one packet back and
903 forth between each other.
904 The attacker can overload both servers and their
905 LANs simply by injecting a few packets in this manner.
907 exist with the internal chargen port.
908 A competent sysadmin will turn off all
910 .Xr inetd 8 Ns -internal
913 Spoofed packet attacks may also be used to overload the kernel route cache.
915 .Va net.inet.ip.rtexpire , net.inet.ip.rtminexpire ,
917 .Va net.inet.ip.rtmaxcache
920 A spoofed packet attack that uses a random source IP will cause
921 the kernel to generate a temporary cached route in the route table, viewable
923 .Dq Li "netstat -rna | fgrep W3" .
924 These routes typically timeout in 1600
926 If the kernel detects that the cached route table has gotten
927 too big it will dynamically reduce the
929 but will never decrease it to
932 There are two problems: (1) The kernel does not react
933 quickly enough when a lightly loaded server is suddenly attacked, and (2) The
935 is not low enough for the kernel to survive a sustained attack.
936 If your servers are connected to the internet via a T3 or better it may be
937 prudent to manually override both
943 Never set either parameter to zero
944 (unless you want to crash the machine :-)).
945 Setting both parameters to 2 seconds should be sufficient to protect the route
947 .Sh ACCESS ISSUES WITH KERBEROS AND SSH
948 There are a few issues with both Kerberos and SSH that need to be addressed
949 if you intend to use them.
950 Kerberos5 is an excellent authentication
951 protocol but the kerberized
956 There are bugs that make them unsuitable for dealing with binary streams.
958 Kerberos does not encrypt a session unless you use the
961 SSH encrypts everything by default.
963 SSH works quite well in every respect except when it is set up to
964 forward encryption keys.
965 What this means is that if you have a secure workstation holding
966 keys that give you access to the rest of the system, and you
969 unsecure machine, your keys become exposed.
970 The actual keys themselves are
973 installs a forwarding port for the duration of your
974 login and if an attacker has broken root on the unsecure machine he can utilize
975 that port to use your keys to gain access to any other machine that your
978 We recommend that you use SSH in combination with Kerberos whenever possible
980 SSH can be compiled with Kerberos support.
982 your reliance on potentially exposable SSH keys while at the same time
983 protecting passwords via Kerberos.
985 should only be used for automated tasks from secure machines (something
986 that Kerberos is unsuited to).
987 We also recommend that you either turn off
988 key-forwarding in the SSH configuration, or that you make use of the
989 .Va from Ns = Ns Ar IP/DOMAIN
990 option that SSH allows in its
992 file to make the key only usable to entities logging in from specific
1001 .Xr xdm 1 Pq Pa ports/x11/xorg-clients ,
1013 manual page was originally written by