Network Working Group N. Kumar
Internet-Draft R. Asati
Intended status: Informational Cisco
Expires: April 28, 2015 M. Chen
X. Xu
Huawei
October 25, 2014

BIER Use Cases
draft-kumar-bier-use-cases-00.txt

Abstract

Bit Index Explicit Replication (BIER) is an architecture that provides optimal multicast forwarding through a "BIER domain" without requiring intermediate routers to maintain any multicast related per-flow state. BIER also does not require any explicit tree-building protocol for its operation. A multicast data packet enters a BIER domain at a "Bit-Forwarding Ingress Router" (BFIR), and leaves the BIER domain at one or more "Bit-Forwarding Egress Routers" (BFERs). The BFIR router adds a BIER header to the packet. The BIER header contains a bit-string in which each bit represents exactly one BFER to forward the packet to. The set of BFERs to which the multicast packet needs to be forwarded is expressed by setting the bits that correspond to those routers in the BIER header.

This document describes some of the use-cases for BIER.

Status of This Memo

This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at http://datatracker.ietf.org/drafts/current/.

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."

This Internet-Draft will expire on April 28, 2015.

Copyright Notice

Copyright (c) 2014 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 (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.


Table of Contents

1. Introduction

Bit Index Explicit Replication (BIER) [I-D.wijnands-bier-architecture] is an architecture that provides optimal multicast forwarding through a "BIER domain" without requiring intermediate routers to maintain any multicast related per-flow state. BIER also does not require any explicit tree-building protocol for its operation. A multicast data packet enters a BIER domain at a "Bit-Forwarding Ingress Router" (BFIR), and leaves the BIER domain at one or more "Bit-Forwarding Egress Routers" (BFERs). The BFIR router adds a BIER header to the packet. The BIER header contains a bit-string in which each bit represents exactly one BFER to forward the packet to. The set of BFERs to which the multicast packet needs to be forwarded is expressed by setting the bits that correspond to those routers in the BIER header.

The obvious advantage of BIER is that there is no per flow multicast state in the core of the network and there is no tree building protocol that sets up tree on demand based on users joining a multicast flow. In that sense, BIER is potentially applicable to many services where Multicast is used and not limited to the examples described in this draft. In this document we are describing a few use-cases where BIER could provide benefit over using existing mechanisms.

2. Specification of Requirements

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 [RFC2119].

3. BIER Use Cases

3.1. Multicast in L3VPN Networks

The Multicast L3VPN architecture [RFC6513] describes many different profiles in order to transport L3 Multicast across a providers network. Each profile has its own different tradeoffs (see section 2.1 [RFC6513]). When using "Multidirectional Inclusive" “Provider Multicast Service Interface” (MI-PMSI) an efficient tree is build per VPN, but causes flooding of egress PE’s that are part of the VPN, but have not joined a particular C-multicast flow. This problem can be solved with the “Selective” PMSI to build a special tree for only those PE’s that have joined the C-multicast flow for that specific VPN. The more S-PMSI’s, the less bandwidth is wasted due to flooding, but causes more state to be created in the providers network. This is a typical problem network operators are faced with by finding the right balance between the amount of state carried in the network and how much flooding (waste of bandwidth) is acceptable. Some of the complexity with L3VPN’s comes due to providing different profiles to accommodate these trade-offs.

With BIER there is no trade-off between State and Flooding. Since the receiver information is explicitly carried within the packet, there is no need to build S-PMSI’s to deliver multicast to a sub-set of the VPN egress PE’s. Due to that behaviour, there is no need for S-PMSI’s.

Mi-PMSI’s and S-PMSI’s are also used to provide the VPN context to the Egress PE router that receives the multicast packet. Also, in some MVPN profiles it is also required to know which Ingress PE forwarded the packet. Based on the PMSI the packet is received from, the target VPN is determined. This also means there is a requirement to have a least a PMSI per VPN or per VPN/Ingress PE. This means the amount of state created in the network is proportional to the VPN and ingress PE’s. Creating PMSI state per VPN can be prevented by applying the procedures as documented in [RFC5331]. This however has not been very much adopted/implemented due to the excessive flooding it would cause to Egress PE’s since *all* VPN multicast packets are forwarded to *all* PE’s that have one or more VPN’s attached to it.

With BIER, the destination PE’s are identified in the multicast packet, so there is no flooding concern when implementing [RFC5331]. For that reason there is no need to create multiple BIER domain’s per VPN, the VPN context can be carry in the multicast packet using the procedures as defined in [RFC5331]. Also see [I-D.rosen-l3vpn-mvpn-bier] for more information.

With BIER only a few MVPN profiles will remain relevant, simplifying the operational cost and making it easier to be interoperable among different vendors.

3.2. IPTV Services

IPTV is a service, well known for its characteristics of allowing both live and on-demand delivery of media traffic over IP. In a typical IPTV environment the egress routers connecting to the receivers will build the tree towards the ingress router connecting to the IPTV servers. The egress routers would rely on IGMP/MLD (static or dynamic) to learn about the receiver's interest in one or more multicast group/channels. Interestingly, BIER could allows provisioning any new multicast group/channel by only modifying the channel mapping on ingress routers. This is deemed beneficial for the linear IPTV video broadcasting in which every receivers behind every egress PE routers would receive the IPTV video traffic.

With BIER, there is no need of tree building from egress to ingress. Further, any addition of new channel or new egress routers can be directly controlled from ingress router. When a new channel is included, the multicast group is mapped to Bit string that includes all egress routers. Ingress router would start sending the new channel and deliver it to all egress routers. As it can be observed, there is no need for static IGMP provisioning in each egress routers whenever a new channel/stream is added. Instead, it can be controlled from ingress router itself by configuring the new group to Bit Mask mapping on ingress router.

3.3. Data center Virtualization/Overlay

Virtual eXtensible Local Area Network (VXLAN) [RFC7348] is a kind of network virtualization overlay technology which is intended for multi-tenancy data center networks. To emulate a layer2 flooding domain across the layer3 underlay, it requires to have a mapping between the VXLAN Virtual Network Instance (VNI) and the IP multicast group in a ratio of 1:1 or n:1. In other words, it requires to enable the multicast capability in the underlay. For instance, it requires to enable PIM-SM [RFC4601] or PIM-BIDIR [RFC5015] multicast routing protocol in the underlay. VXLAN is designed to support 16M VNIs at maximum. In the mapping ratio of 1:1, it would require 16M multicast groups in the underlay which would become a significant challenge to both the control plane and the data plane of the data center switches. In the mapping ratio of n:1, it would result in inefficiency bandwidth utilization which is not optimal in data center networks. More importantly, it is recognized by many data center operators as a unaffordable burden to run multicast in data center networks from network operation and maintenance perspectives. As a result, many VXLAN implementations are claimed to support the ingress replication capability since ingress replication eliminates the burden of running multicast in the underlay. Ingress replication is an acceptable choice in small-sized networks where the average number of receivers per multicast flow is not too large. However, in multi-tenant data center networks, especially those in which the NVE functionality is enabled on a high amount of physical servers, the average number of NVEs per VN instance would be very large. As a result, the ingress replication scheme would result in a serious bandwidth waste in the underlay and a significant replication burden on ingress NVEs.

With BIER, there is no need for maintaining that huge amount of multicast states in the underlay anymore while the delivery efficiency of overlay BUM traffic is the same as if any kind of stateful multicast protocols such as PIM-SM or PIM-BIDIR is enabled in the underlay.

4. Security Considerations

There are no security issues introduced by this draft.

5. IANA Considerations

There are no IANA consideration introduced by this draft.

6. Acknowledgments

The authors would like to thank IJsbrand Wijnands, Greg Shepherd and Christian Martin for their contribution.

7. References

7.1. Normative References

[I-D.rosen-l3vpn-mvpn-bier] Rosen, E., Sivakumar, M., Wijnands, I., Aldrin, S., Dolganow, A. and T. Przygienda, "Multicast VPN Using BIER", Internet-Draft draft-rosen-l3vpn-mvpn-bier-01, October 2014.
[I-D.wijnands-bier-architecture] Wijnands, I., Rosen, E., Dolganow, A., Przygienda, T. and S. Aldrin, "Multicast using Bit Index Explicit Replication", Internet-Draft draft-wijnands-bier-architecture-01, October 2014.
[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997.

7.2. Informative References

[RFC4601] Fenner, B., Handley, M., Holbrook, H. and I. Kouvelas, "Protocol Independent Multicast - Sparse Mode (PIM-SM): Protocol Specification (Revised)", RFC 4601, August 2006.
[RFC5015] Handley, M., Kouvelas, I., Speakman, T. and L. Vicisano, "Bidirectional Protocol Independent Multicast (BIDIR-PIM)", RFC 5015, October 2007.
[RFC5331] Aggarwal, R., Rekhter, Y. and E. Rosen, "MPLS Upstream Label Assignment and Context-Specific Label Space", RFC 5331, August 2008.
[RFC6513] Rosen, E. and R. Aggarwal, "Multicast in MPLS/BGP IP VPNs", RFC 6513, February 2012.
[RFC7348] Mahalingam, M., Dutt, D., Duda, K., Agarwal, P., Kreeger, L., Sridhar, T., Bursell, M. and C. Wright, "Virtual eXtensible Local Area Network (VXLAN): A Framework for Overlaying Virtualized Layer 2 Networks over Layer 3 Networks", RFC 7348, August 2014.

Authors' Addresses

Nagendra Kumar Cisco 7200 Kit Creek Road Research Triangle Park, NC 27709 US EMail: naikumar@cisco.com
Rajiv Asati Cisco 7200 Kit Creek Road Research Triangle Park, NC 27709 US EMail: rajiva@cisco.com
Mach(Guoyi) Chen Huawei EMail: mach.chen@huawei.com
Xiaohu Xu Huawei EMail: xuxiaohu@huawei.com