IPFIX Working Group A. Kobayashi, Ed. Internet-Draft NTT PF Lab. Intended status: Informational B. Claise, Ed. Expires: November 1, 2009 Cisco Systems, Inc. April 30, 2009 IPFIX Mediation: Problem Statement draft-ietf-ipfix-mediators-problem-statement-03 Status of this Memo This Internet-Draft is submitted to IETF in full conformance with the provisions of BCP 78 and BCP 79. This document may contain material from IETF Documents or IETF Contributions published or made publicly available before November 10, 2008. The person(s) controlling the copyright in some of this material may not have granted the IETF Trust the right to allow modifications of such material outside the IETF Standards Process. Without obtaining an adequate license from the person(s) controlling the copyright in such materials, this document may not be modified outside the IETF Standards Process, and derivative works of it may not be created outside the IETF Standards Process, except to format it for publication as an RFC or to translate it into languages other than English. 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 November 1, 2009. Kobayashi, et al. Expires November 1, 2009 [Page 1] Internet-Draft Mediation Problem Statement April 2009 Copyright Notice Copyright (c) 2009 IETF Trust and the persons identified as the document authors. All rights reserved. This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents in effect on the date of publication of this document (http://trustee.ietf.org/license-info). Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Kobayashi, et al. Expires November 1, 2009 [Page 2] Internet-Draft Mediation Problem Statement April 2009 Abstract Flow-based measurement is a popular method for various network monitoring usages. The sharing of flow-based information for monitoring applications having different requirements raises some open issues in terms of measurement system scalability, flow-based measurement flexibility, and export reliability that IPFIX Mediation may help resolve. IPFIX Mediation covers two classes of mediation: context mediation for traffic data and transport mediation for transport protocols. This document describes the IPFIX Mediation applicability examples, along with some problems that network administrators have been facing. Kobayashi, et al. Expires November 1, 2009 [Page 3] Internet-Draft Mediation Problem Statement April 2009 Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 5 2. Terminology and Definition . . . . . . . . . . . . . . . . . . 6 3. IPFIX/PSAMP Documents Overview . . . . . . . . . . . . . . . . 9 3.1. IPFIX Documents Overview . . . . . . . . . . . . . . . . . 9 3.2. PSAMP Documents Overview . . . . . . . . . . . . . . . . . 9 4. Problem Statement . . . . . . . . . . . . . . . . . . . . . . 10 4.1. Coping with IP Traffic Growth . . . . . . . . . . . . . . 10 4.2. Coping with Multipurpose Traffic Measurement . . . . . . . 11 4.3. Coping with Heterogeneous Environments . . . . . . . . . . 11 4.4. Summary . . . . . . . . . . . . . . . . . . . . . . . . . 11 5. Mediation Applicability Examples . . . . . . . . . . . . . . . 12 5.1. Adjusting Flow Granularity . . . . . . . . . . . . . . . . 12 5.2. Hierarchical Collecting Infrastructure . . . . . . . . . . 12 5.3. Correlation for Data Records . . . . . . . . . . . . . . . 13 5.4. Time Composition . . . . . . . . . . . . . . . . . . . . . 13 5.5. Spatial Composition . . . . . . . . . . . . . . . . . . . 14 5.6. Data Record Anonymization . . . . . . . . . . . . . . . . 15 5.7. Data Retention . . . . . . . . . . . . . . . . . . . . . . 15 5.8. IPFIX Export from a Branch Office . . . . . . . . . . . . 16 5.9. Distributing Data Records . . . . . . . . . . . . . . . . 17 5.10. Flow-based Sampling and Selection . . . . . . . . . . . . 18 5.11. Interoperability between Legacy Protocols and IPFIX . . . 19 6. IPFIX Mediators Implementation Specific Problems . . . . . . . 20 6.1. Loss of Original Exporter Information . . . . . . . . . . 20 6.2. Loss of Base Time Information . . . . . . . . . . . . . . 20 6.3. Transport Sessions Management . . . . . . . . . . . . . . 21 6.4. Loss of Option Template Information . . . . . . . . . . . 21 6.5. Template ID Management . . . . . . . . . . . . . . . . . . 21 6.6. Consideration for Network Topology . . . . . . . . . . . . 22 6.7. Exporting the Function Item . . . . . . . . . . . . . . . 22 6.8. Consideration for Aggregation . . . . . . . . . . . . . . 23 7. Summary and Conclusion . . . . . . . . . . . . . . . . . . . . 24 8. Security Considerations . . . . . . . . . . . . . . . . . . . 26 9. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 27 10. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 28 11. References . . . . . . . . . . . . . . . . . . . . . . . . . . 29 11.1. Normative References . . . . . . . . . . . . . . . . . . . 29 11.2. Informative References . . . . . . . . . . . . . . . . . . 29 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 31 Kobayashi, et al. Expires November 1, 2009 [Page 4] Internet-Draft Mediation Problem Statement April 2009 1. Introduction While the IPFIX requirements defined in [RFC3917] mention an intermediate function, such as an IPFIX Proxy or an IPFIX Concentrator, there are no documents defining the function called IPFIX Mediation. IPFIX Mediation is a generic function that covers the manipulation of IPFIX Flow Records, PSAMP Packet Reports, entire IPFIX Messages, or their transmission. This document describes general problems, applicability examples, and defines the terminology (IPFIX Proxy, Concentrator, etc.) for referring to different use cases for IPFIX Mediation. Furthermore, some specific problems related to the IPFIX protocol [RFC5101] when applying IPFIX Mediation are addressed. This document is structured as follows: section 2 describes the terminology used in this document, section 3 gives an IPFIX/PSAMP document overview, section 4 introduces general problems related to flow-based measurement, section 5 describes some applicability examples where IPFIX Mediations would be beneficial, and, finally, section 6 describes some problems an IPFIX Mediation implementation might face. Kobayashi, et al. Expires November 1, 2009 [Page 5] Internet-Draft Mediation Problem Statement April 2009 2. Terminology and Definition The terms in this section are in line with those in the IPFIX Protocol specifications [RFC5101] and the PSAMP specification document [RFC5476]. The terms Observation Point, Observation Domain, Flow Key, Flow Record, Data Record, Exporting Process, Exporter, IPFIX Device, Collecting Process, Collector, IPFIX Message, Metering Process, Transport Session, Information Element, and Template Withdrawal Message, are defined in the IPFIX protocol specifications [RFC5101]. The terms Packet Report, Sampling, Filtering, PSAMP Device, and Configured Selection Fraction are defined in the PSAMP specification document [RFC5476]. Furthermore, new terminology to be used in the context of IPFIX Mediation is defined in this section. All these terms have an initial capital letter in this document. While IPFIX Mediation can process both Flow Records and Packet Reports, this document prefers the more generic "Data Record" term as this is a more generic term, unless the reference to the IPFIX Flow Record or PSAMP Packet Report is required. IPFIX Mediation IPFIX Mediation is a generic function that covers the manipulation of IPFIX Flow Records, PSAMP Packet Reports, or entire IPFIX Messages, or their transmission. The IPFIX Mediation offers one or multiple of the following capabilities: * content mediation that changes Flow Records/Packet Reports information + aggregating Flow Records/Packet Reports based on a new set of Flow Key fields + correlating a set of Flow Records/Packet Reports + filtering and selecting Flow Records/Packet Reports + modifying Flow Records/Packet Reports, including: - changing the value of specified Information Elements - adding new Information Elements by deriving further Flow or packet properties from existing fields (for example: calculating new metrics or new counters) - deleting specified Information Elements Kobayashi, et al. Expires November 1, 2009 [Page 6] Internet-Draft Mediation Problem Statement April 2009 * transport mediation + changing the transport protocol that carries IPFIX Messages + rerouting entire IPFIX Messages to an appropriate Collecting Process + replicating Flow Records/Packet Reports (or the entire IPFIX Messages) IPFIX Mediator An IPFIX Mediator is an IPFIX Device that implements one or more IPFIX Mediation capabilities. Examples are devices such as routers, switches, network management systems (NMS), etc. Original Exporter An Original Exporter is an IPFIX Device that hosts the Observation Points where the metered IP packets are observed. IPFIX Proxy An IPFIX Proxy is a type of IPFIX Mediation that relays incoming Transport Sessions to one or multiple Collectors. The protocols used at the input and the output can be different, which implies that IPFIX Messages, Data Records, and Template Records need to be encoded, e.g., for converting from a legacy protocol to IPFIX. An IPFIX Proxy is not implemented on the Original Exporter, but as a separate Mediator. IPFIX Concentrator An IPFIX Concentrator is a type of IPFIX Mediation that receives Flow Records/Packet Reports, correlates them, aggregates them, or modifies them, then exports the new Data Records. IPFIX Distributor An IPFIX Distributor is a type of IPFIX Mediation that distributes Data Records to one or multiple IPFIX Collectors. The decision as to which IPFIX Collector a Data Record is exported can be determined by filtering certain field values or other properties derived from the Data Record. Kobayashi, et al. Expires November 1, 2009 [Page 7] Internet-Draft Mediation Problem Statement April 2009 IPFIX Masquerading Proxy An IPFIX Masquerading Proxy is a type of IPFIX Mediation that screens out parts of input Flow Records/Packet Reports according to configured policies. It can thus, for example, hide the network topology information or customers' IP addresses. Kobayashi, et al. Expires November 1, 2009 [Page 8] Internet-Draft Mediation Problem Statement April 2009 3. IPFIX/PSAMP Documents Overview 3.1. IPFIX Documents Overview The IPFIX protocol [RFC5101] provides network administrators with access to IP flow information. The architecture for the export of measured IP flow information out of an IPFIX Exporting Process to a Collecting Process is defined in [RFC5470], per the requirements defined in [RFC3917]. The IPFIX protocol [RFC5101] specifies how IPFIX Data Records and Templates are carried via a number of transport protocols from IPFIX Exporting Processes to IPFIX Collecting Processes. IPFIX has a formal description of IPFIX Information Elements, their names, types, and additional semantic information, as specified in [RFC5102]. [I-D.ietf-ipfix-mib] specifies the IPFIX Management Information Base. Finally, [RFC5472] describes what types of applications can use the IPFIX protocol and how they can use the information provided. It furthermore shows how the IPFIX framework relates to other architectures and frameworks. The storage of IPFIX Messages in a file is specified in [I-D.ietf-ipfix-file]. 3.2. PSAMP Documents Overview The framework for packet selection and reporting [RFC5474] enables network elements to select subsets of packets by statistical and other methods and to export a stream of reports on the selected packets to a Collector. The set of packet selection techniques (Sampling, Filtering, and Hashing) standardized by PSAMP are described in [RFC5475]. The PSAMP protocol [RFC5476] specifies the export of packet information from a PSAMP Exporting Process to a Collector. Like IPFIX, PSAMP has a formal description of its Information Elements, their names, types and additional semantic information. The PSAMP information model is defined in [RFC5477]. [I-D.ietf-psamp-mib] describes the PSAMP Management Information Base. Kobayashi, et al. Expires November 1, 2009 [Page 9] Internet-Draft Mediation Problem Statement April 2009 4. Problem Statement Network administrators generally face the problems of measurement system scalability, flow-based measurement flexibility, and export reliability, even if some techniques, such as Sampling, Filtering, Data Records aggregation and export replication, have already been developed. The problems consist of optimizing the resources of the measurement system while pursuing appropriate conditions: data accuracy, flow granularity, and export reliability. These conditions depend on two factors. o measurement systems capacity: This consists of the bandwidth of the management network, the storage capacity, and the performances of the collecting devices and exporting devices. o applications requirements: Different applications, such as traffic engineering, detecting anomaly traffic, and accounting, etc., impose different Flow Record granularity, and data accuracy. The recent continued IP traffic growth has been overwhelming the measurement system capacity, and multi-purposing applications, along with the heterogeneous environments, have further been contributing to a complex situation. The following sub-sections explain different problems in more details. 4.1. Coping with IP Traffic Growth Enterprise or service provider networks already have multiple 10 Gb/s links, their total traffic exceeding 100 Gb/s. In the near future, broadband users' traffic will increase by approximately 40% every year according to [TRAFGRW]. When operators monitor traffic of 500 Gb/s with a Sampling rate of 1/1000, the amount of exported Flow Records from Exporters could exceed 50 kFlows/s. This value is beyond the ability of a single Collector. To deal with this problem, current data reduction techniques, such as Sampling, Filtering, Data Records aggregation have been generally implemented on Exporters. Note that Sampling technique leads to potential loss of small Flows. With both Sampling and aggregation techniques, administrators might no longer be able to investigate very granular traffic change and anomaly detection, both of which can currently be detected. With Filtering, only a subset of the Flow Records are exported. Considering the potential drawbacks of Sampling, Filtering, and Data Records aggregation, there is a need for a large-scale collecting Kobayashi, et al. Expires November 1, 2009 [Page 10] Internet-Draft Mediation Problem Statement April 2009 infrastructure that does not rely on data reduction techniques. 4.2. Coping with Multipurpose Traffic Measurement A set of conditions (flow granularity and data accuracy) may meet the requirements of some applications, such as traffic engineering, but may not meet the requirements of other applications, such as accounting, QoS performance, or even security. Therefore, with a single set of conditions, multipurpose traffic measurements cannot be accomplished. To cope with the issue, an Exporter needs to export traffic data with strictest condition (fine flow granularity and high data accuracy) required by the set of applications. However, this implies an increased load on both the Exporter and Collector. 4.3. Coping with Heterogeneous Environments Network administrators use IPFIX Devices and PSAMP Devices from various vendors, various software versions, various device types (router, switch, or probe) in a single network domain. Even legacy flow export protocols are still deployed in current network. This heterogeneous environment leads to differences in Metering Process capability, Exporting Process capacity (export rate, cache memory, etc.), and data format. For example, probes and switches cannot retrieve packet property information from a route table. To deal with this problem, the collecting infrastructure needs to absorb the differences. However, equipping all collecting devices with this absorption function is difficult. 4.4. Summary In optimizing the resources of a measurement system, it is important to use traffic data reduction techniques at the possible initial phase, e.g., at the Exporter. However, this implementation is made difficult by heterogeneous environment of exporting devices. This implies that a new Mediation functional block is required in typical Exporter-Collector architectures. Based on some applicability examples, the next section shows the limitation of the typical Exporter-Collector architecture model and the IPFIX Mediation benefits. Kobayashi, et al. Expires November 1, 2009 [Page 11] Internet-Draft Mediation Problem Statement April 2009 5. Mediation Applicability Examples 5.1. Adjusting Flow Granularity The simplest types of Flows are those comprised of packets all having a fixed IP-quintuple of protocol, source and destination IP addresses, and source and destination port numbers. However, a shorter set of Flow Keys, such as a triple, a double, or a single Flow Key, (for example network prefix, peering AS number, or BGP Next-Hop fields), creates more aggregated Flow Records. This is especially useful for measuring traffic exchange in an entire network domain and for easily adjusting the performance of Exporters and Collectors. Implementation analysis: Implementations for this case depend on where Flow granularity is adjusted. More suitable implementations use configurable Metering Processes in Original Exporters. The cache in the Metering Process can specify its own set of Flow Keys and extra fields. The Original Exporter thus creates directly aggregated Flow Records. In the case where the Original Exporter contains a Metering Process that creates fixed tuple Flow Records (no possibility to change the Flow Keys), an IPFIX Concentrator can adjust the Flow Keys by aggregation Flow Records. Even if the case where the Original Exporter contains a Metering Process for which the Flow Keys can be configured, an IPFIX Concentrator can further aggregate the Flow Records. 5.2. Hierarchical Collecting Infrastructure The increase of IPFIX Exporters, the increase of the traffic over large-scale networks, and the variety of treatments expected to be performed over the Data Records is more and more difficult to handle within a single Collector. Hence to increase the collecting (e.g. the bandwidth capacity) and processing capacity must be distributed over several IPFIX entities. As a possible approach, a hierarchical structure is useful for increasing the measurement systems capacity, both in export bandwidth capacity and in collecting capacity. Implementation analysis: To cope with the increase of IPFIX Exporters and traffic, one possible implementation uses IPFIX Concentrators to build a hierarchical collection system. To cope with the variety of treatments, one possible implementation uses IPFIX Distributors to Kobayashi, et al. Expires November 1, 2009 [Page 12] Internet-Draft Mediation Problem Statement April 2009 build a distributed collection system. More specific cases are described in section 5.9. 5.3. Correlation for Data Records The correlation amongst Data Records or between Data record and meta data provides new metrics or information, including the following. o One-to-one correlation between Data Records * One way delay and packet arrival interval time etc. One way delay from the correlation of Packet Reports from different Exporters along a specific path. * Treatment from the correlation of Data Records with the same Flow Key(s) observed at incoming/outgoing interfaces. Examples are the rate-limiting ratio, the compression ratio, the optimization ratio, etc. o Correlation amongst Data Records Average/maximum/minimum values from correlating multiple Data Records. Examples are the average/maximum/minimum packets of Flow, the average/maximum/minimum one way delay, the average/ maximum/minimum packet loss, etc. o Correlation between Data Record and other meta data Examples are some BGP attributes associated with Data Record by looking up routing table. Implementation analysis: One possible implementation for the case uses an IPFIX Concentrator located between the Metering Processes and Exporting Processes on the Original Exporter, or alternatively a separate IPFIX Concentrator located between the Original Exporters and IPFIX Collectors. 5.4. Time Composition Time composition is defined as the aggregation of consecutive Data Records with identical Flow Keys. It leads to the same output as setting a longer active interval timer on Original Exporters with one advantage: the creation of new metrics such as average, maximum and minimum values from Flow Records with a shorter time interval enables administrators to keep track of changes that might have happened during the time interval. Kobayashi, et al. Expires November 1, 2009 [Page 13] Internet-Draft Mediation Problem Statement April 2009 Implementation analysis: One possible implementation for this case uses an IPFIX Concentrator located between the Metering Processes and Exporting Processes on the Original Exporter, or alternatively as a separate IPFIX Concentrator located between the Original Exporters and IPFIX Collectors. 5.5. Spatial Composition Spatial composition is defined as the aggregation of Data Records in a set of Observation Points within an Observation Domain, across multiple Observation Domains from a single Exporter, or even across multiple Exporters. The spatial composition is divided into four types. o Case 1: Spatial Composition within one Observation Domain For example, in the case where a link aggregation exists, Data Records observed at physical interfaces belonging to the same trunk can be merged. o Case 2: Spatial Composition across Observation Domains, but within a single Exporter For example, in the case where a link aggregation exists, Data Records observed at physical interfaces belonging to a same trunk grouping beyond the line interface module can be merged. o Case 3: Spatial Composition across Exporters Data Records observed within an administrative domain, such as the west area and east area of an ISP network, can be merged. o Case 4: Spatial Composition across administrative domains Data Records observed across administrative domains, such as across different customer networks or different ISP networks, can be merged. Implementation analysis: One possible implementation for the case 1 and 2 uses an IPFIX Concentrator located between the Metering Processes and Exporting Processes on the Original Exporter. A separate IPFIX Concentrator located between the Original Exporters and IPFIX Collector is a valid solution for the case 1, 2, 3, and 4. Kobayashi, et al. Expires November 1, 2009 [Page 14] Internet-Draft Mediation Problem Statement April 2009 5.6. Data Record Anonymization IPFIX exports across administrative domains can be used to measure traffic for wide-area traffic engineering or to analyze Internet traffic trends, as described in the Spatial Composition across administrative domains in the previous subsection. In such case, administrators need to adhere to privacy protection policies and prevent access to confidential traffic measurements by other people. Typically, anonymization techniques enables the provision of traffic data to other people without violating these policies. Generally, anonymization modifies a data set to protect the identity of the people or entities described by the data set from being disclosed. It also attempts to preserve sets of network traffic properties useful for a given analysis while ensuring the data cannot be traced back to the specific networks, hosts, or users generating the traffic. For example, IP address anonymization is particularly important for avoiding the identification of the users, hosts, and routers. Implementation analysis: One possible implementation for this case uses an anonymization function at the Original Exporter. However, this increases the load on the Original Exporter. A more flexible implementation uses a separate IPFIX Masquerading Proxy between the Original Exporter and Collector. 5.7. Data Retention Data retention refers to the storage of traffic data by service providers and commercial organizations. Network administrators should retain both IP and voice traffic data, in wired and wireless networks, generated by end users while using a service provider's services. The traffic data is required for the purpose of the investigation, detection and prosecution of serious crime, if necessary. Data retention services examples are the following: o Fixed telephony (includes fixed voice calls, voicemail, and conference and data calls) o Mobile telephony (includes mobile voice calls, voicemail, conference and data calls, SMS, and MMS) o Internet telephony (includes every multimedia session associated with IP multimedia services) Kobayashi, et al. Expires November 1, 2009 [Page 15] Internet-Draft Mediation Problem Statement April 2009 o Internet e-mail o Internet access Data retention for Internet access services in particular requires a measurement system with reliable export and huge storage as the data must be available for a long period of time, typically a couple of years. Implementation analysis: Regarding export reliability requirement, the most suitable implementation uses the SCTP transport protocol between the Original Exporter and Collector. If a unreliable transport protocol such as UDP is used, a legacy exporting device exports Data Records to a nearby IPFIX Proxy through UDP, and then an IPFIX Proxy could reliably export them to the top IPFIX Collector through SCTP. If a unreliable transport protocol such as UDP is used and if there is no IPFIX Proxy, the legacy exporting device must duplicate the exports to several Collectors. Regarding huge storage requirement, one possible implementation uses a decentralized set of Collectors. If administrators need to retrieve specific Data Records, these Collectors would need to be equipped with IPFIX Mediations. 5.8. IPFIX Export from a Branch Office Generally, in large enterprise networks, Data Records from branch offices are gathered in a central office. However, in the long distance branch office case, the bandwidth for transport IPFIX is limited. Therefore, even if multiple Flow Records type should be of interest to the Collector (Flow Records in both directions, Flow Records before and after WAN optimization techniques, performance metrics associated with the Flow Records exported on regular interval), the export bandwidth limitation is an important factor to pay attention to. Implementation analysis: One possible implementation for the case uses an IPFIX Concentrator located in a branch office. The IPFIX Concentrator would aggregate and correlate Flow Records to cope with the export bandwidth limitation. Kobayashi, et al. Expires November 1, 2009 [Page 16] Internet-Draft Mediation Problem Statement April 2009 5.9. Distributing Data Records Recently, several networks have shifted towards integrated networks, such as the pure IP and MPLS, which includes IPv4, IPv6, and VPN traffic. Data Record types (IPv4, IPv6, MPLS, and VPN) need to be analyzed separately and from different perspectives for different organizations. A single Collector handling all Data Record types might become a bottleneck in the collecting infrastructure. Data Records distributed based on their respective types can be exported to the appropriate Collector, resulting in the load distribution amongst multiple Collectors. Implementation analysis: One possible implementation for this case uses the replications of the IPFIX Message in an Original Exporter for multiple IPFIX Collectors. Each Collector then extracts the Data Record required by its own applications. However, the replication increases the load of the Exporting Process and the waste of the bandwidth between the Exporter and Collector. A more sophisticated implementation uses an IPFIX Distributor located between the Metering Processes and Exporting Processes in an Original Exporter. The IPFIX Distributor determines respectively to which Collector a Data Record is exported by filtering certain field values. If a Original Exporter does not have IPFIX Distributor capability, it exports Data Records to a nearby separate IPFIX Distributor, and then the IPFIX Distributor could distribute them to the appropriate IPFIX Collectors. For example, in the case of distributing a specific customer's Data Records, an IPFIX Distributor needs to identify the customer networks. The Route Distinguisher (RD), ingress interface, peering AS number, or BGP Next-Hop, or simply the network prefix may be evaluated to distinguish different customer networks. In the following figure, the IPFIX Distributor reroutes Data Records on the basis of the RD value. This system enables each customer's traffic to be inspected independently. Kobayashi, et al. Expires November 1, 2009 [Page 17] Internet-Draft Mediation Problem Statement April 2009 .---------. |Traffic | .---->|Collector|<==>Customer#A | |#1 | | '---------' RD=100:1 .----------. .-----------. | |IPFIX | |IPFIX |----' .---------. |Exporter#1| |Distributor| RD=100:2 |Traffic | | |------->| |--------->|Collector|<==>Customer#B | | | | |#2 | | | | |----. '---------' '----------' '-----------' | RD=100:3 | .---------. | |Traffic | '---->|Collector|<==>Customer#C |#3 | '---------' Figure A: Distributing Data Records to Collectors using IPFIX Distributor 5.10. Flow-based Sampling and Selection Generally, the distribution of the number of packets per Flow seems to be heavy-tailed. Most types of Flow Records are likely to be small Flows consisting of a small number of packets. The measurement system is overwhelmed with a huge amount of these small Flows. If statistics information of small Flows is exported as merged data by applying a policy or threshold, the load on the Exporter is reduced. Furthermore, if the flow distribution is known, exporting only a subset of the Data Records might be sufficient. Implementation analysis: One possible implementation for this case uses an IPFIX Concentrator located between the Metering Processes and Exporting Processes on the Original Exporter, or alternatively as a separate IPFIX Concentrator located between the Original Exporters and IPFIX Collectors. A set of IPFIX Mediation functions, such as filtering, selecting and aggregation is used in the IPFIX Concentrator. Kobayashi, et al. Expires November 1, 2009 [Page 18] Internet-Draft Mediation Problem Statement April 2009 5.11. Interoperability between Legacy Protocols and IPFIX During the migration process from a legacy protocol such as NetFlow [RFC3954] to IPFIX, both NetFlow exporting devices and IPFIX Exporters are likely to coexist in the same network. Operators need to continue measuring the traffic data from legacy exporting devices, even after introducing IPFIX Collectors. Implementation analysis: One possible implementation for this case uses an IPFIX Proxy that converts a legacy protocol to IPFIX. Kobayashi, et al. Expires November 1, 2009 [Page 19] Internet-Draft Mediation Problem Statement April 2009 6. IPFIX Mediators Implementation Specific Problems 6.1. Loss of Original Exporter Information Both the Exporter IP address indicated by the source IP address of the IPFIX Transport Session and the Observation Domain ID included in the IPFIX Message header are likely to be lost by an IPFIX Mediator such as IPFIX Concentrator too. In some cases, a IPFIX Masquerading Proxy might drop the information. In other cases, the Collector must recognize the Original Exporter (and potentially the Observation Domain and Observation Point as well) whether Data Records go through an IPFIX Mediator or not. Note that, if the Mediator can not communicate the Original Exporter IP address, then the top level Collector will wrongly deduce that the IP address of the IPFIX Mediator is that of the Original Exporter. In the following figure, a Collector can identify two IP addresses: 10.1.1.3 (IPFIX Mediator) and 10.1.1.2 (Exporter#2), respectively. The Collector, however, needs to somehow recognize both Exporter#1 and Exporter#2, which are the Original Exporters. The IPFIX Mediator must have a specific way to the Original Exporter IP address to the IPFIX Collector. .----------. .--------. |IPFIX | |IPFIX | |Exporter#1|--------->|Mediator|---+ | | | | | '----------' '--------' | .---------. IP:10.1.1.1 IP:10.1.1.3 '----->|IPFIX | ODID:10 ODID:0 |Collector| +----->| | .----------. | '---------' |IPFIX | | |Exporter#2|-----------------------' | | '----------' IP:10.1.1.2 ODID:20 Figure B: Loss of Original Exporter Information. 6.2. Loss of Base Time Information The Export Time field included in the IPFIX Message header indicates the base time for Data Records. IPFIX Information Elements, described in [RFC5102], have delta time fields that indicate the time difference from the value of the Export Time field. If the Data Records include any delta time fields and the IPFIX Mediator Kobayashi, et al. Expires November 1, 2009 [Page 20] Internet-Draft Mediation Problem Statement April 2009 overwrites the Export Time field when sending IPFIX Messages, the delta time fields become meaningless and, because Collectors cannot recognize this situation, wrong time values are propagated. 6.3. Transport Sessions Management Maintaining relationships between the incoming Transport Sessions and the outgoing ones depends on the Mediator's implementation. If multiple incoming Transport Sessions are relayed to a single outgoing Transport Session, and if the IPFIX Mediators shuts down its outgoing Transport Session, Data Records on other incoming Transport Sessions would not be relayed at all. In the case of resetting of an incoming session, the behavior of the IPFIX Mediator needs to be specified. 6.4. Loss of Option Template Information In some cases, depending on the implementation of the IPFIX Mediators, the information that is reported by the Option Templates could also be lost. If, for example, the Sampling rate is not communicated from the Mediator to the Collector, the Collector would miscalculate the traffic volume. This might lead to crucial problems. Even if an IPFIX Mediator was to simply relay received Option Template Information, the values of its scope fields could become meaningless in the context of a different Transport Sessions. The minimal information to be communicated by an IPFIX Mediator must be specified. 6.5. Template ID Management The Template ID is unique on the basis of the Transport Session and Observation Domain ID. If Mediations are not able to manage the relations amongst the Template IDs and the incoming Transport Session information, and if the Template ID is used in the Options Template scope, the Mediators would, for example, relay wrong values in the scope field and in the Template Withdrawal Message. The Collector would thus not be able to interpret the Template ID in the Template Withdrawal Message and in the Options Template scope. As a consequence, there is a risk that the Collector would then shut down the IPFIX Transport Session. For example, an IPFIX Distributor must maintain the state of the incoming Transport Sessions in order to manage the Template ID on its outgoing Transport Session correctly. In the following figure, even if the Transport Session from Exporter re-initializes, the IPFIX Distributor must manage the association of Template IDs in specific Transport Session. Typically, when the Exporter#1 Transport Session re-initialized, the Template ID 256 replaced the previous Template ID 258, while the IPFIX Distributor will keep exporting the Template ID Kobayashi, et al. Expires November 1, 2009 [Page 21] Internet-Draft Mediation Problem Statement April 2009 256 to the Collector. .----------. OLD: Template ID 258 |IPFIX | NEW: Template ID 256 |Exporter#1|----+ | | | '----------' X .----------. | .-----------. .----------. |IPFIX | '---------->| | | | |Exporter#2|--------------->|IPFIX |-------X------>|IPFIX | | |Template ID 257 |Distributor|Template ID 256| Collector| '----------' +---------->| | | | .----------. | '-----------' '----------' |IPFIX | | |Exporter#3|----' | | Template ID 256 '----------' Figure C: Relaying from Multiple Transport Sessions to Single Transport Session. 6.6. Consideration for Network Topology While IPFIX Mediation can be applied anywhere, caution should be taken as how to aggregate the counters, as there is a potential risk of double-counting. For example, if three Exporters export Flow Records related to the same Flow, the one-way delay can be calculated, while the summing up the number of packets and bytes does not make sense. Alternatively, if three Exporters export Flow Records entering an administrative domain, then the sum of the packets and bytes is a valid operation. Therefore, the possible function to be applied to Flow Records must take into consideration the measurement topology. The information such as the network topology, or at least the Observation Point and measurement direction, is required on the IPFIX Mediation. 6.7. Exporting the Function Item In some case, the top IPFIX Collector needs to recognize which specific function(s) the IPFIX Mediation has executed on the Data Records. The IPFIX Collector cannot distinguish between Time Composition, Spatial Composition, and Flow Key aggregation, if the IPFIX Mediator does not export the applied function. Some parameters related to the function also would need to be exported. For example, in case of Time Composition, the active time of original Flow Records is required to interpret the minimum/maximum counter correctly. In case of Spatial Composition, spatial area information on which Data Kobayashi, et al. Expires November 1, 2009 [Page 22] Internet-Draft Mediation Problem Statement April 2009 Records is aggregated is required. 6.8. Consideration for Aggregation In case of Flow Key aggregation, Time Composition, and Spatial Composition, there are the following considerations: o Aggregation rule for non Flow Keys There are no obvious rules of non Flow Keys. For example, if an IPFIX Mediation receives two Flow Records with different DSCP values, and this DSCP field is not a Flow Key, those two Flow Records can be aggregated based on the Flow Keys value. However, there is no rules for what the DSCP value should be for the aggregated Data Record. Potential solutions are: the value of single of the two DSCP, the value 0 (in this case, the value 0 is a valid DSCP value), or removing a DSCP field in its Data Record. o Configured Selection Fraction on aggregation There is no obvious rules of how to compute Configured Selection Fraction, and whether a Mediator should report Configured Selection Fraction, when aggregation resulting from Sampling. For example, special care must be taken in the following: aggregation resulting from the different Configured Selection Fraction, aggregation resulting from different Sampling techniques, such as Systematic Count-Based Sampling and Random n-out-of-N Sampling etc. Kobayashi, et al. Expires November 1, 2009 [Page 23] Internet-Draft Mediation Problem Statement April 2009 7. Summary and Conclusion This document described the problems that network administrators have been facing, the applicability of IPFIX Mediation to these problems, and the problems related to the implementation of IPFIX Mediators. To assist the operations of the Exporters and Collectors, there are various IPFIX Mediations from which the administrators may select. Examples of the applicability of IPFIX Mediation are as follows. o Regarding large-scale measurement system, IPFIX Concentrators or IPFIX Distributors help to achieve traffic analysis with high data accuracy and fine flow granularity even as IP traffic grows. As IPFIX Mediation capabilities, Flow selection Sampling, aggregation, and composition are effective. o Regarding data retention, IPFIX Mediators enhance the export reliability, and the storage of the measurement system. o Regarding the distribution of Data Records, IPFIX Distributors help to achieve multipurpose traffic analysis for different organizations, or help to achieve respective traffic analysis based on Data Record types(IPv4, IPv6, MPLS, and VPN). o Regarding IPFIX Exporting across domains, IPFIX Masquerading Proxies help administrators to anonymize or filter Flow Records/ Packet Reports, preventing privacy violations. o Regarding interoperability, IPFIX Proxies provide interoperability between legacy protocols and IPFIX, even during the migration period to IPFIX. As a result, the IPFIX Mediation benefits become apparent. However, there are still some open issues with the use of IPFIX Mediators. o Both Observation Point and IPFIX Message header information, such as the Exporter IP address, Observation Domain ID, and Export Time field, might be lost. This data should therefore be communicated between the Original Exporter and Collector via the IPFIX Mediator. o IPFIX Mediators are required to manage Transport Sessions, Template IDs, and Observation Domain IDs. Otherwise, anomalous IPFIX messages could be created. o Data advertised by Option Templates from the Original Exporter, such as the Sampling rate and Sampling algorithm used, might be lost. If a Collector is not informed of current Sampling rates, traffic information might become worthless. Kobayashi, et al. Expires November 1, 2009 [Page 24] Internet-Draft Mediation Problem Statement April 2009 These problems stem from the fact that no standards regarding IPFIX Mediation have been set. In particular, the minimum set of information that should be communicated between Original Exporters and Collectors, the management between different IPFIX Transport Sessions, and the internal components of IPFIX Mediators should be standardized. Kobayashi, et al. Expires November 1, 2009 [Page 25] Internet-Draft Mediation Problem Statement April 2009 8. Security Considerations A flow-based measurement system must prevent potential security threats: the disclosure of confidential traffic data, injection of incorrect data, and unauthorized access to traffic data. These security threats of the IPFIX protocol are covered by the security considerations section in [RFC5101] and are still valid for IPFIX Mediators. And a measurement system must also prevent following security threats related to IPFIX Mediation: o Attacks against IPFIX Mediator IPFIX Mediators can be considered as a prime target for attacks, as an alternative to IPFIX Exporters and Collectors. IPFIX Proxies or Masquerading Proxies need to prevent unauthorized access or denial-of-service (DoS) attacks from untrusted public networks. o Man-in-the-middle attack by untrusted IPFIX Mediator The Collector-Mediator-Exporter structure model would increase the risk of the man-in-the-middle attack. o Configuration on IPFIX Mediation In the case of IPFIX Distributors and IPFIX Masquerading Proxies, an accidental misconfiguration and unauthorized access to configuration data could lead to the crucial problem of disclosure of confidential traffic data. Kobayashi, et al. Expires November 1, 2009 [Page 26] Internet-Draft Mediation Problem Statement April 2009 9. IANA Considerations This document has no actions for IANA. Kobayashi, et al. Expires November 1, 2009 [Page 27] Internet-Draft Mediation Problem Statement April 2009 10. Acknowledgements The authors would like to thank Gerhard Muenz, Keisuke Ishibashi and Nevil Brownlee for providing valuable comments and suggestions. Kobayashi, et al. Expires November 1, 2009 [Page 28] Internet-Draft Mediation Problem Statement April 2009 11. References 11.1. Normative References [RFC5101] Claise, B., "Specification of the IP Flow Information Export (IPFIX) Protocol for the Exchange of IP Traffic Flow Information", January 2008. [RFC5476] Claise, B., "Packet Sampling (PSAMP) Protocol Specifications", March 2009. 11.2. Informative References [I-D.ietf-ipfix-file] Trammell, B., Boschi, E., Mark, L., Zseby, T., and A. Wagner, "Specification of the IPFIX File Format", draft-ietf-ipfix-file-05 (work in progress) , November 2007. [I-D.ietf-ipfix-mib] Dietz, T., Claise, B., and A. Kobayashi, "Definitions of Managed Objects for IP Flow Information Export", draft-ietf-ipfix-mib-06 (work in progress) , March 2009. [I-D.ietf-psamp-mib] Dietz, T. and B. Claise, "Definitions of Managed Objects for Packet Sampling", draft-ietf-psamp-mib-06 (work in progress) , June 2006. [RFC3917] Quittek, J., Zseby, T., Claise, B., and S. Zander, "Requirements for IP Flow Information Export(IPFIX)", October 2004. [RFC3954] Claise, B., "Cisco Systems NetFlow Services Export Version 9", October 2004. [RFC5102] Quittek, J., Bryant, S., Claise, B., Aitken, P., and J. Meyer, "Information Model for IP Flow Information Export", January 2008. [RFC5470] Sadasivan, G., Brownlee, N., Claise, B., and J. Quittek, "Architecture for IP Flow Information Export", March 2009. [RFC5472] Zseby, T., Boschi, E., Brownlee, N., and B. Claise, "IP Flow Information Export (IPFIX) Applicability", March 2009. [RFC5474] Duffield, N., "A Framework for Packet Selection and Kobayashi, et al. Expires November 1, 2009 [Page 29] Internet-Draft Mediation Problem Statement April 2009 Reporting", March 2009. [RFC5475] Zseby, T., Molina, M., Duffield, N., Niccolini, S., and F. Raspall, "Sampling and Filtering Techniques for IP Packet Selection", March 2009. [RFC5477] Dietz, T., Claise, B., Aitken, P., Dressler, F., and G. Carle, "Information Model for Packet Sampling Exports", March 2009. [TRAFGRW] Cho, K., Fukuda, K., Esaki, H., and A. Kato, "The Impact and Implications of the Growth in Residential User-to-User Traffic", SIGCOMM2006, pp. 207-218, Pisa, Italy, September 2006. . Kobayashi, et al. Expires November 1, 2009 [Page 30] Internet-Draft Mediation Problem Statement April 2009 Authors' Addresses Atsushi Kobayashi NTT Information Sharing Platform Laboratories 3-9-11 Midori-cho Musashino-shi, Tokyo 180-8585 Japan Phone: +81-422-59-3978 Email: akoba@nttv6.net URI: http://www3.plala.or.jp/akoba/ Benoit Claise Cisco Systems, Inc. De Kleetlaan 6a b1 Diegem 1831 Belgium Phone: +32 2 704 5622 Email: bclaise@cisco.com Haruhiko Nishida NTT Information Sharing Platform Laboratories 3-9-11 Midori-cho Musashino-shi, Tokyo 180-8585 Japan Phone: +81-422-59-3978 Email: nishida.haruhiko@lab.ntt.co.jp Christoph Sommer University of Erlangen-Nuremberg Department of Computer Science 7 Martensstr. 3 Erlangen 91058 Germany Phone: +49 9131 85-27993 Email: christoph.sommer@informatik.uni-erlangen.de URI: http://www7.informatik.uni-erlangen.de/~sommer/ Kobayashi, et al. Expires November 1, 2009 [Page 31] Internet-Draft Mediation Problem Statement April 2009 Falko Dressler University of Erlangen-Nuremberg Department of Computer Science 7 Martensstr. 3 Erlangen 91058 Germany Phone: +49 9131 85-27914 Email: dressler@informatik.uni-erlangen.de URI: http://www7.informatik.uni-erlangen.de/~dressler/ Stephan Emile France Telecom 2 avenue Pierre Marzin Lannion, F-22307 Fax: +33 2 96 05 18 52 Email: emile.stephan@orange-ftgroup.com Kobayashi, et al. Expires November 1, 2009 [Page 32]