idnits 2.17.1 draft-ietf-rtfm-applicability-statement-02.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** Looks like you're using RFC 2026 boilerplate. This must be updated to follow RFC 3978/3979, as updated by RFC 4748. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- ** Missing expiration date. The document expiration date should appear on the first and last page. == 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 Introduction section. ** 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.) ** The document seems to lack separate sections for Informative/Normative References. All references will be assumed normative when checking for downward references. ** There is 1 instance of too long lines in the document, the longest one being 1 character in excess of 72. Miscellaneous warnings: ---------------------------------------------------------------------------- -- 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 (October 1999) is 8958 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) ** Downref: Normative reference to an Informational RFC: RFC 1272 (ref. '1') ** Obsolete normative reference: RFC 2063 (ref. '2') (Obsoleted by RFC 2722) ** Obsolete normative reference: RFC 2064 (ref. '3') (Obsoleted by RFC 2720) -- Possible downref: Non-RFC (?) normative reference: ref. '4' -- Possible downref: Non-RFC (?) normative reference: ref. '5' ** Downref: Normative reference to an Informational RFC: RFC 2123 (ref. '6') Summary: 10 errors (**), 0 flaws (~~), 1 warning (==), 4 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 1 Internet Engineering Task Force Nevil Brownlee 2 INTERNET-DRAFT The University of Auckland 3 April 1999 4 Expires October 1999 6 RTFM: Applicability Statement 8 10 Status of this Memo 12 This document is an Internet-Draft and is in full conformance with all 13 provisions of Section 10 of RFC2026. 15 Internet-Drafts are working documents of the Internet Engineering Task 16 Force (IETF), its areas, and its working groups. Note that other groups 17 may also distribute working documents as Internet-Drafts. 19 Internet-Drafts are draft documents valid for a maximum of six months 20 and may be updated, replaced, or obsoleted by other documents at any 21 time. It is inappropriate to use Internet-Drafts as reference material 22 or to cite them other than as "work in progress." 24 The list of current Internet-Drafts can be accessed at 25 http://www.ietf.org/ietf/1id-abstracts.txt 27 The list of Internet-Draft Shadow Directories can be accessed at 28 http://www.ietf.org/shadow.html. 30 This Internet Draft is a product of the Realtime Traffic Flow 31 Measurement Working Group of the IETF. 33 Abstract 35 This document provides an overview covering all aspects of Realtime 36 Traffic Flow Measurement, including its area of applicability and its 37 limitations. 39 Contents 41 1 The RTFM Documents 2 43 2 Brief Technical Specification (TS) 4 45 3 Applicability Statement (AS) 4 47 4 Limitations 5 49 5 Security Considerations 6 51 6 Policy Considerations 6 53 7 Soundness 7 55 8 Appendix A: WG Report on the Meter MIB 8 57 9 References 9 59 10 Author's Address 10 61 1 The RTFM Documents 63 The RTFM Traffic Measurement System has been developed by the Realtime 64 Traffic Flow Measurement Working Group. It is described in six other 65 documents (references [1] to [6]), as follows: 67 [1] Internet Accounting: Background (Informational) 69 Sets out the requirements for a usage reporting system for 70 network traffic. Sketches out the RTFM Architecture (meters, 71 meter readers and managers) allowing for multiple meters 72 and meter readers, with asynchronous reading from the meters. 73 Proposes methods of classifying traffic flows, the need for 74 flows to be bi-directional (with separate sets of counters 75 for each direction) and the need for each packet to be counted 76 in a single flow (the 'count in one bucket' principle). 78 [2] RTFM Architecture (Informational) 80 Defines the RTFM Architecture, giving descriptions of each 81 component. Explains how traffic flows are viewed as logical 82 entities described in terms of their address-attribute values, 83 so that each is defined by the attributes of its end-points. 84 Gives a detailed description of the RTFM traffic meter, with 85 full details of how flows are stored in the meter's flow table, 86 and how packets are matched in accordance with rules stored in 87 a ruleset. 89 [3] RTFM Meter MIB (Proposed Standard) 91 Describes the SNMP Management Information Base for an RTFM 92 meter, including its flow table, rule table (storing the 93 meter's rulesets) and the control tables used for managing 94 a meter and reading flow data from it. 96 [4] SRL: A Language for Describing Traffic Flows (Informational) 97 and Specifying Actions for Flow Groups 99 An RTFM ruleset is an array of rules, used by the meter to 100 decide which flows are of interest, which end-point is the 101 flow source, and how much detail (i.e. what attribute values) 102 must be saved for each flow. SRL is a high-level language 103 providing a clear, logical way to write rulesets. It should 104 also be useful for other applications which select flows 105 and perform actions upon them, e.g. packet-marking gateways, 106 RSVP policy agents, etc. 108 [5] RTFM New Attributes (Experimental) 110 There has been considerable interest from users in extending 111 the RTFM Architecture so as to allow a meter to report on 112 an increased number of flow-related measures. This 113 RFC documents work on specifying such measures (the 'new' 114 attributes) and reports on experience of implementing them. 116 [6] RTFM: Experiences with NeTraMet (Informational) 118 NeTraMet is a free software implementation of the RTFM 119 Architecture which has been available since 1993. This RFC 120 records RTFM implementation experience gained with NeTraMet 121 up to late 1996. One particularly important result is the 122 realisation that groups of rules which test the same attribute 123 using the same mask can be implemented as a single hashed 124 comparison, allowing the meter to rapidly determine whether a 125 packet belongs to one of a large number of networks. 127 2 Brief Technical Specification (TS) 129 RTFM provides for the measurement of network traffic 'flows,' i.e. 131 - a method of specifying traffic flows within 132 a network 133 - a hierarchy of devices (meters, meter readers, managers) 134 for measuring the specified flows 135 - a mechanism for configuring meters and meter readers, 136 and for collecting the flow data from remote meters 138 RTFM provides high time resolution for flow first- and last-packet 139 times. Counters for long-duration flows may be read at intervals 140 determined by a manager. The RTFM Meter is designed so as to do as much 141 data reduction work as possible, which minimizes the amount of data to 142 be read and the amount of processing needed to produce useful reports 143 from it. 145 RTFM flow data can be used for a wide range of purposes, such as usage 146 accounting, long-term recording of network usage (classified by IP 147 address attributes) and real-time analysis of traffic flows at remote 148 metering points. 150 3 Applicability Statement (AS) 152 To use RTFM for collecting network traffic information one must first 153 consider where in the network traffic flows are to be measured. Once 154 that is decided, an RTFM Meter must be installed at each chosen 155 measurement point. 157 At least one Meter Reader is needed to collect the measured data from 158 the meters, and a single Manager is needed to control the meters and 159 meter readers. 161 RTFM Meters may be single- or multi-user hosts running a meter program 162 (one such program is available as free software, a second is under 163 development at IBM Research). Alternatively, meters could be run as 164 firmware in switches or routers. A hybrid approach in which an RTFM 165 meter takes raw traffic data from a router provides another useful 166 implementation path. 168 RTFM Managers are programs running on a host, communicating with meters 169 and meter readers via the network. For this purpose meters are SNMP 170 agents implementing the RTFM Meter MIB, and managers are SNMP clients 171 using the Meter MIB to store and access the flow data. 173 4 Limitations 175 RTFM is designed to measure traffic flows for traffic passing a point in 176 a network. If packets for a flow pass the metering point in both 177 directions the meter will match them up, providing counters for each 178 direction. If packets only pass in one direction the meter can only 179 provide counts for that direction. 181 Users of RTFM should note that installing meters, meter readers and 182 managers merely provides one with the capability to collect flow data. 183 Further installation work will be needed to develop configuration files 184 (RTFM rulesets) for each meter, data processing applications to analyse 185 the flow data, and various scripts, cron jobs, etc. so as to create a 186 useful production-quality measurement system which suits a user's 187 particular needs. 189 One of the strengths of RTFM is its ability to collect flow data at 190 whatever level of detail (or 'granularity') is required. It can be 191 tempting to simply collect 'all possible data,' but there are severe 192 resource constraints. If one tries to save the complete 193 address-attribute value for all attributes of every possible flow a very 194 large amount of data may be produced rapidly, but the meter has only a 195 finite amount of memory for its flow table. A better approach is to 196 save the minimum amount of data required to achieve the measurement 197 system goals. 199 For example, to collect usage data so as to bill subscribers identified 200 by their IP address one could just save the full IP address, nothing 201 more. The RTFM meter would produce flow data for each subscriber IP 202 address, with PDU and Octet counts for data sent and received, which 203 would be the minimum needed to produce bills. In practice one would 204 probably want to save at least part of the Destination IP address, which 205 would allow the production of usage logs showing subscriber activity 206 over time. 208 The simplest way to determine how much detail can be collected is to 209 create an initial ruleset which collects the minimum amount, then to 210 modify it step by step, gradually increasing the amount of information 211 saved for each flow. An RTFM meter ought to provide some measures of 212 its own performance (e.g. number of active flows, percentage idle 213 processor time, packets metered, packets not metered). Such measures 214 will be implementation-specific, but should allow a user to assess the 215 impact of each change to the ruleset. 217 If the network data rate is too high, i.e. the meter reports that it 218 cannot meter all the packets even with the initial ruleset above, one 219 may be able to use other strategies. For example one could 220 - run the meter on a faster computer, e.g. move 221 from a DOS PC to a workstation, or perhaps use a 222 meter implemented in firmware within a switch or router. 224 - use sampling. The details of such sampling are not defined within 225 the RTFM Architecture, but the Meter MIB provides one simple 226 method by allowing one to specify that only every nth packet 227 on an interface will be metered. This would probably not be 228 acceptable for producing billing data, but might well be 229 acceptable for traffic engineering purposes. 231 5 Security Considerations 233 These are discussed in detail in the Architecture and Meter MIB 234 documents. In brief, an RTFM Meter is an SNMP agent which observes a 235 network and collects flow data from it. Since it doesn't control the 236 network directly, it has no direct effect on network security. 238 On the other hand, the flow data itself may well be valuable - to the 239 network operator (as billing data) or to an attacker (who may wish to 240 modify that data, or the meter's ruleset(s). It is therefore important 241 to take proper precautions to ensure that access to the meter and its 242 data is sufficiently secure. 244 For example, a meter port attached to a network should be passive, so 245 that it cannot respond to login attempts of any kind. Control and data 246 connections to a meter should be via a secure management network. 247 Finally, suitable security should be established for the meter, as it 248 would be for any other SNMP agent. 250 Meters may, like any other network component, be subjected to Denial of 251 Service and other attacks. These are outside the RTFM Architecture - 252 countermeasures for them are available, but are also outside RTFM. 254 6 Policy Considerations 256 When collecting traffic data, one must have well-defined operations 257 policies covering points such as: 259 - Exactly what data is to be collected, at what level of detail? 260 - How long will the data be kept? 261 - What may the data be used for? 262 - Who will be allowed to see the raw data? 263 - May summaries of the data be shown to other people? 265 Policy issues such as these should normally be considered as part of an 266 organisation's Network Security Policy. 268 Other policy issues relating more directly to the traffic data are 269 essentially part of the measurement system design, such as: 271 - How much time resolution is required for the data? 272 (Less resolution implies longer collection intervals, 273 but that may require more memory in the meters to hold 274 flow data between collections). 275 - What level of hardware redundancy is needed? 276 (A single meter and meter reader is generally enough. 277 For greater reliability, meters and meter readers can be 278 duplicated). 279 - Who is allowed to use the system? (Approved users will need 280 permissions to download rulesets to the meters, and to 281 collect their data, possibly via their own meter readers). 283 7 Soundness 285 NeTraMet, the first implementation of the RTFM Architecture, has been in 286 use worldwide since 1994. Currently there are many organisations, large 287 and small, using it to collect traffic data for billing purposes. 289 One example of these is Kawaihiko, the New Zealand Universities' 290 Network, which has seven RTFM meters located at sites throughout New 291 Zealand. One of the sites is NZIX, the New Zealand Internet eXchange at 292 the University of Waikato, where Kawaihiko has a meter (attached to a 293 100baseT network) observing traffic flows across the exchange to each of 294 Kawaihiko's three international Internet Service Providers. 5-minute 295 Octet counts are collected from all the Kawaihiko meters by a single 296 meter reader at Auckland. Traffic data from the meters is used to 297 determine the cost per month for each of the Kawaihiko sites. 299 It is difficult to estimate how many organisations are using RTFM 300 traffic measurement. There are about 250 people on the NeTraMet mailing 301 list, which often carries questions like 'why doesn't this ruleset do 302 what I meant?' Once new users have the system running, however, they 303 tend to simply use it without further comment. 305 >From time to time the list provides useful feedback. For example, early 306 in 1998 there were two very significant user contributions: 308 - Jacek Kowalski (Telstra, Melbourne) described an improved hash 309 algorithm for NeTraMet's flow table, which provided almost an 310 order of magnitude improvement in packet-handling performance. 312 - Kevin Hoadley (JANET, U.K.) reported having problems with very 313 large rulesets. These were resolved, and better methods of 314 downloading rules developed, allowing NeTraMet to work well 315 for rulesets with more than 32,000 rules. 317 Perhaps one reason why there is little discussion of NeTraMet's use in 318 collecting billing data is that users may consider that the way collect 319 their data is a commercially sensitive matter. 321 8 Appendix A: WG Report on the Meter MIB 323 The Meter MIB (in its current form) was developed early in 1996. It was 324 produced as an SNMPv2 MIB, following a number of detailed (and 325 continuing) discussions with David Perkins beginning at the Dallas IETF 326 meeting in December 1995. 328 There are two current implementations: 330 - NeTraMet (Nevil Brownlee, The University of Auckland) 332 - IBM Meter (Sig Handelman & Stephen Stibler, IBM Research, N.Y, 333 Bert Wijnen provided further help with SNMP) 335 The NeTraMet meter is a stand-alone SNMP agent using an SNMPv2C 336 implementation derived from CMU SNMPv2. 338 The IBM meter runs as a sub-agent on an AIX system. All the meter code 339 has been written by Stephen Stibler - it was not derived from the 340 NeTraMet code. Stephen has found it useful to use nifty, one of 341 NeTraMet's manager/reader programs, to test the IBM meter. 343 As indicated above, there have only been two implementors to date, and 345 the Working Group consensus has been very strong. 347 The MIB has one unusual aspect: the method used to read large amounts 348 of data from its Flow Table. An earlier SNMPv1 version of the MIB was 349 in use from 1992 to 1997; it used opaque objects to read column slices 350 from the flow table for flows which had been active since a specified 351 time. This was very non-standard (or at least very 352 application-specific). 354 With the change to SNMPv2 we were able to use 64-bit counters for PDUs 355 and Octets, RowStatus variables for control tables and GETBULK requests 356 to read rows from the flow table. We also use the TimeFilter convention 357 from the RMON2 MIB to select flows to be read; this gives the meter MIB 358 a strong resemblance to RMON2. 360 The current MIB introduces a better way of reading large amounts of data 361 from the flow table. This is the 'DataPackage' convention, which 362 specifies the attribute values to be read from a flow table row. The 363 meter returns the values for each required attribute within a 364 BER-encoded sequence. This means there is only one object identifier 365 for the whole sequence, greatly reducing the number of bytes required to 366 retrieve the data. The combination of 368 TimeFilter: to select the flows to be read 369 DataPackage: to select the attributes required for each flow 370 GetBulk: to read many flows with a single SNMP PDU 372 provides a very effective way to read flow data from a traffic meter. 374 9 References 376 [1] Mills, C., Hirsch, G. and Ruth, G., "Internet Accounting 377 Background", RFC 1272, Bolt Beranek and Newman Inc., Meridian 378 Technology Corporation, November 1991. 380 [2] Brownlee, N., Mills, C., and Ruth, G., "Traffic Flow 381 Measurement: Architecture", RFC 2063, The University of 382 Auckland, GTE Laboratories Inc, January 1997. 384 [3] Brownlee, N., "Traffic Flow Measurement: Meter MIB", RFC 385 2064, The University of Auckland, January 1997. 387 [4] Brownlee, N., "SRL: A Language for Describing Traffic Flows 388 and Specifying Actions for Flow Groups," Internet Draft, 389 'Working draft' to become an Informational RFC, The University 390 of Auckland. 392 [5] Handelman, S.W., Brownlee, N., Ruth, G., Stibler, S., "New 393 Attributes for Traffic Flow Measurment," Internet Draft, 394 'Working draft' to become an Experimental RFC, IBM, The 395 University of Auckland, GTE Laboratories Inc, IBM. 397 [6] Brownlee, N., "Traffic Flow Measurement: Experiences with 398 NeTraMet," RFC 2123, The University of Auckland, March 1997. 400 10 Author's Address 402 Nevil Brownlee 403 Information Technology Systems & Services 404 The University of Auckland 406 Phone: +64 9 373 7599 x8941 407 E-mail: n.brownlee@auckland.ac.nz 409 Expires October 1999