IPFIX Working Group L. Peluso Internet-Draft University of Napoli/Fraunhofer Intended status: Standards Track Institute FOKUS Expires: May 21, 2008 T. Zseby Fraunhofer Institute FOKUS S. D'Antonio CINI Consortium/ITeM Laboratory M. Molina DANTE November 18, 2007 Flow selection Techniques draft-peluso-flowselection-tech-01.txt Status of this Memo By submitting this Internet-Draft, each author represents that any applicable patent or other IPR claims of which he or she is aware have been or will be disclosed, and any of which he or she becomes aware will be disclosed, in accordance with Section 6 of BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet- Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt. The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html. This Internet-Draft will expire on May 21, 2008. Copyright Notice Copyright (C) The IETF Trust (2007). Abstract Flow selection is the process in charge of electing a limited number of flows from all of those observed at an observation point to be Peluso, et al. Expires May 21, 2008 [Page 1] Internet-Draft Flow selection Techniques November 2007 considered into the measurement process chain. The flow selection process can be enabled at different stages of the monitoring reference model. It can be performed at metering time once the packet classification has been executed, i.e. flow state dependent packet sampling, or at recording/exporting time by limiting the number of flows to be stored and/or exported to the collector applications. This document illustrates the motivations which might lead flow selection to be performed and presents a classification of the related techniques. The document furthermore provides the basis for the definition of information models for configuring flow selection techniques and discusses what information about the flow selection process is beneficial to be exported by adopting a suitable information model. Requirements Language The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 [RFC2119]. Peluso, et al. Expires May 21, 2008 [Page 2] Internet-Draft Flow selection Techniques November 2007 Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 2. Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 3. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4 3.1. Selection process related terminology . . . . . . . . . . 4 4. Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . 5 5. Flow selection techniques . . . . . . . . . . . . . . . . . . 6 5.1. Flow selection on flow record content . . . . . . . . . . 8 5.2. Flow selection on flow record arrival time . . . . . . . . 8 5.3. Flow selection on external events . . . . . . . . . . . . 8 6. Flow selection causes and relevant exportable information . . 9 6.1. Flow selection in the metering process . . . . . . . . . . 9 6.2. Flow selection in the flow recording process . . . . . . . 9 6.3. Flow selection in the exporting process . . . . . . . . . 11 6.4. Information model for flow selection information exporting . . . . . . . . . . . . . . . . . . . . . . . . 12 6.4.1. Meter process related . . . . . . . . . . . . . . . . 12 6.4.2. Flow recording process related . . . . . . . . . . . . 13 6.4.3. Flow exporting process related . . . . . . . . . . . . 14 6.5. Requirements put on implementations . . . . . . . . . . . 15 7. Information model for flow selection configuration . . . . . . 15 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 16 9. References . . . . . . . . . . . . . . . . . . . . . . . . . . 16 9.1. Normative References . . . . . . . . . . . . . . . . . . . 16 9.2. Informative References . . . . . . . . . . . . . . . . . . 16 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 17 Intellectual Property and Copyright Statements . . . . . . . . . . 19 Peluso, et al. Expires May 21, 2008 [Page 3] Internet-Draft Flow selection Techniques November 2007 1. Introduction 2. Scope The main aim of this document is to present flow selection techniques that can be performed by an IPFIX device and to define additional information that can be reported to keep under control/complement the flow selection process. This document does not intend to deal with the flow selection that might result from the sampling of packets in the metering process before that the classification process is performed. Although that approach leads to a natural selection of the flows that are generated once the classification process has been performed, packet sampling techniques are widely analysed in [PSAMP-TECH] and, therefore, outside the scope of this document. Instead, it describes those selection techniques that might be considered in order to enable flow selection by directly acting on traffic flows during the metering phase and/or the exporting phase. 3. Terminology The terminology used here is fully consistent with all terms listed in [IPFIX-ARCH] and [PSAMP-TECH] and includes additional terms required for the description of flow selection techniques. 3.1. Selection process related terminology In this section, some additional terms are presented which extend the terminology introduced in [PSAMP-TECH]. * Flow Selection Process A Flow Selection Process takes the set of the accounted Flow Records as its input and selects a subset of that set as its output. * Flow Selection State A Flow Selection Process may maintain state information for use by the Flow Selection Process. At a given time, the Flow Selection State may depend on flows observed at and before that time, and other variables. Examples include: Peluso, et al. Expires May 21, 2008 [Page 4] Internet-Draft Flow selection Techniques November 2007 (i) number of accounted flow records; (ii) memory space available for flow recording; (iii) state of the pseudorandom number generators; (iv) hash values calculated during selection. * Flow Selector A Flow Selector defines the action of a Flow Selection Process on a single flow of its input. The Flow Selector can make use of the following information in determining whether a flow is selected: (i) the content of the flow record; (ii) any information state related to the flow recording; (iii) any selection state that may be maintained by the Flow Selection Process. 4. Motivation As stated in [PSAMP-TECH], packet selection is in charge of electing a representative subset of packets that allow accurate estimates of properties of the unsampled traffic to be formed. Its main application consists in performing some forms of data reduction on observed Internet traffic in order to limit the processing overhead at measurement devices. Despite its proven ability in achieving this objective, the mechanism responsible for steering the selection process is generally driven by a packet-based decision strategy. It means that, the element on which this selection mechanism is performed is a packet and the decision of which packets are suitable to be sampled depends on packets only. As a consequence, depending on the specific adopted selection strategy, packet selection may not take in consideration potential effects of its actions on subsequent measurement tasks, such as flow recording and exporting processes, which are instead considering flows rather than packets. In this perspective, flow selection differs from packet selection in that the basis elements on which the selection process is applied is not a packet but a flow. In the IPFIX architecture the object of the selection process would be the so-colled flow records. It has been observed that the distribution of the number of packets per flow or the number of bytes per flow are heavy-tailed. That means, most flows consist only of a small number of packets and only few flows have a large number of packets. The few large flows contribute to the majority of the overall traffic volume [DuLT01a], [DuLT01b]. Peluso, et al. Expires May 21, 2008 [Page 5] Internet-Draft Flow selection Techniques November 2007 This observation on the flow size distributions in Internet traffic is also referred to as "Quasi-Zipf-Law" [KuXW04] or as "elephant and mice phenomenon". The large flows are referred to as elephant flows or heavy hitters. Obviously, distributions characterizing the flow size strictly depends on the flow definition in use and can change with regard to the profile of future applications. For several applications it makes sense to select only the flows of interest. 5. Flow selection techniques Figure 1 shows the IPFIX reference model as defined in [IPFIX-ARCH], and extends it by introducing the functional components where flow selection can take place. As previously mentioned, traffic flows can be selected at different stages of the measurement chain. The first possibility is to perform flow selection by analysing flow packets during the metering process. The second option is to act directly on the traffic flows during the flow recording process and/or the flow exporting process. Peluso, et al. Expires May 21, 2008 [Page 6] Internet-Draft Flow selection Techniques November 2007 Packet(s) coming in to Observation Point(s) | | v v +----------------+---------------------------+ +-----+-------+ | Metering Process on an | | | | Observation Point | | | | | | | | packet header capturing | | | | | |...| Metering | | timestamping | | Process N | | | | | | | packet selection | | | | | | | | | classification | | | | | | | | | flow state dependent packet sampling (*) | | | | | | | | | aggregation | | | | | | | | | flow recording (*) | | | | | | | | | | Timing out Flows | | | | | Handle resource overloads | | | +--------|-----------------------------------+ +-----|-------+ | | Flow Records (selected by Observation Domain) Flow Records | | +----------------------+----------------------+ | +----------------------|---------------+ | Exporting Process v | | +---------------+-----------+ | | | flow export (*) | | | +---------------+-----------+ | | | | +----------------------+---------------+ | v IPFIX export packet to Collector (*) indicates where flow selection can take place. Figure 1 In case the flow selection is performed during the metering process, then it consists in accounting only a subset of all the incoming packets which are collected at the observation point. However, unlike the selection process executed before the packet Peluso, et al. Expires May 21, 2008 [Page 7] Internet-Draft Flow selection Techniques November 2007 classification is performed, the flow selection applies to only the incoming packets which somehow satisfy certain conditions regarding flows state information available from the flow recording process. This kind of selection is considered as a packet sampling technique, in accordance with [PSAMP-TECH] where such technique is referred to as flow state dependent sampling. The state of the stored flow records is thus taken into account while performing packet selection, so that the process responsible for generating or updating flow records might be influenced by the selective collection of packets which feed it. Under this perspective, unlike the flow selection performed at the flow recording and exporting processes, this flow selection technique operates at a very early stage of the flow monitoring process, as it acts at packet level. The adoption of such technique allows to prevent that some observed/observable packets might enforce the flow recording process to account, for instance, not representative or not expected flow records. The flow selection that might be provided is executed during the flow recording and/or exporting processes, it is done at flow level, once packets are classified and assigned to the correspondent flows. More exactly, the flow selection process can be carried out by storing new flow records only in those cases whem enough resources are available at the monitoring device or by discarding already accounted records which, under certain circumstances and at a certain point in time, are not anymore significant. Finally, at the flow exporting time it might be required that not all of the stored flow records are actually exported to the collectors. We can distinguish the following selection techniques: 1. based on flow record content (i.e. all reported flow characteristics); 2. based on flow record arrival time; 3. based on external events like the exhaustion of local resources. 5.1. Flow selection on flow record content 5.2. Flow selection on flow record arrival time 5.3. Flow selection on external events Peluso, et al. Expires May 21, 2008 [Page 8] Internet-Draft Flow selection Techniques November 2007 6. Flow selection causes and relevant exportable information In this section we identify and describe in more detail some possible causes of flow selection, along with the information that can be beneficial to make available to applications about it. 6.1. Flow selection in the metering process The main reason for applying in the metering process a flow state dependent sampling is that the flow recording process may not have, at a certain point in time, enough positions to record all observable flows. Another reason may be that there may not be enough processing resources to create and manage a new flow record. To overcome with these limitations, a number of possible policies can be applied, the simplest one being not to consider for measurement the new packets that do not belong to already existing flow records (i.e. that would require the creation of a new one). More refined policies are however possible, mainly aimed at the so called elephant flow detection, i.e. to give priority in the flow recording process to flows carrying more traffic. For instance, [EsVa01] proposes criteria to define a packet eligible to create a new flow record (sample and hold, multistage filters). Independently of the specific algorithms, we are concerned here about defining what information it makes sense to keep about the flow state dependent packet sampling and make available to applications (by exporting it out of an IPFIX device). It is certainly possible to keep a cumulative counter of the total number of packets and bytes that were not considered for measurement because of flow state dependent sampling. Also, it is possible to keep a timestamp for the first and last of these non measured packets. This means, in practice, to aggregate all these packets in a macro flow, and keep track of its volume and duration. Imagining keeping more detailed information about packets not measured because of flow state dependent sampling would contradict the fact that the sampling is done because of lack of memory and/or processing resources. 6.2. Flow selection in the flow recording process This block is optional in the IPFIX framework architecture. However, we address here the case where it is present. We already described in the previous section that because of lack of memory positions in the flow recording process some incoming packets may be discarded if they lead to the opening of a new flow record. However, under certain circumstances, it may be advantageous to discard an existing flow record in the flow recording process to make room for the new record opened by an arriving packet. For example, an algorithm for taking the decision whether to discard the new arriving packet or an existing flow record is described in [Moli03]. In this section we Peluso, et al. Expires May 21, 2008 [Page 9] Internet-Draft Flow selection Techniques November 2007 are not concerned about the algorithm details but about what information to store about this record removal. For the same reasons expressed before, we argue that it does not make sense to store separate information for each discarded flow record, as it would contradict the motivation itself for which the discarding is done (i.e. lack of memory resources). The information that is certainly possible to keep with a limited effort is a cumulative counter of the total number of not yet exported packets and bytes belonging to flow records that were eliminated from the flow recording process. Ideally, we would also like to keep a timestamp for the first (T_fd) and last (T_ld) not yet exported packets belonging to all these discarded flow records. This would mean, in practice, to aggregate all these packets in a macro flow, and keep track of its volume and duration. To do so precisely, we would need to keep in each flow record a timestamp for the first and last non-exported packets, and whenever a record is discarded look at these timestamps to see if they are smaller or larger (respectively) of T_fd and T_ld and if yes update them. Another information that can be easily kept is the number of these discarding events, along with a timestamp of the first and last of them. This information should not be used by applications to re- normalize their received per flow statistics (because a flow may be discarded and re-created multiple times) but rather to keep under control the good functioning of the implemented policy. Note that we consider a discarding event only when the discarded flow record contains some not exported traffic. Otherwise, the removal of a record whose traffic was fully exported (after a timeout or after the arrival of specific packets, e.g. TCP FIN or RST) is part of the normal functioning of an IPFIX flow metering system. Note also that we consider only the case when an elimination of a flow record from the flow recording process leads to the complete loss of all the information contained in the flow record. If on the contrary another policy is implemented, like immediate exporting of the flow record before elimination, or freezing of the flow record and moving it in an area of memory different from which is considered the flow recording process for later exporting, this is not considered an elimination and therefore is out of the scope of this document. In parallel to the information about the number of discarded flow records and associated packets and bytes, it is useful to keep cumulative information about the number of flow records containing not yet exported traffic that exist in the flow recording process, along with the cumulative number of not exported packets and bytes contained in them. This information is useful also for exporting process related reasons, as clarified in the following paragraph. Peluso, et al. Expires May 21, 2008 [Page 10] Internet-Draft Flow selection Techniques November 2007 6.3. Flow selection in the exporting process The exporting process may implement policies for not exporting the whole set of flow records of the flow recording process. In case of absence of the flow recording process, when the metering process directly feeds the exporting process (i.e. directly put the exported packets in the IPFIX format), the following reasoning does not apply. The motivations for not exporting some flow records (containing non exported traffic) can be two: there are explicit configured policies or the exporting process faces resource limitation. An example of explicit policy can be not to export the flows whose accounted traffic is below a certain threshold, or a more complex mechanism such as the one described in [DuLT01a] or [DuLT01b]. An example of resource limitation is that the exporting process has an assigned, limited time slot to operate or a limited predefined number of export packets that it can send. There can also be hybrid cases where there are resource limitations and policies are applied in order to optimize the exported information (e.g. given that we want to export only N flow records, select a subset so that the overall number of reported packets and bytes belonging to the subset is maximized). Coming to the issue of which information it makes sense to keep about this flow selection, there are two cases to consider. If a flow is not exported and because of this decision is deleted from the flow recording process, we are in the same case described before (where the deletion was triggered by the need to make room for another record). The information to keep is then naturally the same as described before (cumulative packets and bytes for all the flows not exported, timestamps of the first and last packets belonging to non exported flow records, counter of dropping events and timestamp of first and last dropping event). Only the reason for this removal is different. If on the contrary a record eligible for exporting is not exported but it remains in the flow recording process it has always a chance to be exported in the future. For an application, however, it would be beneficial to know what it is not currently being exported because of exporting process policies/resource limitations, in terms of flow records, packets and bytes. This, not to re-normalize its estimates (it would be dangerous and error prone because the exporting of these records may be simply delayed), but rather to keep under control what is happening: for example, understand if there are pathologic situations where a large number of flow records and/or associated traffic are never exported, or if the number of flow records in the flow recording process is growing, etc. When it comes to understanding if this information can be easily available, however, we recognize that there is the problem that in order to be aware that it has not exported a flow record, an exporting process should at least have browsed through it. In other words, we would have to assume that there is always a full scanning of the flow recording process associated to the exporting process selection Peluso, et al. Expires May 21, 2008 [Page 11] Internet-Draft Flow selection Techniques November 2007 decision. However, there may be more efficient implementations where this does not happen. Therefore, even if we provide support in the information model for this information, defining it as mandatory in the protocol definition would put a constraint on the exporting process implementation, which is undesirable. 6.4. Information model for flow selection information exporting We formally define the elements to contain the information described in the previous section. Some elements have an associated couple of timestamps, which we reference for brevity (when it is not ambiguous) as Tfirst and Tlast (instead of element_nameTfirst, element_nameTlast). Note that all the following information elements are aimed at describing macro flows (e.g. the total number of packets and bytes contained in all dropped or not created flow records). Some of these macro flows are additive only, in the sense that they only add contributions to them, but never subtract. E.g. the macro flow of the packets contained in flow records that are discarded from the flow reporting process receives a contribution when a flow record is discarded, and this contribution can never be subtracted. On the contrary, some of the macro flows can dynamically receive and loose contributions. E.g. the macro flows of packets not yet exported receives a contribution when a new packets arrives, and looses some contribution when there is an exporting event. Associating a timestamp for the oldest and most recent contributions to additive only flow is easy, while for the others is not (would require to maintain full state) and that is why we did not define timestamps for these information elements. 6.4.1. Meter process related 6.4.1.1. FsMeter_UnmeasPacketCount Contains the count of packets that were not measured because of flow state dependent sampling, in terms of: TsFirst: timestamp of the first packet not measured because of flow state dependent sampling (Type: dateTime) TsLast: timestamp of the last packet not measured because of flow state dependent sampling (Type: dataTime) 6.4.1.2. FsMeter_UnmeasBytesCount Contains the count of bytes that were not measured because of flow state dependent sampling Type: unsignedInt Peluso, et al. Expires May 21, 2008 [Page 12] Internet-Draft Flow selection Techniques November 2007 6.4.2. Flow recording process related 6.4.2.1. FsFrec_PacketInDroppedRecsCount Contains the count of non exported packets that were contained in flow records eliminated from the flow recording process because of resource limitations/policies in the flow recording process. It is defined in terms of: TsFirst: timestamp of the first non-exported packet belonging to a eliminated flow record (Type: dateTime) TsLast: timestamp of the last non-exported packet belonging to a eliminated flow record (Type: dateTime) 6.4.2.2. FsFrec_ByteInDroppedRecsCount Contains the count of non exported bytes that were contained in flow records eliminated from the flow recording process because of resource limitations/policies in the flow recording process. Type: unsignedInt 6.4.2.3. FsFrec_FrecDroppedCount Contains the count of flow records containing non exported packets eliminated from the flow recording process because of resources limitations/policies in the flow recording process. It is defined in terms of: TsFirst: timestamp of the first flow record elimination event from the flow recording process (Type: dateTime) TsLast: timestamp of the last flow record elimination event from the flow recording process (Type: dateTime) 6.4.2.4. FsFrec_UnexportedFrecCount Contains the count of the flow records currently existing in the flow recording process containing at least one non exported packet Type: unsignedInt 6.4.2.5. FsFrec_UnexportedPacketInFrecCount Contains the count of non exported packets contained in flow records of the flow recording process. Peluso, et al. Expires May 21, 2008 [Page 13] Internet-Draft Flow selection Techniques November 2007 Type: unsignedInt 6.4.2.6. FsFrec_UnexportedBytesInFrecCount Contains the count of non exported bytes contained in flow records of the flow recording process. Type: unsignedInt 6.4.3. Flow exporting process related 6.4.3.1. FsExp_PacketInDroppedRecsCount Contains the count of non exported packets that were contained in flow records eliminated from the flow recording process because of resource limitations/policies in the exporting process. It is defined in terms of: TsFirst: timestamp of the first non exported packet belonging to a eliminated flow record (Type: dateTime) TsLast: timestamp of the last non exported packet belonging to a eliminated flow record (Type: dateTime) 6.4.3.2. FsExp_ByteInDroppedRecsCount Contains the count of non exported bytes that were contained in flow records eliminated from the flow recording process because of resource limitations/policies in the exporting process. Type: unsignedInt 6.4.3.3. FsExp_FrecDroppedCount Contains the count of flow records containing non exported packets eliminated from the flow recording process because of resource limitations/policies in the exporting process. It is defined in terms of: TsFirst: timestamp of the first flow record elimination event from the flow recording process (Type: dateTime) TsLast: timestamp of the last flow record elimination event from the flow recording process (Type: dateTime) Peluso, et al. Expires May 21, 2008 [Page 14] Internet-Draft Flow selection Techniques November 2007 6.4.3.4. FsExp_UnexportedCount Contains the count of the flow records currently existing in the flow recording process containing non-exported traffic and not being exported because of exporting process resource lmitations/policies. Type: unsignedInt 6.4.3.5. FsExp_UnexportedPacketCount Contains the count of non exported packets contained in flow records of the flow recording process not being exported because of exporting process resource limitations/policies. Type: unsignedInt 6.4.3.6. FsFrec_UnexportedByteInFrecCount Contains the count of non exported bytes contained in flow records of the flow recording process not being exported because of exporting process resource limitations/policies. Type: unsignedInt 6.5. Requirements put on implementations To support the described information model an implementation must keep, in the flow records, counts for non-exported packets and bytes. Sometimes these are referred as delta counts. An implementation may also keep absolute counts for scopes not specified in this information model (it appears that both delta and absolute counters can be exported in the IPFIX information model, see [IPFIX-INFO]). In addition, to fully support this information model, it would be required to keep in a flow record a timestamp for the first and last non-exported packets. An implementation may need to keep timestamps for the first and last exported packets as well for scopes not specified in this information model, or to join the two timers for the last exported and first exported packets (which is of course an approximation) or to approximate them with the time of the exporting event. 7. Information model for flow selection configuration This section aims at describing the representative parameters of the above presented flow selection techniques. To this regard, this section provides the basis for an information model to adopt in order to configure the flow selection process at an IPFIX device. Peluso, et al. Expires May 21, 2008 [Page 15] Internet-Draft Flow selection Techniques November 2007 8. IANA Considerations This document makes no request of IANA. 9. References 9.1. Normative References [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. 9.2. Informative References [DuLT01a] Duffield, N., Lund, C., and M. Thorup, "Charging from Sampled Network Usage", ACM Internet Measurement Workshop IMW 2001, San Francisco, USA, November 2001. [DuLT01b] Duffield, N., Lund, C., and M. Thorup, "Properties and Prediction of Flow Statistics from Sampled Packet Streams", ACM SIGCOMM Internet Measurement Workshop 2002, November 2002. [DuLT01c] Duffield, N., Lund, C., and M. Thorup, "Learn More, sample less: control of volume and variance in network measurement", IEEE Transactions on Information Theory, May 2005. [DuLT01d] Duffield, N., Lund, C., and M. Thorup, "Flow Sampling under Hard Resource Constraints", ACM IFIP Conference on Measurement and Modeling of Computer Systems SIGMETRICS, June 2004. [EsVa01] Estan, C. and G,. Varghese, "New Directions in Traffic Measurement and Accounting: Focusing on the Elephants, Ignoring the Mice", ACM SIGCOMM Internet Measurement Workshop 2001, San Francisco (CA), November 2001. [FeGL98] Feldmann, A., Rexford, J., and R. Caceres, "Efficient Policies for Carrying Web Traffic over Flow-Switched Networks", IEEE/ACM Transaction on Networking, December 1998. [IPFIX-ARCH] Sadasivan, G., Bownlee, N., Claise, B., and J. Quittek, "Architecture for IP Flow Information Export", Internet Draft draft-ietf-ipfix-architecture-12.txt, work in progress, September 2006. Peluso, et al. Expires May 21, 2008 [Page 16] Internet-Draft Flow selection Techniques November 2007 [IPFIX-INFO] Quittek, J., Bryant, S., Claise, B., Aitken, P., and J. Meyer, "Information Model for IP Flow Information Export", Internet Draft draft-ietf-ipfix-info-15.txt, work in progress, February 2007. [KuXW04] Kumar, K., Xu, J., Wang, J., Spatschek, O., and L. Li, "Space-code bloom filter for efficient per-flow traffic measurement", INFOCOM 2004 Twenty-third AnnualJoint Conference of the IEEE Computer and Communications Societies, March 2004. [Moli03] Molina, M., "A scalable and efficient methodology for flow monitoring in the Internet", International Teletraffic Congress (ITC-18), Berlin, September 2003. [PSAMP-TECH] Zseby, T., Molina, M., Raspall, F., Duffield, N., and S. Niccolini, "Sampling and Filtering techniques for IP Packet Selection", Internet Draft draft-ietf-psamp-sample-tech-10.txt, work in progress, June 2007. Authors' Addresses Lorenzo Peluso University of Napoli/Fraunhofer Institute FOKUS Via Claudio 21 Napoli 80125 Italy Phone: +39 081 7683821 Email: lorenzo.peluso@unina.it, lorenzo.peluso@fokus.fraunhofer.de Tanja Zseby Fraunhofer Institute FOKUS Kaiserin-Augusta-Allee 31 Berlin 10589 Germany Phone: +49 30 3463 7153 Email: zseby@fokus.fraunhofer.de Peluso, et al. Expires May 21, 2008 [Page 17] Internet-Draft Flow selection Techniques November 2007 Salvatore D'Antonio CINI Consortium/ITeM Laboratory Monte S.Angelo, Via Cinthia Napoli 80126 Italy Phone: +39 081 679944 Email: saldanto@unina.it Maurizio Molina DANTE Hills Road 126-130 Cambridge CB2 1PQ United Kingdom Phone: +44 1223 371300 Email: maurizio.molina@dante.org.uk Peluso, et al. Expires May 21, 2008 [Page 18] Internet-Draft Flow selection Techniques November 2007 Full Copyright Statement Copyright (C) The IETF Trust (2007). This document is subject to the rights, licenses and restrictions contained in BCP 78, and except as set forth therein, the authors retain all their rights. This document and the information contained herein are provided on an "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Intellectual Property The IETF takes no position regarding the validity or scope of any Intellectual Property Rights or other rights that might be claimed to pertain to the implementation or use of the technology described in this document or the extent to which any license under such rights might or might not be available; nor does it represent that it has made any independent effort to identify any such rights. Information on the procedures with respect to rights in RFC documents can be found in BCP 78 and BCP 79. Copies of IPR disclosures made to the IETF Secretariat and any assurances of licenses to be made available, or the result of an attempt made to obtain a general license or permission for the use of such proprietary rights by implementers or users of this specification can be obtained from the IETF on-line IPR repository at http://www.ietf.org/ipr. The IETF invites any interested party to bring to its attention any copyrights, patents or patent applications, or other proprietary rights that may cover technology that may be required to implement this standard. Please address the information to the IETF at ietf-ipr@ietf.org. Acknowledgment Funding for the RFC Editor function is provided by the IETF Administrative Support Activity (IASA). Peluso, et al. Expires May 21, 2008 [Page 19]