RFC 3346

Applicability Statement for Traffic Engineering with MPLS

📅 published August 2002, from ietf

RFC 3346: Applicability Statement for Traffic Engineering with MPLS ✨

Okay bestie, let's do the vibe check first 💅. This is RFC 3346 from the Network Working Group, and its title is "Applicability Statement for Traffic Engineering with MPLS." The authors are J. Boyle (PD Nets), V. Gill (AOL Time Warner, Inc.), A. Hannan (RoutingLoop), D. Cooper (Global Crossing), D. Awduche (Movaz Networks), B. Christian (Worldcom), and W.S. Lai (AT&T). It's marked Informational and dated August 2002 🎀. The header doesn't list anything it obsoletes or updates, so it's a standalone vibe 🦋.

Status of this Memo 🌸

This memo gives information to the Internet community. It is not an Internet standard of any kind, and distribution of it is unlimited 💖.

Copyright (C) The Internet Society (2002). All Rights Reserved.

Abstract ✨

This doc is all about how Multiprotocol Label Switching (MPLS) applies to traffic engineering in IP networks 🧚. It covers the special things to think about when you deploy MPLS for traffic engineering in real operational setups, and it also spills the tea on the limitations of the MPLS approach to traffic engineering 👀.

Table of Contents 🍓

   1.  Introduction....................................................2
   2.  Technical Overview of ISP Traffic Engineering...................3
   3.  Applicability of Internet Traffic Engineering...................4
   3.1 Avoidance of Congested Resources................................4
   3.2 Resource Utilization in Network Topologies with Parallel Links..5
   3.3 Implementing Routing Policies using Affinities..................5
   3.4 Re-optimization After Restoration...............................6
   4.  Implementation Considerations...................................6
   4.1 Architectural and Operational Considerations....................6
   4.2 Network Management Aspects......................................7
   4.3 Capacity Engineering Aspects....................................8
   4.4 Network Measurement Aspects.....................................8
   5.  Limitations.....................................................9
   6.  Conclusion.....................................................11
   7.  Security Considerations........................................11
   8.  References.....................................................11
   9.  Acknowledgments................................................12
   10. Authors' Addresses.............................................13
   11. Full Copyright Statement.......................................14

1. Introduction 💫

Everyone agrees that one of the biggest first uses of Multiprotocol Label Switching (MPLS) is traffic engineering (TE) [1][2] in IP networks 🔥. A big community of IP service providers has found that traffic engineering their networks can be super valuable, both tactically and strategically [2, 3, 4] 💅. To support this, extensions have been specified for the Interior Gateway Protocols (IGP) IS-IS [5] and OSPF [6], and for the signaling protocols RSVP [7] and LDP [8] 🎀. The IS-IS, OSPF, and RSVP extensions have all been built and rolled out at large scale in lots of networks made of multi-vendor equipment, so they're not just theory, they're the real deal 😭.

This document talks about how TE applies to Internet service provider networks, with the focus on the MPLS-based approach 🦋. It adds to the existing protocol applicability statements and specifically relates to the operational applicability of RSVP-TE [9] 🌸. It also covers the special considerations for deploying MPLS in operational contexts, and it highlights the limitations of this approach to traffic engineering 🫶.

2. Technical Overview of ISP Traffic Engineering 🧚

Traffic engineering (TE) is generally about optimizing the performance of operational networks [2] ✨. These days, TE means mapping IP traffic flows onto the existing physical network topology in the most effective way possible, so you hit your operational goals 💖. The techniques people use today include, but aren't limited to:

  1. Manipulation of IGP cost (metrics)
  2. Explicit routing using constrained virtual-circuit switching techniques such as ATM or Frame Relay SPVCs
  3. Explicit routing using constrained path setup techniques such as MPLS

This document is mainly about MPLS techniques 🌈. Specifically, it's about using paths other than the shortest paths the IGP picks, to get more balanced network utilization. For example, you can move traffic off the IGP-selected shortest paths onto alternate paths to dodge congestion in the network 🏃‍♀️. You can do this by using explicitly signaled LSP-tunnels 🔥. The explicit routes can be computed offline, then downloaded and configured on the routers using an appropriate mechanism 👑. Or you can configure the desired characteristics of an LSP (like endpoints, bandwidth, and affinities) on a router, and it will use an appropriate algorithm to compute a path through the network that satisfies those characteristics, subject to various types of constraints 🧠. In general, the characteristics associated with LSPs may include:

  • Ingress and egress nodes
  • Bandwidth required
  • Priority
  • Nodes to include or exclude in the path
  • Affinities to include or exclude in the path
  • Resilience requirements

Affinities are arbitrary, provider-assigned attributes applied to links, and they're carried in the TE extensions for the IGPs 🎀. They give links a class structure, so different links can be logically grouped together 🧚. You can use them to set up various types of policies or route preferences, like including or excluding groups of links from the path of LSPs 💅. Affinities are unique to MPLS, and the original requirement for them was documented in [2] 🌸.

3. Applicability of Internet Traffic Engineering 🦋

As mentioned in [2] and [7], traffic engineering with MPLS is a good fit for setting up and maintaining explicitly routed paths in an IP network, so you can place traffic effectively 💜. LSP-tunnels can forward subsets of traffic along paths that are independent of the routes computed by conventional IGP Shortest Path First (SPF) algorithms 🚀. This gives network operators a lot of flexibility in controlling the paths of traffic flows across their networks, and it lets them implement policies that can improve network performance 🙌. Some examples of scenarios where MPLS-based TE capabilities apply in service provider environments are below, but MPLS is definitely not limited to these scenarios 🫶.

3.1 Avoidance of Congested Resources 🔥

To lower the utilization of congested link(s), an operator can use TE methods to route a subset of traffic away from those links onto less congested topological elements 🌊. These techniques work well when some parts of the network are congested while others are underutilized 😤.

Operators who don't make heavy use of LSP-tunnels might take a tactical approach to MPLS TE, creating LSP-tunnels only when they're needed to fix specific congestion problems 🎯. For example, an LSP can be created between two nodes (a source and a destination) that are known to send traffic into a congested network element, and the LSP can be explicitly routed through a separate path to divert some traffic away from the congestion 🦋. On the other hand, operators who make heavy use of LSP-tunnels, whether for measurement or automated traffic control, may decide to explicitly route a subset of the LSPs that cross the point of congestion onto alternate paths 🧚. This can be a quick response when the bandwidth parameter on the LSPs doesn't accurately reflect the traffic they actually carry, and the operator decides that changing the bandwidth parameter values might not work or might not have a lasting impact 💅.

There are other approaches that measure traffic workloads on LSPs and use those empirical statistics to configure various characteristics of LSPs 📊. For example, they can use the derived statistics to configure explicit routes for LSPs (also known as offline TE [10]) 🗺️. They can also use the statistics to set values for LSP attributes like bandwidths, priority, and affinities (online TE) ⚡. All of these approaches can be used both tactically and strategically to react to periods of congestion in a network 🫶. Congestion can happen for lots of reasons: equipment or facility failure, longer-than-expected provisioning cycles for new circuits, and surprise surges in traffic demand 😭.

In practice, lots of service provider networks have multiple parallel links between nodes 🌈. One example is transoceanic connectivity, which is often provisioned as numerous low-capacity circuits, such as NxDS-3 (N parallel DS-3 circuits) and NxSTM-1 (N parallel STM-1 circuits) 🌊. Parallel circuits also show up a lot in bandwidth-constrained cities 🏙️. MPLS TE methods can be applied to spread the aggregate traffic workload across these parallel circuits effectively 💖.

The MPLS-based approaches people commonly use for parallel links include using LSP bandwidth parameters to control the proportion of demand that traverses each link, explicitly configuring routes for LSP-tunnels so they're distributed across the parallel links, and using affinities to map different LSPs onto different links 🧠. These solutions also apply in networks with parallel and replicated topologies, like an NxOC-3/12/48 topology 🎀.

3.3 Implementing Routing Policies using Affinities 👑

Sometimes you want to restrict certain types of traffic to certain types of links, or explicitly exclude certain types of links from the paths for some types of traffic 🙅‍♀️. This might be needed to meet a business policy or a network engineering objective 💼. MPLS resource affinities give you a powerful way to implement these kinds of objectives 💪.

As a concrete example, imagine a global service provider with a flat (non-hierarchical) IGP 🌍. MPLS TE affinities can be used to explicitly keep continental traffic (traffic originating and terminating within a continent) from crossing transoceanic resources 🌊.

Another example is using MPLS TE affinities to keep certain traffic off a subset of circuits. For instance, you might keep inter-regional LSPs off circuits that are reserved for intra-regional traffic 🚫.

Still another example is a heterogeneous network made of links with different capacities, like OC-12, OC-48, and OC-192 🌈. In networks like that, affinities can force some types of traffic to only traverse links with a given capacity, like OC-48 🦋.

3.4. Re-optimization After Restoration ✨

okay so after a network failure goes down, it may be desirable to calculate a new set of paths for LSPs to optimize performance over the residual topology 🛣️ this re-optimization is complementary to fast-reroute, which is used to reduce packet losses during routing transients under network restoration 💖 traffic protection can also be accomplished by associating a primary LSP with a set of secondary LSPs, hot-standby LSPs, or a combination thereof [11] 🦋

4. Implementation Considerations 🎀

4.1. Architectural and Operational Considerations 💅

when deploying TE solutions in a service provider environment, the impact of administrative policies and the selection of nodes that will serve as endpoints for LSP-tunnels SHOULD be carefully considered 👑 as noted in [9], when devising a virtual topology for LSP-tunnels, special consideration SHOULD be given to the tradeoff between the operational complexity associated with a large number of LSP-tunnels and the control granularity that large numbers of LSP-tunnels allow 🧠 in other words, a large number of LSP-tunnels allows greater control over the distribution of traffic across the network, but increases network operational complexity 😭 in large networks, it MAY be advisable to start with a simple LSP-tunnel virtual topology and then introduce additional complexity based on observed or anticipated traffic flow patterns [9] 🌸

administrative policies SHOULD guide the amount of bandwidth to be allocated to an LSP 🍓 one MAY choose to set the bandwidth of a particular LSP to a statistic of the measured observed utilization over an interval of time, e.g., peak rate, or a particular percentile or quartile of the observed utilization ✨ sufficient over-subscription (of LSPs) or under-reporting bandwidth on the physical links SHOULD be used to account for flows that exceed their normal limits on an event-driven basis 🔥 flows SHOULD be monitored for trends that indicate when the bandwidth parameter of an LSP should be resized 👀 flows SHOULD be monitored constantly to detect unusual variance from expected levels 💜 if an unpoliced flow greatly exceeds its assigned bandwidth, action SHOULD be taken to determine the root cause and remedy the problem 🚨 traffic policing is an option that MAY be applied to deal with congestion problems, especially when some flows exceed their bandwidth parameters and interfere with other compliant flows 🧚 however, it is usually more prudent to apply policing actions at the edge of the network rather than within the core, unless under exceptional circumstances 🌈

when creating LSPs, a hierarchical network approach MAY be used to alleviate scalability problems associated with flat full mesh virtual topologies 🫶 in general, operational experience has shown that very large flows (between city pairs) are long-lived and have stable characteristics, while smaller flows (edge to edge) are more dynamic and have more fluctuating statistical characteristics 🌟 a hierarchical architecture can be devised consisting of core and edge networks in which the core is traffic engineered and serves as an aggregation and transit infrastructure for edge traffic 💖

however, over-aggregation of flows can result in a stream so large that it precludes the constraint-based routing algorithm from finding a feasible path through a network 😱 splitting a flow by using two or more parallel LSPs and distributing the traffic across the LSPs can solve this problem, at the expense of introducing more state in the network 🍬

failure scenarios SHOULD also be addressed when splitting a stream of traffic over several links 🛠️ it is of little value to establish a finely balanced set of flows over a set of links only to find that upon link failure the balance reacts poorly, or does not revert to the original situation upon restoration 💀

4.2. Network Management Aspects 📊

networks planning to deploy MPLS for traffic engineering MUST consider network management aspects, particularly performance and fault management [12] 🎀 with the deployment of MPLS in any infrastructure, some additional operational tasks are required, such as constant monitoring to ensure that the performance of the network is not impacted in the end-to-end delivery of traffic 🌷 in addition, traffic characteristics, such as latency across an LSP, MAY also need to be assessed on a regular basis to ensure that service-level guarantees are achieved 💫

obtaining information on LSP behavior is critical in determining the stability of an MPLS network 🔍 when LSPs transition or path changes occur, packets may be dropped, which impacts network performance 😬 it SHOULD be the goal of any network deploying MPLS to minimize the volatility of LSPs and reduce the root causes that induce this instability 🎯 unfortunately, there are very few, if any, NMS systems that are available at this time with the capability to provide the correct level of management support, particularly root cause analysis 🙃 consequently, most early adopters of MPLS develop their own management systems in-house for the MPLS domain 🛠️ the lack of availability of commercial network management systems that deal specifically with MPLS-related aspects is a significant impediment to the large-scale deployment of MPLS networks 🚧

the performance of an MPLS network is also dependent on the configured values of bandwidth for each LSP 💅 since congestion is a common cause of performance degradation in operational networks, it is important to proactively avoid these situations 🌊 while MPLS was designed to minimize congestion on links by utilizing bandwidth reservations, it is still heavily reliant on user configurable data 🥹 if the LSP bandwidth value does not properly represent the traffic demand of that LSP, over-utilization MAY occur and cause significant congestion within the network 🚦 therefore, it is important to develop, deploy, and maintain a good modeling tool for determining LSP bandwidth size 📐 lack of this capability MAY result in sub-optimal network performance 🫠

4.3. Capacity Engineering Aspects 🏗️

traffic engineering has a goal of ensuring traffic performance objectives for different services 🎯 this requires that the different network elements be dimensioned properly to handle the expected load 📦 more specifically, in mapping given user demands onto network resources, network dimensioning involves the sizing of the network elements, such as links, processors, and buffers, so that performance objectives can be met at minimum cost 💸 major inputs to the dimensioning process are cost models, characterization of user demands, and specification of performance objectives 📝

in using MPLS, dimensioning involves the assignment of resources such as bandwidth to a set of pre-selected LSPs for carrying traffic, and mapping the logical network of LSPs onto a physical network of links with capacity constraints 🗺️ the dimensioning process also determines the link capacity parameters or thresholds associated with the use of some bandwidth reservation scheme for service protection 🛡️ service protection controls the QoS for certain service types by restricting access to bandwidth, or by giving priority access to one type of traffic over another 👑 such methods are essential, e.g., to prevent starvation of low-priority flows, to guarantee a minimum amount of resources for flows with expected short duration, to improve the acceptance probability for flows with high bandwidth requirements, or to maintain network stability by preventing performance degradation in case of a local overload 🌈

4.4. Network Measurement Aspects 📈

network measurement entails robust statistics collection and systems development 🧪 knowing what to do with these measurements is often where the secret sauce is 🤫 examples for different applications of measurements are described in [13] 📚 for instance, to ensure that the QoS objectives have been met, performance measurements and performance monitoring are required so that real-time traffic control actions, or policy-based actions, can be taken 🚀 also, to characterize the traffic demands, traffic measurements are used to estimate the offered loads from different service classes and to provide forecasting of future demands for capacity planning purposes 🔮 forecasting and planning MAY result in capacity augmentation or MAY lead to the introduction of new technology and architecture 🏙️

to avoid QoS degradation at the packet level, measurement-based admission control can be employed by using online measurements of actual usage 🚪 this is a form of preventive control to ensure that the QoS requirements of different service classes can be met simultaneously, while maintaining network efficiency at a high level 🧘‍♀️ however, it requires proper network dimensioning to keep the probability for the refusal of connection/flow requests sufficiently low 💭

5. Limitations 🥺

okay so this is the reality check bestie 💅 getting a real understanding of how MPLS works takes a LOT of effort, and you might need serious engineering and testing time just to find problems with vendor implementations 🔍 starting MPLS software and sorting out the configuration management side of TE can eat up serious engineering, operations, and system development resources 😩 building automated systems to generate router configurations for network elements can need major software development and hardware resources too, and getting to the point where router configs update automatically is a slow journey 🐢 tracking manual tweaks to router configs (or the problems they cause) can feel like an endless task 💀 so yeah, you're gonna need a lot more tools to simplify and automate the MPLS TE function 🛠️

certain architectural choices can cause scalability problems in operations, in protocols, and in router implementations 📉 this gets worse as the number of LSP-tunnels or router configuration data grows, and it gets especially messy with full meshes, which create O(N^2) LSPs, where N is the number of network-edge nodes 🕸️ in these cases, shrinking N through hierarchy, regionalization, or picking the right tunnel termination points can help the network's ability to scale ✨ losing scale can look like protocol instability, not being able to change configs to handle growth, routers not being updatable, routers not being able to hold or properly process configs, or just losing the ability to manage the network properly 😵‍💫

MPLS TE is widely deployed, but it's still a newer technology compared to the classic IP routing protocols like IS-IS, OSPF, and BGP 👵 it also comes with way more configuration and protocol options 🎛️ so some implementations aren't battle-hardened yet, and automated testing of different configurations is difficult, if not basically impossible 🙈 multi-vendor environments are starting to show up, but usually you need extra effort to make sure everything fully interoperates 🤝

common TE approaches in service provider environments switch the forwarding paradigm from connectionless to connection oriented 🔄 so analyzing the network operationally can get complicated in some ways (and get easier in others) 🧩 inconsistencies in forwarding state cause dropped packets, whereas with connectionless methods a packet will either loop and get dropped, or get misdirected onto another branch in the routing tree 🌿

currently deployed MPLS TE approaches can be negatively affected by both internal and external router and link failures ⚡ this can create a mismatch between the signaled capacity and the traffic an LSP-tunnel actually carries 😬

Many routers in service provider environments are already stressed out just processing the software workload of running IGP, BGP, and IPC 😫 turning on TE in an MPLS environment means adding traffic engineering databases and processes, adding extra information for the routing processes to carry, and adding signaling state and processing to these network elements 📦 extra traffic measurements might also need support. in some environments, this extra load just isn't feasible 🚫

MPLS in general, and MPLS-TE in particular, is NOT a magic fix for lack of network capacity, or for lack of proper capacity planning and provisioning in the network dimensioning process 🙅‍♀️ MPLS-TE can make traffic travel greater distances or take paths with more network elements, which means more latency 🐌 usually this added inefficiency is there to cover shortcomings in capacity planning or available resources, or to route around hot spots 🔥 TE's ability to fit more traffic on a given topology can also be a short-term gain during periods of persistent traffic growth ⏳ these approaches can't achieve impossible mappings of traffic onto topologies 🚫 if you don't capacity plan and execute properly, congestion is coming, no matter what technology helpers you use 💥

6. Conclusion 👑

so that's the wrap-up 🎀 this document talked about the applicability of traffic engineering in Internet service provider environments, with the focus on MPLS-based approaches to get TE done in this context ✨ we covered the applicability of traffic engineering, the management and deployment considerations that come with it, and the limitations we highlighted 📝

MPLS is kinda iconic because it combines the ability to monitor point-to-point traffic statistics between two routers with the capability to control the forwarding paths of subsets of traffic through a given network topology 🦋 that's what makes TE with MPLS useful for improving network performance by effectively mapping traffic flows onto links within service provider networks 🗺️ tools that simplify and automate the MPLS TE functions and activation help you actually reach the full potential 🚀

7. Security Considerations 🔐

okay this part is serious, no cap 🫡 this document does not introduce new security issues ✅ but when it's deployed in service provider networks, it is mandatory to ensure that only authorized entities are permitted to initiate establishment of LSP-tunnels 🚨 like, only the right people get to start those tunnels, periodt 🔒

8. References 📚

  1. Rosen, E., Viswanathan, A. and R. Callon, "Multiprotocol Label Switching Architecture," RFC 3031, January 2001.
  2. Awduche, D., Malcolm, J., Agogbua, J., O'Dell, M. and J. McManus, "Requirements for Traffic Engineering Over MPLS," RFC 2702, September 1999.
  3. X. Xiao, A. Hannan, B. Bailey, and L. Ni, "Traffic Engineering with MPLS in the Internet," IEEE Network, March/April 2000.
  4. V. Springer, C. Pierantozzi, and J. Boyle, "Level3 MPLS Protocol Architecture," Work in Progress.
  5. T. Li, and H. Smit, "IS-IS Extensions for Traffic Engineering," Work in Progress.
  6. D. Katz, D. Yeung, and K. Kompella, "Traffic Engineering Extensions to OSPF," Work in Progress.
  7. Awduche, D., Berger, L., Gan, D.H., Li, T., Srinivasan, V. and G. Swallow, "RSVP-TE: Extensions to RSVP for LSP Tunnels," RFC 3209, December 2001.
  8. Jamoussi, B. (Editor), "Constraint-Based LSP Setup using LDP," RFC 3212, January 2002.
  9. Awduche, D., Hannan, A. and X. Xiao, "Applicability Statement for Extensions to RSVP for LSP-Tunnels," RFC 3210, December 2001.
  10. Awduche, D., Chiu, A., Elwalid, A., Widjaja, I. and X. Xiao, "Overview and Principles of Internet Traffic Engineering", RFC 3272, May 2002.
  11. W.S. Lai, D. McDysan, J. Boyle, M. Carlzon, R. Coltun, T. Griffin, E. Kern, and T. Reddington, "Network Hierarchy and Multilayer Survivability," Work in Progress.
  12. D. Awduche, "MPLS and Traffic Engineering in IP Networks," IEEE Communications Magazine, December 1999.
  13. W.S. Lai, B. Christian, R.W. Tibbs, and S. Van den Berghe, "A Framework for Internet Traffic Engineering Measurement," Work in Progress.

9. Acknowledgments 🫶

the reason MPLS protocols work so well for traffic engineering in service provider networks is mostly thanks to the experience and foresight of network engineers and developers who already knew traffic engineering with ATM in these environments 🙌 the authors specifically want to thank the authors of RFC 2702 for clearly articulating the requirements 💖 and also the developers and testers of code in deployment today, for keeping their focus locked in 🔥

10. Authors' Addresses 💌

Jim Boyle
Protocol Driven Networks
Tel: +1 919-852-5160
EMail: jboyle@pdnets.com

Vijay Gill
AOL Time Warner, Inc.
12100 Sunrise Valley Drive
Reston, VA 20191
EMail: vijay@umbc.edu

Alan Hannan
RoutingLoop Intergalactic
112 Falkirk Court
Sunnyvale, CA 94087, USA
Tel: +1 408-666-2326
EMail: alan@routingloop.com

Dave Cooper
Global Crossing
960 Hamlin Court
Sunnyvale, CA 94089, USA
Tel: +1 916-415-0437
EMail: dcooper@gblx.net

Daniel O. Awduche
Movaz Networks
7926 Jones Branch Drive, Suite 615
McLean, VA 22102, USA
Tel: +1 703-298-5291
EMail: awduche@movaz.com

Blaine Christian
Worldcom
22001 Loudoun County Parkway, Room D1-2-737
Ashburn, VA 20147, USA
Tel: +1 703-886-4425
EMail: blaine@uu.net

Wai Sum Lai
AT&T
200 Laurel Avenue
Middletown, NJ 07748, USA
Tel: +1 732-420-3712
EMail: wlai@att.com

bestie, the legal stuff stays word-for-word so nothing gets lost in translation 🙏

Copyright (C) The Internet Society (2002).  All Rights Reserved.

This document and translations of it may be copied and furnished to
others, and derivative works that comment on or otherwise explain it
or assist in its implementation may be prepared, copied, published
and distributed, in whole or in part, without restriction of any
kind, provided that the above copyright notice and this paragraph are
included on all such copies and derivative works.  However, this
document itself may not be modified in any way, such as by removing
the copyright notice or references to the Internet Society or other
Internet organizations, except as needed for the purpose of
developing Internet standards in which case the procedures for
copyrights defined in the Internet Standards process must be
followed, or as required to translate it into languages other than
English.

The limited permissions granted above are perpetual and will not be
revoked by the Internet Society or its successors or assigns.

This document and the information contained herein is provided on an
"AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
TASK FORCE DISCLAIMS 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.

Acknowledgement 💜

funding for the RFC Editor function is currently provided by the Internet Society 🌈

📦 where this page comes from