idnits 2.17.1 draft-ietf-opsec-current-practices-06.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** It looks like you're using RFC 3978 boilerplate. You should update this to the boilerplate described in the IETF Trust License Policy document (see https://trustee.ietf.org/license-info), which is required now. -- Found old boilerplate from RFC 3978, Section 5.1 on line 14. -- Found old boilerplate from RFC 3978, Section 5.5 on line 1662. -- Found old boilerplate from RFC 3979, Section 5, paragraph 1 on line 1639. -- Found old boilerplate from RFC 3979, Section 5, paragraph 2 on line 1646. -- Found old boilerplate from RFC 3979, Section 5, paragraph 3 on line 1652. ** This document has an original RFC 3978 Section 5.4 Copyright Line, instead of the newer IETF Trust Copyright according to RFC 4748. ** This document has an original RFC 3978 Section 5.5 Disclaimer, instead of the newer disclaimer which includes the IETF Trust according to RFC 4748. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- == No 'Intended status' indicated for this document; assuming Proposed Standard Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack an IANA Considerations section. (See Section 2.2 of https://www.ietf.org/id-info/checklist for how to handle the case when there are no actions for IANA.) Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the RFC 3978 Section 5.4 Copyright Line does not match the current year -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (July 20, 2006) is 6487 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Missing Reference: 'IRR' is mentioned on line 917, but not defined == Unused Reference: 'RFC2119' is defined on line 1459, but no explicit reference was found in the text == Unused Reference: 'I-D.lewis-infrastructure-security' is defined on line 1487, but no explicit reference was found in the text == Unused Reference: 'I-D.savola-bcp84-urpf-experiences' is defined on line 1492, but no explicit reference was found in the text == Unused Reference: 'I-D.savola-rtgwg-backbone-attacks' is defined on line 1497, but no explicit reference was found in the text ** Obsolete normative reference: RFC 2828 (Obsoleted by RFC 4949) ** Obsolete normative reference: RFC 3682 (Obsoleted by RFC 5082) == Outdated reference: A later version (-03) exists of draft-ietf-v6ops-icmpv6-filtering-recs-02 == Outdated reference: A later version (-03) exists of draft-savola-bcp84-urpf-experiences-01 == Outdated reference: A later version (-03) exists of draft-savola-rtgwg-backbone-attacks-02 Summary: 6 errors (**), 0 flaws (~~), 10 warnings (==), 7 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 OPSEC M. Kaeo 3 Internet-Draft Double Shot Security, Inc. 4 Expires: January 21, 2007 July 20, 2006 6 Operational Security Current Practices 7 draft-ietf-opsec-current-practices-06 9 Status of this Memo 11 By submitting this Internet-Draft, each author represents that any 12 applicable patent or other IPR claims of which he or she is aware 13 have been or will be disclosed, and any of which he or she becomes 14 aware will be disclosed, in accordance with Section 6 of BCP 79. 16 Internet-Drafts are working documents of the Internet Engineering 17 Task Force (IETF), its areas, and its working groups. Note that 18 other groups may also distribute working documents as Internet- 19 Drafts. 21 Internet-Drafts are draft documents valid for a maximum of six months 22 and may be updated, replaced, or obsoleted by other documents at any 23 time. It is inappropriate to use Internet-Drafts as reference 24 material or to cite them other than as "work in progress." 26 The list of current Internet-Drafts can be accessed at 27 http://www.ietf.org/ietf/1id-abstracts.txt. 29 The list of Internet-Draft Shadow Directories can be accessed at 30 http://www.ietf.org/shadow.html. 32 This Internet-Draft will expire on January 21, 2007. 34 Copyright Notice 36 Copyright (C) The Internet Society (2006). 38 Abstract 40 This document is a survey of the current practices used in today's 41 large ISP operational networks to secure layer 2 and layer 3 42 infrastructure devices. The information listed here is the result of 43 information gathered from people directly responsible for defining 44 and implementing secure infrastructures in Internet Service Provider 45 environments. 47 Table of Contents 49 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 50 1.1. Scope . . . . . . . . . . . . . . . . . . . . . . . . . . 3 51 1.2. Threat Model . . . . . . . . . . . . . . . . . . . . . . . 3 52 1.3. Attack Sources . . . . . . . . . . . . . . . . . . . . . . 4 53 1.4. Operational Security Impact from Threats . . . . . . . . . 6 54 1.5. Document Layout . . . . . . . . . . . . . . . . . . . . . 7 55 2. Protected Operational Functions . . . . . . . . . . . . . . . 9 56 2.1. Device Physical Access . . . . . . . . . . . . . . . . . . 9 57 2.2. Device Management - In-Band and Out-of-Band (OOB) . . . . 11 58 2.3. Data Path . . . . . . . . . . . . . . . . . . . . . . . . 17 59 2.4. Routing Control Plane . . . . . . . . . . . . . . . . . . 19 60 2.5. Software Upgrades and Configuration Integrity / 61 Validation . . . . . . . . . . . . . . . . . . . . . . . . 23 62 2.6. Logging Considerations . . . . . . . . . . . . . . . . . . 26 63 2.7. Filtering Considerations . . . . . . . . . . . . . . . . . 30 64 2.8. Denial of Service Tracking / Tracing . . . . . . . . . . . 31 65 3. Security Considerations . . . . . . . . . . . . . . . . . . . 33 66 4. References . . . . . . . . . . . . . . . . . . . . . . . . . . 34 67 4.1. Normative References . . . . . . . . . . . . . . . . . . . 34 68 4.2. Informational References . . . . . . . . . . . . . . . . . 34 69 Appendix A. Acknowledgments . . . . . . . . . . . . . . . . . . . 35 70 Appendix B. Protocol Specific Attacks . . . . . . . . . . . . . . 36 71 B.1. Layer 2 Attacks . . . . . . . . . . . . . . . . . . . . . 36 72 B.2. IPv4 Protocol Based Attacks . . . . . . . . . . . . . . . 36 73 B.3. IPv6 Attacks . . . . . . . . . . . . . . . . . . . . . . . 38 74 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . . 39 75 Intellectual Property and Copyright Statements . . . . . . . . . . 40 77 1. Introduction 79 Security practices are well understood by the network operators who 80 have for many years gone through the growing pains of securing their 81 network infrastructures. However, there does not exist a written 82 document that enumerates these security practices. Network attacks 83 are continually increasing and although it is not necessarily the 84 role of an ISP to act as the Internet police, each ISP has to ensure 85 that certain security practices are followed to ensure that their 86 network is operationally available for their customers. This 87 document is the result of a survey conducted to find out what current 88 security practices are being deployed to secure network 89 infrastructures. 91 1.1. Scope 93 The scope for this survey is restricted to security practices that 94 mitigate exposure to risks with the potential to adversely impact 95 network availability and reliability. Securing the actual data 96 traffic is outside the scope of the conducted survey. This document 97 focuses solely on documenting currently deployed security mechanisms 98 for layer 2 and layer 3 network infrastructure devices. Although 99 primarily focused on IPv4, many of the same practices can (and 100 should) apply to IPv6 networks. Both IPv4 and IPv6 network 101 infrastructures are taken into account in this survey. 103 1.2. Threat Model 105 A threat is a potential for a security violation, which exists when 106 there is a circumstance, capability, action, or event that could 107 breach security and cause harm [RFC2828]. Every operational network 108 is subject to a multitude of threat actions, or attacks, i.e. an 109 assault on system security that derives from an intelligent act that 110 is a deliberate attempt to evade security services and violate the 111 security policy of a system [RFC2828]. All of the threats in any 112 network infrastructure is an instantiation or combination of the 113 following: 115 Reconnaissance: An attack whereby information is gathered to 116 ascertain the network topology or specific device information which 117 can be further used to exploit known vulnerabilities 119 Man-In-The-Middle: An attack where a malicious user impersonates 120 either the sender or recipient of a communication stream while 121 inserting, modifying or dropping certain traffic. This type of 122 attack also covers phishing and session hijacks. 124 Protocol Vulnerability Exploitation: An attack which takes advantage 125 of known protocol vulnerabilities due to design or implementation 126 flaws to cause inappropriate behavior. 128 Message Insertion: This can be a valid message (which could be a 129 reply attack, which is a scenario where a message is captured and 130 resent at later time). A message can also be inserted with any of 131 the fields in the message being OspoofedO, such as IP addresses, port 132 numbers, header fields or even packet content. Flooding is also part 133 of this threat instantiation. 135 Message Diversion/Deletion: An attack where legitimate messages are 136 removed before they can reach the desired recipient or are re- 137 directed to a network segment that is normally not part of the data 138 path. 140 Message Modification: This is a subset of a message insertion attack 141 where a previous message has been captured and modified before being 142 retransmitted. The message can be captured by using a man-in-the- 143 middle attack or message diversion. 145 Note that sometimes Denial of service attacks are listed as separate 146 categories. A denial of service is a consequence of an attack and 147 can be the result of too much traffic (i.e. flooding), or exploiting 148 protocol exploitation or inserting/deleting/diverting/modifying 149 messages. 151 1.3. Attack Sources 153 These attacks can be sourced in a variety of ways: 155 Active vs passive attacks 157 An active attack involves writing data to the network. It is 158 common practice in active attacks to disguise one's address and 159 conceal the identity of the traffic sender. A passive attack 160 involves only reading information off the network. This is 161 possible if the attacker has control of a host in the 162 communications path between two victim machines or has compromised 163 the routing infrastructure to specifically arrange that traffic 164 pass through a compromised machine. In general, the goal of a 165 passive attack is to obtain information that the sender and 166 receiver would prefer to remain private. [RFC3552] 168 On-path vs off-path attacks 170 In order for a datagram to be transmitted from one host to 171 another, it generally must traverse some set of intermediate links 172 and routers. Such routers are naturally able to read, modify, or 173 remove any datagram transmitted along that path. This makes it 174 much easier to mount a wide variety of attacks if you are on-path. 175 Off-path hosts can transmit arbitrary datagrams that appear to 176 come from any hosts but cannot necessarily receive datagrams 177 intended for other hosts. Thus, if an attack depends on being 178 able to receive data, off-path hosts must first subvert the 179 topology in order to place themselves on-path. This is by no 180 means impossible but is not necessarily trivial. [RFC3552] 182 Insider or outsider attacks 184 An "insider attack" is one which is initiated from inside a given 185 security perimeter, by an entity that is authorized to access 186 system resources but uses them in a way not approved by those who 187 granted the authorization. An "outside attack" is initiated from 188 outside the perimeter, by an unauthorized or illegitimate user of 189 the system. 191 Deliberate attacks vs unintentional events 193 A deliberate attack is one where a miscreant intentionally 194 performs an assault on system security. However, there are also 195 instances where unintentional events cause the same harm yet are 196 performed without malice in mind. Configuration errors and 197 software bugs can be as devastating to network availability as any 198 deliberate attack on the network infrastructure. 200 The attack source can be a combination of any of the above, all of 201 which need to be considered when trying to ascertain what impact any 202 attack can have on the availability and reliability of the network. 203 It is nearly impossible to stop insider attacks or unintentional 204 events. However, if appropriate monitoring mechanisms are in place, 205 these attacks can also be detected and mitigated as with any other 206 attack source. The amount of effort it takes to identify and trace 207 an attack is of course dependent on the resourcefulness of the 208 attacker. Any of the specific attacks discussed further in this 209 document will elaborate on malicious behavior which are sourced by an 210 "outsider" and are deliberate attacks. Some further elaboration will 211 be given to the feasibility of passive vs active and on-path vs off- 212 path attacks to show the motivation behind deploying certain security 213 features. 215 1.4. Operational Security Impact from Threats 217 The main concern for any of the potential attack scenarios is the 218 impact and harm it can cause to the network infrastructure. The 219 threat consequences are the security violations which results from a 220 threat action, i.e. an attack. These are typically classified as 221 follows: 223 (Unauthorized) Disclosure 225 A circumstance or event whereby an entity gains access to data for 226 which the entity is not authorized. 228 Deception 230 A circumstance or event that may result in an authorized entity 231 receiving false data and believing it to be true. 233 Disruption 235 A circumstance or event that interrupts or prevents the correct 236 operation of system services and functions. A broad variety of 237 attacks, collectively called denial of service attacks, threaten 238 the availability of systems and bandwidth to legitimate users. 239 Many such attacks are designed to consume machine resources, 240 making it difficult or impossible to serve legitimate users. 241 Other attacks cause the target machine to crash, completely 242 denying service to users. 244 Usurpation 246 A circumstance or event that results in control of system services 247 or functions by an unauthorized entity. Most network 248 infrastructure systems are only intended to be completely 249 accessible to certain authorized individuals. Should an 250 unauthorized person gain access to critical layer 2 / layer 3 251 infrastructure devices or services, they could cause great harm to 252 the reliability and availability of the network. 254 A complete description of threat actions that can cause these threat 255 consequences can be found in [RFC2828]. Typically, a number of 256 different network attacks are used in combination to cause one or 257 more of the above mentioned threat consequences. An example would be 258 a malicious user who has the capability to eavesdrop on traffic. 260 First, he may listen in on traffic for a while to do some 261 reconnaissance work and ascertain which IP addresses belonged to 262 specific devices such as routers. Were this miscreant to obtain 263 information such as a router password sent in cleartext, he can then 264 proceed to compromise the actual router. From there, the miscreant 265 can launch various active attacks such as sending bogus routing 266 updates to redirect traffic or capture additional traffic to 267 compromise other network devices. 269 1.5. Document Layout 271 This document is a survey of current operational practices that 272 mitigate the risk of being susceptible to any threat actions. As 273 such, the main focus is on the currently deployed security practices 274 used to detect and/or mitigate attacks. The top-level categories in 275 this document are based on operational functions for ISPs and 276 generally relate to what is to be protected. This is followed by a 277 description of which attacks are possible and the security practices 278 currently deployed which will provide the necessary security services 279 to help mitigate these attacks. These security services are 280 classified as: 282 o User Authentication 284 o User Authorization 286 o Data Origin Authentication 288 o Access Control 290 o Data Integrity 292 o Data Confidentiality 294 o Auditing / Logging 296 o DoS Mitigation 298 In many instances, a specific protocol currently deployed will offer 299 a combination of these services. For example, AAA can offer user 300 authentication, user authorization and audit / logging services while 301 SSH can provide data origin authentication, data integrity and data 302 confidentiality. The services offered are more important than the 303 actual protocol used. Note that access control will refer basically 304 to logical access control, i.e. filtering. Each section ends with an 305 additional considerations section which explains why specific 306 protocols may or may not be used and also gives some information 307 regarding capabilities which are not possible today due to bugs or 308 lack of ease of use. 310 2. Protected Operational Functions 312 2.1. Device Physical Access 314 Device physical access pertains to protecting the physical location 315 and access of the layer 2 or layer 3 network infrastructure device. 316 Physical security is a large field of study/practice in and of 317 itself, arguably the largest, oldest and most well understood area of 318 security. Although it is important to have contingency plans for 319 natural disasters such as earthquakes and floods which can cause 320 damage to networking devices, this is out-of-scope for this document. 321 Here we concern ourselves with protecting access to the physical 322 location and how a device can be further protected from unauthorized 323 access if the physical location has been compromised, i.e protecting 324 the console access. This is aimed largely at stopping an intruder 325 with physical access from gaining operational control of the 326 device(s). Note that nothing will stop an attacker with physical 327 access from effecting a denial of service attack, which can be easily 328 accomplished by powering off the device or just unplugging some 329 cables. 331 2.1.1. Threats / Attacks 333 If any intruder gets physical access to a layer 2 or layer 3 device, 334 the entire network infrastructure can be under the control of the 335 intruder. At a minimum, the intruder can take the compromised device 336 out-of-service, causing network disruption, the extent of which 337 depends on the network topology. A worse scenario is where the 338 intruder can use this device to crack the console password and have 339 complete control of the device, perhaps without anyone detecting such 340 a compromise, or to attach another network device onto a port and 341 siphon off data with which the intruder can ascertain the network 342 topology and take control of the entire network. 344 The threat of gaining physical access can be realized in a variety of 345 ways even if critical devices are under high-security. There still 346 occur cases where attackers have impersonated maintenance workers to 347 gain physical access to critical devices that have caused major 348 outages and privacy compromises. Insider attacks from authorized 349 personnel also pose a real threat and must be adequately recognized 350 and dealt with. 352 2.1.2. Security Practices 354 For physical device security, equipment is kept in highly restrictive 355 environments. Only authorized users with cardkey badges have access 356 to any of the physical locations that contain critical network 357 infrastructure devices. These cardkey systems keep track of who 358 accessed which location and at what time. Most cardkey systems have 359 a fail back "master key" in case the card system is down. This 360 "master key" usually has limited access and its use is also carefully 361 logged (which should only happen if the cardkey system is NOT online/ 362 functional). 364 All console access is always password protected and the login time is 365 set to time out after a specified amount of inactivity - typically 366 between 3-10 minutes. The type of privileges that you obtain from a 367 console login varies between separate vendor devices. In some cases 368 you get initial basic access and need to perform a second 369 authentication step to get more privileged (i.e. enable or root) 370 access. In other vendors you get the more privileged access when you 371 log into the console as root, without requiring a second 372 authentication step. 374 How ISPs manage these logins vary greatly although many of the larger 375 ISPs employ some sort of AAA mechanism to help automate privilege 376 level authorization and can utilize the automation to bypass the need 377 for a second authentication step. Also, many ISPs define separate 378 classes of users to have different privileges while logged onto the 379 console. Typically all console access is provided via an out-of-band 380 (OOB) management infrastructure which is discussed in the section on 381 OOB management. 383 2.1.3. Security Services 385 The following security services are offered through the use of the 386 practices described in the previous section: 388 o User Authentication - All individuals who have access to the 389 physical facility are authenticated. Console access is 390 authenticated. 392 o User Authorization - An authenticated individual has implicit 393 authorization to perform commands on the device. In some cases 394 multiple authentication is required to differentiate between basic 395 and more privileged access. 397 o Data Origin Authentication - Not applicable 399 o Access Control - Not applicable 401 o Data Integrity - Not applicable 403 o Data Confidentiality - Not applicable 404 o Auditing / Logging - All access to the physical locations of the 405 infrastructure equipment is logged via electronic card-key 406 systems. All console access is logged (refer to the OOB 407 management section for more details) 409 o DoS Mitigation - Not applicable 411 2.1.4. Additional Considerations 413 Physical security is relevant to operational security practices as 414 described in this document mostly from a console access perspective. 415 Most ISPs provide console access via an OOB management infrastructure 416 which is discussed in the OOB management section of this document. 418 The physical and logical authentication and logging systems should be 419 run independently of each other and reside in different physical 420 locations. These systems need to be secured to ensure that they 421 themselves will not be compromised which could give the intruder 422 valuable authentication and logging information. 424 Social engineering plays a big role in many physical access 425 compromises. Most ISPs have set up training classes and awareness 426 programs to educate company personnel to deny physical access to 427 people who are not properly authenticated or authorized to have 428 physical access to critical infrastructure devices. 430 2.2. Device Management - In-Band and Out-of-Band (OOB) 432 In-band management is generally considered to be device access where 433 the control traffic takes the same data path as the data which 434 traverses the network. Out-of-band management is generally 435 considered to be device access where the control traffic takes a 436 separate path as the data which traverses the network. In many 437 environments, device management for layer 2 and layer 3 438 infrastructure devices is deployed as part of an out-of-band 439 management infrastructure although there are some instances where it 440 is deployed in-band as well. Note that while many of the security 441 concerns and practices are the same for OOB management and in-band 442 management, most ISPs prefer an OOB management system since access to 443 the devices which make up this management network are more vigilantly 444 protected and considered to be less susceptible to malicious 445 activity. 447 Console access is always architected via an OOB network. Presently, 448 the mechanisms used for either in-band management or OOB are via 449 virtual terminal access (i.e. Telnet or SSH), SNMP, or HTTP. In all 450 large ISPs that were interviewed, HTTP management is never used and 451 is explicitly disabled. Note that file transfer protocols (TFTP, 452 FTP, SCP) will be covered in the 'Software Upgrades and Configuration 453 Integrity/Validation' section. 455 2.2.1. Threats / Attacks 457 For device management, passive attacks are possible if someone has 458 the capability to intercept data between the management device and 459 the managed device. The threat is possible if a single 460 infrastructure device is somehow compromised and can act as a network 461 sniffer or if it is possible to insert a new device which acts as a 462 network sniffer. 464 Active attacks are possible for both on-path and off-path scenarios. 465 For on-path active attacks, the situation is the same as for a 466 passive attack, where either a device has to already be compromised 467 or a device can be inserted into the path. For off-path active 468 attacks, where a topology subversion is required to reroute traffic 469 and essentially bring the attacker on-path, the attack is generally 470 limited to message insertion or modification. 472 2.2.1.1. Confidentiality Violations 474 Confidentiality violations can occur when a miscreant intercepts any 475 management data that has been sent in cleartext or with weak 476 encryption. This includes interception of usernames and passwords 477 with which an intruder can obtain unauthorized access to network 478 devices. It can also include other information such as logging or 479 configuration information if an administrator is remotely viewing 480 local logfiles or configuration information. 482 2.2.1.2. Offline Cryptographic Attacks 484 If username/password information was encrypted but the cryptographic 485 mechanism used made it easy to capture data and break the encryption 486 key, the device management traffic could be compromised. The traffic 487 would need to be captured either by eavesdropping on the network or 488 by being able to divert traffic to a malicious user. 490 2.2.1.3. Replay Attacks 492 For a replay attack to be successful, the management traffic would 493 need to first be captured either on-path or diverted to an attacker 494 to later be replayed to the intended recipient. 496 2.2.1.4. Message Insertion/Deletion/Modification 498 Data can be manipulated by someone in control of intermediary hosts. 499 Forging data is also possible with IP spoofing, where a remote host 500 sends out packets which appear to come from another, trusted host. 502 2.2.1.5. Man-In-The-Middle 504 A man-in-the-middle attack attacks the identity of a communicating 505 peer rather than the data stream itself. The attacker intercepts 506 traffic that is sent from a management system to the networking 507 infrastructure device and traffic that is sent from the network 508 infrastructure device to the management system. 510 2.2.2. Security Practices 512 OOB management is done via a terminal server at each location. SSH 513 access is used to get to the terminal server from where sessions to 514 the devices are initiated. Dial-in access is deployed as a backup if 515 the network is not available however, it is common to use dial-back, 516 encrypting modems and/or one-time-password (OTP) modems to avoid the 517 security weaknesses of plain dial-in access. 519 All in-band management and OOB management access to layer 2 and layer 520 3 devices is authenticated. The user authentication and 521 authorization is typically controlled by a AAA server (i.e. RADIUS 522 and/or TACACS+). Credentials used to determine the identity of the 523 user vary from static username/password to one-time username/password 524 scheme such as Secure-ID. Static username/passwords are expired 525 after a specified period of time, usually 30 days. Every 526 authenticated entity via AAA is an individual user for greater 527 granularity of control. Note that often the AAA server used for OOB 528 management authentication is a separate physical device from the AAA 529 server used for in-band management user authentication. In some 530 deployments, the AAA servers used for device management 531 authentication/authorization/accounting are on separate networks to 532 provide a demarcation for any other authentication functions. 534 For backup purposes, there is often a single local database entry for 535 authentication which is known to a very limited set of key personnel. 536 It is usually the highest privilege level username/password 537 combination, which in most cases is the same across all devices. 538 This local device password is routinely regenerated once every 2-3 539 months and is also regenerated immediately after an employee who had 540 access to that password leaves the company or is no longer authorized 541 to have knowledge of that password. 543 Each individual user in the AAA database is configured with specific 544 authorization capability. Specific commands are either individually 545 denied or permitted depending on the capability of the device to be 546 accessed. Multiple privilege levels are deployed. Most individuals 547 are authorized with basic authorization to perform a minimal set of 548 commands while a subset of individuals are authorized to perform more 549 privileged commands. Securing the AAA server is imperative and 550 access to the AAA server itself is strictly controlled. When an 551 individual leaves the company, his/her AAA account is immediately 552 deleted and the TACACS/RADIUS shared secret is reset for all devices. 554 Some management functions are performed using command line interface 555 (CLI) scripting. In these scenarios, a dedicated user is used for 556 the identity in scripts that perform CLI scripting. Once 557 authenticated, these scripts control which commands are legitimate 558 depending on authorization rights of the authenticated individual. 560 SSH is always used for virtual terminal access to provide for an 561 encrypted communication channel. There are exceptions due to 562 equipment limitations which are described in the additional 563 considerations section. 565 If SNMP is used for management, it is for read queries only and 566 restricted to specific hosts. If possible, the view is also 567 restricted to only send the information that the management station 568 needs rather than expose the entire configuration file with the read- 569 only SNMP community. The community strings are carefully chosen to 570 be difficult to crack and there are procedures in place to change 571 these community strings between 30-90 days. If systems support two 572 SNMP community strings, the old string is replaced by first 573 configuring a second newer community string and then migrating over 574 from the currently used string to the newer one. Most large ISPs 575 have multiple SNMP systems accessing their routers so it takes more 576 then one maintenance period to get all the strings fixed in all the 577 right systems. SNMP RW is not used and is disabled by configuration. 579 Access control is strictly enforced for infrastructure devices by 580 using stringent filtering rules. A limited set of IP addresses are 581 allowed to initiate connections to the infrastructure devices and are 582 specific to the services which they are to limited to (i.e. SSH and 583 SNMP). 585 All device management access is audited and any violations trigger 586 alarms which initiate automated email, pager and/or telephone 587 notifications. AAA servers keeps track of the authenticated entity 588 as well as all the commands that were carried out on a specific 589 device. Additionally, the device itself logs any access control 590 violations (i.e. if an SSH request comes in from an IP address which 591 is not explicitly permitted, that event is logged so that the 592 offending IP address can be tracked down and investigations made as 593 to why it was trying to access a particular infrastructure device) 595 2.2.3. Security Services 597 The security services offered for device OOB management are nearly 598 identical to those of device in-band management. Due to the critical 599 nature of controlling and limiting device access, many ISPs feel that 600 physically separating the management traffic from the normal customer 601 data traffic will provide an added level of risk mitigation and limit 602 the potential attack vectors. The following security services are 603 offered through the use of the practices described in the previous 604 section: 606 o User Authentication - All individuals are authenticated via AAA 607 services. 609 o User Authorization - All individuals are authorized via AAA 610 services to perform specific operations once successfully 611 authenticated. 613 o Data Origin Authentication - Management traffic is strictly 614 filtered to allow only specific IP addresses to have access to the 615 infrastructure devices. This does not alleviate risk from spoofed 616 traffic, although when combined with edge filtering using BCP38 617 [RFC2827] and BCP84 [RFC3704] guidelines (discussed in the section 618 2.5), then the risk of spoofing is mitigated barring a compromised 619 internal system. Also, using SSH for device access ensures that 620 noone can spoof the traffic during the SSH session. 622 o Access Control - Management traffic is filtered to allow only 623 specific IP addresses to have access to the infrastructure 624 devices. 626 o Data Integrity - Using SSH provides data integrity and ensures 627 that no one has altered the management data in transit. 629 o Data Confidentiality - Using SSH provides data confidentiality. 631 o Auditing / Logging - Using AAA provides an audit trail for who 632 accessed which device and which operations were performed. 634 o DoS Mitigation - Using packet filters to allow only specific IP 635 addresses to have access to the infrastructure devices. This 636 limits but does not prevent spoofed DoS attacks directed at an 637 infrastructure device. However, the risk is lowered by using a 638 separate physical network for management purposes. 640 2.2.4. Additional Considerations 642 Password selection for any device management protocol used is 643 critical to ensure that the passwords are hard to guess or break 644 using a brute-force attack. 646 IPsec is considered too difficult to deploy and the common protocol 647 to provide for confidential management access is SSH. There are 648 exceptions for using SSH due to equipment limitations since SSH may 649 not be supported on legacy equipment. In some cases changing the 650 hostname of a device requires an SSH rekey event since the key is 651 based on some combination of host name, MAC address and time. Also, 652 in the case where the SSH key is stored on a route processor card, a 653 re-keying of SSH would be required whenever the route processor card 654 needs to be swapped. Some providers feel that this operational 655 impact exceeds the security necessary and instead use Telnet from 656 trusted inside hosts (called 'jumphosts' or 'bastion hosts') to 657 manage those devices. An individual would first SSH to the jumphost 658 and then Telnet from the jumphost to the actual infrastructure 659 device, fully understanding that any passwords will be sent in the 660 clear between the jumphost and the device it is connecting to. All 661 authentication and authorization is still carried out using AAA 662 servers. 664 In instances where Telnet access is used, the logs on the AAA servers 665 are more verbose and more attention is paid to them to detect any 666 abnormal behavior. The jumphosts themselves are carefully controlled 667 machines and usually have limited access. Note that Telnet is NEVER 668 allowed to an infrastructure device except from specific jumphosts; 669 i.e. packet filters are used at the console server and/or 670 infrastructure device to ensure that Telnet is only allowed from 671 specific IP addresses. 673 With thousands of devices to manage, some ISPs have created automated 674 mechanisms to authenticate to devices. As an example, Kerberos has 675 been used to automate the authentication process for devices that 676 have support for Kerberos. An individual would first log in to a 677 Kerberized UNIX server using SSH and generate a Kerberos 'ticket'. 678 This 'ticket' is generally set to have a lifespan of 10 hours and is 679 used to automatically authenticate the individual to the 680 infrastructure devices. 682 In instances where SNMP is used, some legacy devices only support 683 SNMPv1 which then requires the provider to mandate its use across all 684 infrastructure devices for operational simplicity. SNMPv2 is 685 primarily deployed since it is easier to set up than v3. 687 2.3. Data Path 689 This section refers to how traffic is handled which traverses the 690 network infrastructure device. The primary goal of ISPs is to 691 forward customer traffic. However, due to the large amount of 692 malicious traffic that can cause DoS attacks and render the network 693 unavailable, specific measures are sometimes deployed to ensure the 694 availability to forward legitimate customer traffic. 696 2.3.1. Threats / Attacks 698 Any data traffic can potentially be attack traffic and the challenge 699 is to detect and potentially stop forwarding any of the malicious 700 traffic. The deliberately sourced attack traffic can consist of 701 packets with spoofed source and/or destination addresses or any other 702 malformed packet which mangle any portion of a header field to cause 703 protocol-related security issues (such as resetting connections, 704 causing unwelcome ICMP redirects, creating unwelcome IP options or 705 packet fragmentations). 707 2.3.2. Security Practices 709 Filtering and rate limiting are the primary mechanism to provide risk 710 mitigation of malicious traffic rendering the ISP services 711 unavailable. However, filtering and rate limiting of data path 712 traffic is deployed in a variety of ways depending on how automated 713 the process is and what the capabilities and performance limitations 714 of existing deployed hardware are. 716 The ISPs which do not have performance issues with their equipment 717 follow BCP38 [RFC2827] and BCP84 [RFC3704] guidelines for ingress 718 filtering. BCP38 recommends filtering ingress packets with obviously 719 spoofed and/or 'reserved' source addresses to limit the effects of 720 denial of service attacks while BCP84 extends the recommendation for 721 multi-homed environments. Filters are also used to help alleviate 722 issues between service providers. Without any filtering, an inter- 723 exchange peer could steal transit just by using static routes and 724 essentially redirect data traffic. Therefore, some ISPs have 725 implemented ingress/egress filters which block unexpected source and 726 destination addresses not defined in the above-mentioned documents. 727 Null routes and black-hole triggered routing are used to deter any 728 detected malicious traffic streams. These two techniques are 729 described in more detail in section 2.8 below. 731 Most ISPs consider layer 4 filtering useful but it is only 732 implemented if performance limitations allow for it. Layer 4 733 filtering is typically only when no other option exists since it does 734 pose a large administrative overhead and ISPs are very much opposed 735 to acting as the Internet firewall. Netflow is used for tracking 736 traffic flows but there is some concern whether sampling is good 737 enough to detect malicious behavior. 739 Unicast RPF is not consistently implemented. Some ISPs are in 740 process of doing so while other ISPs think that the perceived benefit 741 of knowing that spoofed traffic comes from legitimate addresses are 742 not worth the operational complexity. Some providers have a policy 743 of implementing uRPF at link speeds of DS3 and below. 745 2.3.3. Security Services 747 o User Authentication - Not applicable 749 o User Authorization - Not applicable 751 o Data Origin Authentication - When IP address filtering per BCP38, 752 BCP84 and uRPF are deployed at network edges it can ensure that 753 any spoofed traffic comes from at least a legitimate IP address 754 and can be tracked. 756 o Access Control - IP address filtering and layer 4 filtering is 757 used to deny forbidden protocols and limit traffic destined for 758 infrastructure device itself. Filters are also used to block 759 unexpected source/destination addresses. 761 o Data Integrity - Not applicable 763 o Data Confidentiality - Not applicable 765 o Auditing / Logging - Filtering exceptions are logged for potential 766 attack traffic. 768 o DoS Mitigation - Black-hole triggered filtering and rate-limiting 769 are used to limit the risk of DoS attacks. 771 2.3.4. Additional Considerations 773 For layer 2 devices, MAC address filtering and authentication is not 774 used in large-scale deployments. This is due to the problems it can 775 cause when troubleshooting networking issues. Port security becomes 776 unmanageable at a large scale where 1000s of switches are deployed. 778 Rate limiting is used by some ISPs although other ISPs believe it is 779 not really useful since attackers are not well behaved and it doesn't 780 provide any operational benefit over the complexity. Some ISPs feel 781 that rate limiting can also make an attacker's job easier by 782 requiring the attacker to send less traffic to starve legitimate 783 traffic that is part of a rate limiting scheme. Rate limiting may be 784 improved by developing flow-based rate-limiting capabilities with 785 filtering hooks. This would improve the performance as well as the 786 granularity over current capabilities. 788 Lack of consistency regarding the ability to filter, especially with 789 respect to performance issues cause some ISPs to not implement BCP38 790 and BCP84 guidelines for ingress filtering. One such example is at 791 edge boxes where you have up to 1000 T1's connecting into a router 792 with an OC-12 uplink. Some deployed devices experience a large 793 performance impact with filtering which is unacceptable for passing 794 customer traffic through, though ingress filtering (uRPF) might be 795 applicable at the devices connecting these aggregation routers. 796 Where performance is not an issue, the ISPs make a tradeoff between 797 management versus risk. 799 2.4. Routing Control Plane 801 The routing control plane deals with all the traffic which is part of 802 establishing and maintaining routing protocol information. 804 2.4.1. Threats / Attacks 806 Attacks on the routing control plane can be both from passive or 807 active sources. Passive attacks are possible if someone has the 808 capability to intercept data between the communicating routing peers. 809 This can be accomplished if a single routing peer is somehow 810 compromised and can act as a network sniffer or if it is possible to 811 insert a new device which acts as a network sniffer. 813 Active attacks are possible for both on-path and off-path scenarios. 814 For on-path active attacks, the situation is the same as for a 815 passive attack, where either a device has to already be compromised 816 or a device can be inserted into the path. This may lead to an 817 attacker impersonating a legitimate routing peer and exchanging 818 routing information. Unintentional active attacks are more common 819 due to configuration errors, which cause legitimate routing peers to 820 feed invalid routing information to other neighboring peers. 822 For off-path active attacks, the attacks are generally limited to 823 message insertion or modification which can divert traffic to 824 illegitimate destinations and cause traffic to never reach its 825 intended destination. 827 2.4.1.1. Confidentiality Violations 829 Confidentiality violations can occur when a miscreant intercepts any 830 of the routing update traffic. This is becoming more of a concern 831 because many ISPs are classifying addressing schemes and network 832 topologies as private and proprietary information. It is also a 833 concern because the routing protocol packets contain information that 834 may show ways in which routing sessions could be spoofed or hijacked. 835 This in turn could lead into a man-in-the-middle attack where the 836 miscreants can insert themselves into the traffic path or divert the 837 traffic path and violate the confidentiality of user data. 839 2.4.1.2. Offline Cryptographic Attacks 841 If any cryptographic mechanism was used to provide for data integrity 842 and confidentiality, an offline cryptographic attack could 843 potentially compromise the data. The traffic would need to be 844 captured either by eavesdropping on the network or by being able to 845 divert traffic to a malicious user. Note that by using 846 cryptographically protected routing information, the latter would 847 require the cryptographic key to already be compromised anyway so 848 this attack is only feasible if a device was able eavesdrop and 849 capture the cryptographically protected routing information. 851 2.4.1.3. Replay Attacks 853 For a replay attack to be successful, the routing control plane 854 traffic would need to first be captured either on-path or diverted to 855 an attacker to later be replayed to the intended recipient. 856 Additionally, since many of these protocols include replay protection 857 mechanisms, these would also need to be subverted if applicable. 859 2.4.1.4. Message Insertion/Deletion/Modification 861 Routing control plane traffic can be manipulated by someone in 862 control of intermediate hosts. In addition, traffic can be injected 863 by forging IP addresses, where a remote router sends out packets 864 which appear to come from another, trusted router. If enough traffic 865 is injected to be processed by limited memory routers it can cause a 866 DoS attack. 868 2.4.1.5. Man-In-The-Middle 870 A man-in-the-middle attack attacks the identity of a communicating 871 peer rather than the data stream itself. The attacker intercepts 872 traffic that is sent from one routing peer to the other and 873 communicates on behalf of one of the peers. This can lead to 874 diversion of the user traffic to either an unauthorized receiving 875 party or cause legitimate traffic to never reach its intended 876 destination. 878 2.4.2. Security Practices 880 Securing the routing control plane takes many features which are 881 generally deployed as a system. MD5 authentication is used by some 882 ISPs to validate the sending peer and to ensure that the data in 883 transit has not been altered. Some ISPs only deploy MD5 884 authentication at customer's request. Additional sanity checks to 885 ensure with reasonable certainty that the received routing update was 886 originated by a valid routing peer include route filters and the 887 Generalized TTL Security Mechanism (GTSM) feature [RFC3682] 888 (sometimes also referred to as the TTL-Hack). The GTSM feature is 889 used for protocols such as BGP and makes use of a packet's Time To 890 Live (TTL) field (IPv4) or Hop Limit (IPv6) to protect communicating 891 peers. 893 Packet filters are used to limit which systems can appear as a valid 894 peer while route filters are used to limit which routes are believed 895 from a valid peer. In the case of BGP routing, a variety of policies 896 are deployed to limit the propagation of invalid routing information. 897 These include: incoming and outgoing prefix filters for BGP 898 customers, incoming and outgoing prefix filters for peers and 899 upstream neighbors, incoming AS-PATH filter for BGP customers, 900 outgoing AS-PATH filter towards peers and upstream neighbors, route 901 dampening and rejecting selected attributes and communities. 902 Consistency between these policies varies greatly and there is a 903 definite distinction whether the other end is an end-site vs an 904 internal peer vs another big ISP or customer. Mostly ISPs do prefix- 905 filter their end-site customers but due to the operational 906 constraints of maintaining large prefix filter lists, many ISPs are 907 starting to depend on BGP AS-PATH filters to/from their peers and 908 upstream neighbors. 910 In cases where prefix lists are not used, operators often define a 911 maximum prefix limit per peer to prevent misconfiguration (e.g., 912 unintentional de-aggregation) or overload attacks. When the limit is 913 exceeded, the session is either reset or further updates are denied. 914 Typically a lower warning threshold is also configured. 916 Some large ISPs require that routes be registered in an Internet 917 Routing Registry [IRR] which can then be part of the RADB - a public 918 registry of routing information for networks in the Internet that can 919 be used to generate filter lists. Some ISPs, especially in europe, 920 require registered routes before agreeing to become an eBGP peer with 921 someone. 923 Many ISPs also do not propagate interface IP addresses to further 924 reduce attack vectors on routers and connected customers. 926 2.4.3. Security Services 928 o User Authentication - Not applicable 930 o User Authorization - Not applicable 932 o Data Origin Authentication - By using MD5 authentication and/or 933 the TTL-hack a routing peer can be reasonably certain that traffic 934 originated from a valid peer. 936 o Access Control - Route filters, AS-PATH filters and prefix limits 937 are used to control access to specific parts of the network. 939 o Data Integrity - By using MD5 authentication a peer can be 940 reasonably certain that the data has not been modified in transit 941 but there is no mechanism to prove the validity of the routing 942 information itself. 944 o Data Confidentiality - Not implemented 946 o Auditing / Logging - Filter exceptions are logged. 948 o DoS Mitigation - Many DoS attacks are mitigated using a 949 combination of techniques including: MD5 authentication, the GTSM 950 feature, filtering routing advertisements to bogons and filtering 951 routing advertisements to one's own network. 953 2.4.4. Additional Considerations 955 So far the primary concern to secure the routing control plane has 956 been to validate the sending peer and to ensure that the data in 957 transit has not been altered. Although MD5 routing protocol 958 extensions have been implemented which can provide both services, 959 they are not consistently deployed amongst ISPs. Two major 960 deployment concerns have been implementation issues where both 961 software bugs and the lack of graceful re-keying options have caused 962 significant network down times. Also, some ISPs express concern that 963 deploying MD5 authentication will itself be a worse DoS attack victim 964 and prefer to use a combination of other risk mitigation mechanisms 965 such as GTSM (for BGP) and route filters. An issue with GTSM is that 966 it is not supported on all devices across different vendors 967 products'. 969 IPsec is not deployed since the operational management aspects of 970 ensuring interoperability and reliable configurations is too complex 971 and time consuming to be operationally viable. There is also limited 972 concern to the confidentiality of the routing information. The 973 integrity and validity of the updates are of much greater concern. 975 There is concern for manual or automated actions which introduce new 976 routes and can affect the entire routing domain. 978 2.5. Software Upgrades and Configuration Integrity / Validation 980 Software upgrades and configuration changes are usually performed as 981 part of either in-band or OOB management functions. However, there 982 are additional considerations to be taken into account which are 983 enumerated in this section. 985 2.5.1. Threats / Attacks 987 Attacks performed on system software and configurations can be both 988 from passive or active sources. Passive attacks are possible if 989 someone has the capability to intercept data between the network 990 infrastructure device and the system which is downloading or 991 uploading the software or configuration information. This can be 992 accomplished if a single infrastructure device is somehow compromised 993 and can act as a network sniffer or if it is possible to insert a new 994 device which acts as a network sniffer. 996 Active attacks are possible for both on-path and off-path scenarios. 997 For on-path active attacks, the situation is the same as for a 998 passive attack, where either a device has to already be compromised 999 or a device can be inserted into the path. For off-path active 1000 attacks, the attacks are generally limited to message insertion or 1001 modification where the attacker may wish to load illegal software or 1002 configuration files to an infrastructure device. 1004 Note that similar issues are relevant when software updates are 1005 downloaded from a vendor site to an ISPs network management system 1006 that is responsible for software updates and/or configuration 1007 information. 1009 2.5.1.1. Confidentiality Violations 1011 Confidentiality violations can occur when a miscreant intercepts any 1012 of the software image or configuration information. The software 1013 image may give an indication of exploits which the device is 1014 vulnerable to while the configuration information can inadvertently 1015 lead attackers to identify critical infrastructure IP addresses and 1016 passwords. 1018 2.5.1.2. Offline Cryptographic Attacks 1020 If any cryptographic mechanism was used to provide for data integrity 1021 and confidentiality, an offline cryptographic attack could 1022 potentially compromise the data. The traffic would need to be 1023 captured either by eavesdropping on the communication path or by 1024 being able to divert traffic to a malicious user. 1026 2.5.1.3. Replay Attacks 1028 For a replay attack to be successful, the software image or 1029 configuration file would need to first be captured either on-path or 1030 diverted to an attacker to later be replayed to the intended 1031 recipient. Additionally, since many protocols do have replay 1032 protection capabilities, these would have to be subverted as well in 1033 applicable situations. 1035 2.5.1.4. Message Insertion/Deletion/Modification 1037 Software images and configuration files can be manipulated by someone 1038 in control of intermediate hosts. By forging an IP address and 1039 impersonating a valid host which can download software images or 1040 configuration files, invalid files can be downloaded to an 1041 infrastructure device. This can also be the case from trusted 1042 vendors who may unbeknownst to them have compromised trusted hosts. 1043 An invalid software image or configuration file can cause a device to 1044 hang and become inoperable. Spoofed configuration files can be hard 1045 to detect, especially when the only added command is to allow a 1046 miscreant access to that device by entering a filter allowing a 1047 specific host access and configuring a local username/password 1048 database entry for authentication to that device. 1050 2.5.1.5. Man-In-The-Middle 1052 A man-in-the-middle attack attacks the identity of a communicating 1053 peer rather than the data stream itself. The attacker intercepts 1054 traffic that is sent between the infrastructure device and the host 1055 used to upload/download the system image or configuration file. He/ 1056 she can then act on behalf of one or both of these systems. 1058 If an attacker obtained a copy of the software image being deployed, 1059 he could potentially exploit a known vulnerability and gain access to 1060 the system. From a captured configuration file, he could obtain 1061 confidential network topology information or even more damaging 1062 information if any of the passwords in the configuration file were 1063 not encrypted. 1065 2.5.2. Security Practices 1067 Images and configurations are stored on specific hosts which have 1068 limited access. All access and activity relating to these hosts are 1069 authenticated and logged via AAA services. When uploaded/downloading 1070 any system software or configuration files, either TFTP, FTP or SCP 1071 can be used. Where possible, SCP is used to secure the data transfer 1072 and FTP is generally never used. All SCP access is username/password 1073 authenticated but since this requires an interactive shell, most ISPs 1074 will use shared key authentication to avoid the interactive shell. 1075 While TFTP access does not have any security measures, it is still 1076 widely used especially in OOB management scenarios. Some ISPs 1077 implement IP-based restriction on the TFTP server while some custom 1078 written TFTP servers will support MAC-based authentication. The MAC- 1079 based authentication is more common when using TFTP to bootstrap 1080 routers remotely using TFTP. 1082 In most environments scripts are used for maintaining the images and 1083 configurations of a large number of routers. To ensure the integrity 1084 of the configurations, every hour the configuration files are polled 1085 and compared to the previously polled version to find discrepancies. 1086 In at least one environment these tools are Kerberized to take 1087 advantage of automated authentication (not confidentiality). 1088 'Rancid' is one popular publicly available tool for detecting 1089 configuration and system changes. 1091 Filters are used to limit access to uploading/downloading 1092 configuration files and system images to specific IP addresses and 1093 protocols. 1095 The software images perform CRC-checks and the system binaries use 1096 the MD5 algorithm to validate integrity. Many ISPs expressed 1097 interest in having software image integrity validation based on the 1098 MD5 algorithm for enhanced security. 1100 In all configuration files, most passwords are stored in an encrypted 1101 format. Note that the encryption techniques used in varying products 1102 can vary and that some weaker encryption schemes may be subject to 1103 off-line dictionary attacks. This includes passwords for user 1104 authentication, MD5-authentication shared secrets, AAA server shared 1105 secrets, NTP shared secrets, etc. For older software which may not 1106 support this functionality, configuration files may contain some 1107 passwords in readable format. Most ISPs mitigate any risk of 1108 password compromise by either storing these configuration files 1109 without the password lines or by requiring authenticated and 1110 authorized access to the configuration files which are stored on 1111 protected OOB management devices. 1113 Automated security validation is performed on infrastructure devices 1114 using nmap and nessus to ensure valid configuration against many of 1115 the well-known attacks. 1117 2.5.3. Security Services 1119 o User Authentication - All users are authenticated before being 1120 able to download/upload any system images or configuration files. 1122 o User Authorization - All authenticated users are granted specific 1123 privileges to download or upload system images and/or 1124 configuration files. 1126 o Data Origin Authentication - Filters are used to limit access to 1127 uploading/downloading configuration files and system images to 1128 specific IP addresses. 1130 o Access Control - Filters are used to limit access to uploading/ 1131 downloading configuration files and system images to specific IP 1132 addresses and protocols. 1134 o Data Integrity - All systems use either a CRC-check or MD5 1135 authentication to ensure data integrity. Also tools such as 1136 rancid are used to automatically detect configuration changes. 1138 o Data Confidentiality - If the SCP protocol is used then there is 1139 confidentiality of the downloaded/uploaded configuration files and 1140 system images. 1142 o Auditing / Logging - All access and activity relating to 1143 downloading/uploading system images and configuration files are 1144 logged via AAA services and filter exception rules. 1146 o DoS Mitigation - TBD 1148 2.5.4. Additional Considerations 1150 Where the MD5 algorithm is not used to perform data integrity 1151 checking of software images and configuration files, ISPs have 1152 expressed an interest in having this functionality. IPsec is 1153 considered too cumbersome and operationally difficult to use for data 1154 integrity and confidentiality. 1156 2.6. Logging Considerations 1158 Although logging is part of all the previous sections, it is 1159 important enough to be covered as a separate item. The main issues 1160 revolve around what gets logged, how long are logs kept and what 1161 mechanisms are used to secure the logged information while it is in 1162 transit and while it is stored. 1164 2.6.1. Threats / Attacks 1166 Attacks on the logged data can be both from passive or active 1167 sources. Passive attacks are possible if someone has the capability 1168 to intercept data between the recipient logging server and the device 1169 the logged data originated from. This can be accomplished if a 1170 single infrastructure device is somehow compromised and can act as a 1171 network sniffer or if it is possible to insert a new device which 1172 acts as a network sniffer. 1174 Active attacks are possible for both on-path and off-path scenarios. 1175 For on-path active attacks, the situation is the same as for a 1176 passive attack, where either a device has to already be compromised 1177 or a device can be inserted into the path. For off-path active 1178 attacks, the attacks are generally limited to message insertion or 1179 modification which can alter the logged data to keep any compromise 1180 from being detected or to destroy any evidence which could be used 1181 for criminal prosecution. 1183 2.6.1.1. Confidentiality Violations 1185 Confidentiality violations can occur when a miscreant intercepts any 1186 of the logging data which is in transit on the network. This could 1187 lead to privacy violations if some of the logged data has not been 1188 sanitized to disallow any data that could be a violation of privacy 1189 to be included in the logged data. 1191 2.6.1.2. Offline Cryptographic Attacks 1193 If any cryptographic mechanism was used to provide for data integrity 1194 and confidentiality, an offline cryptographic attack could 1195 potentially compromise the data. The traffic would need to be 1196 captured either by eavesdropping on the network or by being able to 1197 divert traffic to a malicious user. 1199 2.6.1.3. Replay Attacks 1201 For a replay attack to be successful, the logging data would need to 1202 first be captured either on-path or diverted to an attacker and later 1203 replayed to the recipient. 1205 2.6.1.4. Message Insertion/Deletion/Modification 1207 Logging data could be injected, deleted or modified by someone in 1208 control of intermediate hosts. Logging data can also be injected by 1209 forging packets from either legitimate or illegitimate IP addresses. 1211 2.6.1.5. Man-In-The-Middle 1213 A man-in-the-middle attack attacks the identity of a communicating 1214 peer rather than the data stream itself. The attacker intercepts 1215 traffic that is sent between the infrastructure device and the 1216 logging server or traffic sent between the logging server and the 1217 database which is used to archive the logged data. Any unauthorized 1218 access to logging information could lead to knowledge of private and 1219 proprietary network topology information which could be used to 1220 compromise portions of the network. An additional concern is having 1221 access to logging information which could be deleted or modified so 1222 as to cover any traces of a security breach. 1224 2.6.2. Security Practices 1226 Logging is mostly performed on an exception auditing basis when it 1227 comes to filtering (i.e. traffic which is NOT allowed is logged). 1228 This is to assure that the logging servers are not overwhelmed with 1229 data which would render most logs unusable. Typically the data 1230 logged will contain the source and destination IP addresses and layer 1231 4 port numbers as well as a timestamp. The syslog protocol is used 1232 to transfer the logged data between the infrastructure device to the 1233 syslog server. Many ISPs use the OOB management network to transfer 1234 syslog data since there is virtually no security performed between 1235 the syslog server and the device. All ISPs have multiple syslog 1236 servers - some ISPs choose to use separate syslog servers for varying 1237 infrastructure devices (i.e. one syslog server for backbone routers, 1238 one syslog server for customer edge routers, etc.) 1240 The timestamp is derived from NTP which is generally configured as a 1241 flat hierarchy at stratum1 and stratum2 to have less configuration 1242 and less maintenance. Consistency of configuration and redundancy is 1243 the primary goal. Each router is configured with several stratum1 1244 server sources, which are chosen to ensure that proper NTP time is 1245 available even in the event of varying network outages. 1247 In addition to logging filtering exceptions, the following is 1248 typically logged: Routing protocol state changes, all device access 1249 (regardless of authentication success or failure), all commands 1250 issued to a device, all configuration changes and all router events 1251 (boot-up/flaps). 1253 The main function of any of these log messages is to see what the 1254 device is doing as well as to try and ascertain what certain 1255 malicious attackers are trying to do. Since syslog is an unreliable 1256 protocol, when routers boot or lose adjacencies, not all messages 1257 will get delivered to the remote syslog server. Some vendors may 1258 implement syslog buffering (e.g., buffer the messages until you have 1259 a route to the syslog destination) but this is not standard. 1260 Therefore, operators often have to look at local syslog information 1261 on a device (which typically has very little memory allocated to it) 1262 to make up for the fact that the server-based syslog files can be 1263 incomplete. Some ISPs also put in passive devices to see routing 1264 updates and withdrawals and do not rely solely on the device for log 1265 files. This provides a backup mechanism to see what is going on in 1266 the network in the event that a device may 'forget' to do syslog if 1267 the CPU is busy. 1269 The logs from the various syslog server devices are generally 1270 transferred into databases at a set interval which can be anywhere 1271 from every 10 minutes to every hour. One ISP uses Rsync to push the 1272 data into a database and then the information is sorted manually by 1273 someone SSH'ing to that database. 1275 2.6.3. Security Services 1277 o User Authentication - Not applicable 1279 o User Authorization - Not applicable 1281 o Data Origin Authentication - Not implemented 1283 o Access Control - Filtering on logging host and server IP address 1284 to ensure that syslog information only goes to specific syslog 1285 hosts. 1287 o Data Integrity - Not implemented 1289 o Data Confidentiality - Not implemented 1291 o Auditing / Logging - This entire section deals with logging. 1293 o DoS Mitigation - An OOB management system is used and sometimes 1294 different syslog servers are used for logging information from 1295 varying equipment. Exception logging tries to keep information to 1296 a minimum. 1298 2.6.4. Additional Considerations 1300 There is no security with syslog and ISPs are fully cognizant of 1301 this. IPsec is considered too operationally expensive and cumbersome 1302 to deploy. Syslog-ng and stunnel are being looked at for providing 1303 better authenticated and integrity protected solutions. Mechanisms 1304 to prevent unauthorized personnel from tampering with logs is 1305 constrained to auditing who has access to the logging servers and 1306 files. 1308 ISPs expressed requirements for more than just UDP syslog. 1309 Additionally, they would like more granular and flexible facilities 1310 and priorities, i.e. specific logs to specific servers. Also, a 1311 common format for reporting standard events so that they don't have 1312 to modify parsers after each upgrade of vendor device or software. 1314 2.7. Filtering Considerations 1316 Although filtering has been covered under many of the previous 1317 sections, this section will provide some more insights to the 1318 filtering considerations that are currently being taken into account. 1319 Filtering is now being categorized into three specific areas: data 1320 plane, management plane and routing control plane. 1322 2.7.1. Data Plane Filtering 1324 Data plane filters control the traffic that traverses through a 1325 device and affect transit traffic. Most ISPs deploy these kinds of 1326 filters at the customer facing edge devices to mitigate spoofing 1327 attacks using BCP38 and BCP84 guidelines. 1329 2.7.2. Management Plane Filtering 1331 Management filters control the traffic to and from a device. All of 1332 the protocols which are used for device management fall under this 1333 category and includes SSH, Telnet, SNMP, NTP, HTTP, DNS, TFTP, FTP, 1334 SCP and Syslog. This type of traffic is often filtered per interface 1335 and is based on any combination of protocol, source and destination 1336 IP address and source and destination port number. Some devices 1337 support functionality to apply management filters to the device 1338 rather than to the specific interfaces (e.g. receive ACL or loopback 1339 interface ACL) which is gaining wider acceptance. Note that logging 1340 the filtering rules can today place a burden on many systems and more 1341 granularity is often required to more specifically log the required 1342 exceptions. 1344 Any services that are not specifically used are turned off. 1346 IPv6 networks require the use of specific ICMP messages for proper 1347 protocol operation. Therefore, ICMP cannot be completely filtered to 1348 and from a device. Instead, granular ICMPv6 filtering is always 1349 deployed to allow for specific ICMPv6 types to be sourced or destined 1350 to a network device. A good guideline for IPv6 filtering is in the 1351 draft work in progress on Recommendations for Filtering ICMPv6 1352 Messages in Firewalls [I-D.ietf-v6ops-icmpv6-filtering-recs]. 1354 2.7.3. Routing Control Plane Filtering 1356 Routing filters are used to control the flow of routing information. 1357 In IPv6 networks, some providers are liberal in accepting /48s due to 1358 the still unresolved multihoming issues while others filter at 1359 allocation boundaries which are typically at /32. Any announcement 1360 received that is longer than a /48 for IPv6 routing and a /24 for 1361 IPv4 routing is filtered out of eBGP. Note that this is for non- 1362 customer traffic. Most ISPs will accept any agreed upon prefix 1363 length from its customer(s). 1365 2.8. Denial of Service Tracking / Tracing 1367 Denial of Service attacks are an ever increasing problem and require 1368 vast amounts of resources to combat effectively. Some large ISPs do 1369 not concern themselves with attack streams that are less than 1G in 1370 bandwidth - this is on the larger pipes where 1G is essentially less 1371 than 5% of offered load. This is largely due to the large amounts of 1372 DDoS traffic which continually requires investigation and mitigation. 1373 At last count the number of hosts making up large distributed DoS 1374 botnets exceeded 1 million hosts. 1376 New techniques are continually evolving to automate the process of 1377 detecting DoS sources and mitigating any adverse effects as quickly 1378 as possible. At this time, ISPs are using a variety of mitigation 1379 techniques including: sink hole routing, black-hole triggered 1380 routing, uRPF, rate limiting and specific control plane traffic 1381 enhancements. Each of these techniques will be detailed below. 1383 2.8.1. Sink Hole Routing 1385 Sink hole routing refers to injecting a more specific route for any 1386 known attack traffic which will ensure that the malicious traffic is 1387 redirected to a valid device or specific system where it can be 1388 analyzed. 1390 2.8.2. Black-Hole Triggered Routing 1392 Black-hole triggered routing (also referred to as Remote Triggered 1393 Black Hole Filtering) is a technique where the BGP routing protocol 1394 is used to propagate routes which in turn redirects attack traffic to 1395 the null interface where it is effectively dropped. This technique 1396 is often used in large routing infrastructures since BGP can 1397 propagate the information in a fast effective manner as opposed to 1398 using any packet-based filtering techniques on hundreds or thousands 1399 of routers. [refer to the following NANOG presentation for a more 1400 complete description http://www.nanog.org/mtg-0402/pdf/morrow.pdf] 1401 Note that this black-holing technique may actually fulfill the goal 1402 of the attacker if the goal was to instigate blackholing traffic 1403 which appeared to come from a certain site. On the other hand, this 1404 blackhole technique can decrease the collateral damage caused by an 1405 overly large attack aimed at something other than critical services. 1407 2.8.3. Unicast Reverse Path Forwarding 1409 Unicast Reverse Path Forwarding (uRPF) is a mechanism for validating 1410 whether an incoming packet has a legitimate source address or not. 1411 It has two modes: strict mode and loose mode. In strict mode, uRPF 1412 checks whether the incoming packet has a source address that matches 1413 a prefix in the routing table, and whether the interface expects to 1414 receive a packet with this source address prefix. If the incoming 1415 packet fails the unicast RPF check, the packet is not accepted on the 1416 incoming interface. Loose mode uRPF is not as specific and the 1417 incoming packet is accepted if there is any route in the routing 1418 table for the source address. 1420 While BCP84 [RFC3704] and a study on uRPF experiences [I-D.savola- 1421 bcp84-urpf-experiences] detail how asymmetry, i.e. multiple routes to 1422 the source of a packet, does not preclude applying feasible paths 1423 strict uRPF, it is generally not used on interfaces that are likely 1424 to have routing asymmetry. Usually for the larger ISPs, uRPF is 1425 placed at the customer edge of a network. 1427 2.8.4. Rate Limiting 1429 Rate limiting refers to allocating a specific amount of bandwidth or 1430 packets per second to specific traffic types. This technique is 1431 widely used to mitigate well-known protocol attacks such as the TCP- 1432 SYN attack where a large number of resources get allocated for 1433 spoofed TCP traffic. Although this technique does not stop an 1434 attack, it can sometimes lessen the damage and impact on a specific 1435 service. However, it can also make the impact of a DDoS attack much 1436 worse if the rate limiting is impacting (i.e. discarding) more 1437 legitimate traffic. 1439 2.8.5. Specific Control Plane Traffic Enhancements 1441 Some ISPs are starting to use capabilities which are available from 1442 some vendors to simplify the filtering and rate-limiting of control 1443 traffic. Control traffic here refers to the routing control plane 1444 and management plane traffic that requires CPU cycles. A DoS attack 1445 against any control plane traffic can therefore be much more damaging 1446 to a critical device than other types of traffic. No consistent 1447 deployment of this capability was found at the time of this writing. 1449 3. Security Considerations 1451 This entire document deals with current security practices in large 1452 ISP environments. It lists specific practices used in today's 1453 environments and as such does not in itself pose any security risk. 1455 4. References 1457 4.1. Normative References 1459 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1460 Requirement Levels", BCP 14, RFC 2119, March 1997. 1462 [RFC2827] Ferguson, P. and D. Senie, "Network Ingress Filtering: 1463 Defeating Denial of Service Attacks which employ IP Source 1464 Address Spoofing", BCP 38, RFC 2827, May 2000. 1466 [RFC2828] Shirey, R., "Internet Security Glossary", RFC 2828, 1467 May 2000. 1469 [RFC3552] Rescorla, E. and B. Korver, "Guidelines for Writing RFC 1470 Text on Security Considerations", BCP 72, RFC 3552, 1471 July 2003. 1473 [RFC3682] Gill, V., Heasley, J., and D. Meyer, "The Generalized TTL 1474 Security Mechanism (GTSM)", RFC 3682, February 2004. 1476 [RFC3704] Baker, F. and P. Savola, "Ingress Filtering for Multihomed 1477 Networks", BCP 84, RFC 3704, March 2004. 1479 4.2. Informational References 1481 [I-D.ietf-v6ops-icmpv6-filtering-recs] 1482 Davies, E. and J. Mohacsi, "Recommendations for Filtering 1483 ICMPv6 Messages in Firewalls", 1484 draft-ietf-v6ops-icmpv6-filtering-recs-02 (work in 1485 progress), July 2006. 1487 [I-D.lewis-infrastructure-security] 1488 Lewis, D., "Service Provider Infrastructure Security", 1489 draft-lewis-infrastructure-security-00 (work in progress), 1490 June 2006. 1492 [I-D.savola-bcp84-urpf-experiences] 1493 Savola, P., "Experiences from Using Unicast RPF", 1494 draft-savola-bcp84-urpf-experiences-01 (work in progress), 1495 June 2006. 1497 [I-D.savola-rtgwg-backbone-attacks] 1498 Savola, P., "Backbone Infrastructure Attacks and 1499 Protections", draft-savola-rtgwg-backbone-attacks-02 (work 1500 in progress), July 2006. 1502 Appendix A. Acknowledgments 1504 The editor gratefully acknowledges the contributions of: George 1505 Jones, who has been instrumental in providing guidance and direction 1506 for this document and the insighful comments from Ross Callon, Ron 1507 Bonica, Gaurab Upadhaya, Warren Kumari, Pekka Savola, Fernando Gont, 1508 Chris Morrow, Donald Smith and the numerous ISP operators who 1509 supplied the information which is depicted in this document. 1511 Appendix B. Protocol Specific Attacks 1513 This section will list many of the traditional protocol based attacks 1514 which have been observed over the years to cause malformed packets 1515 and/or exploit protocol deficiencies. Note that they all exploit 1516 vulnerabilities in the actual protocol itself and often, additional 1517 authentication and auditing mechanisms are now used to detect and 1518 mitigate the impact of these attacks. The list is not exhaustive but 1519 is a fraction of the representation of what types of attacks are 1520 possible for varying protocols. 1522 B.1. Layer 2 Attacks 1524 o ARP Flooding 1526 B.2. IPv4 Protocol Based Attacks 1528 o IP Addresses, either source or destination, can be spoofed which 1529 in turn can circumvent established filtering rules. 1531 o IP Source Route Option can allows attackers to establish stealth 1532 TCP connections 1534 o IP Record Route Option can discloses information about the 1535 topology of the network. 1537 o IP header that is too long or too short can cause DoS attacks to 1538 devices. 1540 o IP Timestamp Option can leak information which can be used to 1541 discern network behavior. 1543 o Fragmentation attacks which can vary widely - more detailed 1544 information can be found at http://www-src.lip6.fr/homepages/ 1545 Fabrice.Legond-Aubry/www.ouah.org/fragma.html 1547 o IP ToS field (or the Differentiated Services (DSCP) field) can be 1548 used to reroute or reclassify traffic based on specified 1549 precedence. 1551 o IP checksum field has been used for scanning purposes, for example 1552 when some firewalls did not check the checksum and allowed an 1553 attacker to differentiate when the response came from an end- 1554 system, and when from a firewall 1556 o IP TTL field can be used to bypass certain network based intrusion 1557 detection systems and to map network behavior. 1559 B.2.1. Higher Layer Protocol Attacks 1561 The following lists additional attacks but does not explicitly 1562 numerate them in detail. It is for informational purposes only. 1564 o IGMP oversized packet 1566 o ICMP Source Quench 1568 o ICMP Mask Request 1570 o ICMP Large Packet (> 1472) 1572 o ICMP Oversized packet (>65536) 1574 o ICMP Flood 1576 o ICMP Broadcast w/ Spoofed Source (Smurf Attack) 1578 o ICMP Error Packet Flood 1580 o ICMP Spoofed Unreachable 1582 o TCP Packet without Flag 1584 o TCP Oversized Packet 1586 o TCP FIN bit with no ACK bit 1588 o TCP Packet with URG/OOB flag (Nuke Attack) 1590 o SYN Fragments 1592 o SYN Flood 1594 o SYN with IP Spoofing (Land Attack) 1596 o SYN and FIN bits set 1598 o TCP port scan attack 1600 o UDP spoofed broadcast echo (Fraggle Attack) 1601 o UDP attack on diagnostic ports (Pepsi Attack) 1603 B.3. IPv6 Attacks 1605 Any of the above-mentioned IPv4 attacks could be used in IPv6 1606 networks with the exception of any fragmentation and broadcast 1607 traffic, which operate differently in IPv6. Note that all of these 1608 attacks are based on either spoofing or misusing any part of the 1609 protocol field(s). 1611 Today, IPv6 enabled hosts are starting to be used to create IPv6 1612 tunnels which can effectively hide botnet and other malicious traffic 1613 if firewalls and network flow collection tools are not capable of 1614 detecting this traffic. The security measures used for protecting 1615 IPv6 infrastructures should be the same as in IPv4 networks but with 1616 additional considerations for IPv6 network operations which may be 1617 different from IPv4. 1619 Author's Address 1621 Merike Kaeo 1622 Double Shot Security, Inc. 1623 3518 Fremont Avenue North #363 1624 Seattle, WA 98103 1625 U.S.A. 1627 Phone: +1 310 866 0165 1628 Email: merike@doubleshotsecurity.com 1630 Intellectual Property Statement 1632 The IETF takes no position regarding the validity or scope of any 1633 Intellectual Property Rights or other rights that might be claimed to 1634 pertain to the implementation or use of the technology described in 1635 this document or the extent to which any license under such rights 1636 might or might not be available; nor does it represent that it has 1637 made any independent effort to identify any such rights. Information 1638 on the procedures with respect to rights in RFC documents can be 1639 found in BCP 78 and BCP 79. 1641 Copies of IPR disclosures made to the IETF Secretariat and any 1642 assurances of licenses to be made available, or the result of an 1643 attempt made to obtain a general license or permission for the use of 1644 such proprietary rights by implementers or users of this 1645 specification can be obtained from the IETF on-line IPR repository at 1646 http://www.ietf.org/ipr. 1648 The IETF invites any interested party to bring to its attention any 1649 copyrights, patents or patent applications, or other proprietary 1650 rights that may cover technology that may be required to implement 1651 this standard. Please address the information to the IETF at 1652 ietf-ipr@ietf.org. 1654 Disclaimer of Validity 1656 This document and the information contained herein are provided on an 1657 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS 1658 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET 1659 ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, 1660 INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE 1661 INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED 1662 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 1664 Copyright Statement 1666 Copyright (C) The Internet Society (2006). This document is subject 1667 to the rights, licenses and restrictions contained in BCP 78, and 1668 except as set forth therein, the authors retain all their rights. 1670 Acknowledgment 1672 Funding for the RFC Editor function is currently provided by the 1673 Internet Society.