idnits 2.17.1 draft-dunbar-trill-server-assisted-edge-00.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** You're using the IETF Trust Provisions' Section 6.b License Notice from 12 Sep 2009 rather than the newer Notice from 28 Dec 2009. (See https://trustee.ietf.org/license-info/) Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack separate sections for Informative/Normative References. All references will be assumed normative when checking for downward references. ** There are 6 instances of too long lines in the document, the longest one being 62 characters in excess of 72. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == The document doesn't use any RFC 2119 keywords, yet seems to have RFC 2119 boilerplate text. -- The document date (March 7, 2011) is 4798 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Unused Reference: 'ARMD-Problem' is defined on line 208, but no explicit reference was found in the text -- Possible downref: Non-RFC (?) normative reference: ref. 'ARMD-Problem' Summary: 3 errors (**), 0 flaws (~~), 3 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 1 TRILL working group L. Dunbar 2 Internet Draft Huawei 3 Intended status: Standard Track 4 Expires: Sept 2011 5 March 7, 2011 7 Directory Server Assisted TRILL edge 8 draft-dunbar-trill-server-assisted-edge-00.txt 10 Status of this Memo 12 This Internet-Draft is submitted to IETF in full conformance with the 13 provisions of BCP 78 and BCP 79. 15 Internet-Drafts are working documents of the Internet Engineering 16 Task Force (IETF), its areas, and its working groups. Note that 17 other groups may also distribute working documents as Internet- 18 Drafts. 20 Internet-Drafts are draft documents valid for a maximum of six months 21 and may be updated, replaced, or obsoleted by other documents at any 22 time. It is inappropriate to use Internet-Drafts as reference 23 material or to cite them other than as "work in progress." 25 The list of current Internet-Drafts can be accessed at 26 http://www.ietf.org/ietf/1id-abstracts.txt 28 The list of Internet-Draft Shadow Directories can be accessed at 29 http://www.ietf.org/shadow.html 31 This Internet-Draft will expire on September 7, 2011. 33 Copyright Notice 35 Copyright (c) 2011 IETF Trust and the persons identified as the 36 document authors. All rights reserved. 38 This document is subject to BCP 78 and the IETF Trust's Legal 39 Provisions Relating to IETF Documents 40 (http://trustee.ietf.org/license-info) in effect on the date of 41 publication of this document. Please review these documents 42 carefully, as they describe your rights and restrictions with respect 43 to this document. Code Components extracted from this document must 44 include Simplified BSD License text as described in Section 4.e of 45 the Trust Legal Provisions and are provided without warranty as 46 described in the BSD License. 48 Abstract 50 TRILL edge nodes currently learn the mapping between MAC address and 51 its corresponding TRILL edge node address by observing the data 52 packets traversed through. 54 This document describes why and how directory based server(s) can 55 optimize TRILL network in data center environment. 57 Conventions used in this document 59 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 60 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 61 document are to be interpreted as described in RFC-2119 0. 63 Table of Contents 65 1. Introduction ................................................ 2 66 2. Terminology ................................................. 3 67 3. Impact to TRILL by massive number of hosts ................... 3 68 4. Directory Server for TRILL in Data Center environment. ........ 4 69 5. Conclusion and Recommendation ................................ 5 70 6. Manageability Considerations ................................. 5 71 7. Security Considerations ...................................... 5 72 8. IANA Considerations ......................................... 5 73 9. Acknowledgments ............................................. 5 74 10. References ................................................. 5 75 Authors' Addresses ............................................. 6 76 Intellectual Property Statement ................................. 6 77 Disclaimer of Validity ......................................... 6 79 1. Introduction 81 Data center networks are different from campus networks in several 82 ways. Main differences include: 83 VM (host) to server assignment is done by Server (or VM) 84 Manager, which means that the host location is arranged by 85 management system(s). 86 Topology is based on racks and rows; 87 There could be massive number of virtual machines (hosts), but 88 relatively small number of switches. 90 This draft describes why Data Center TRILL networks can be optimized 91 by utilizing directory server based approach. 92 2. Terminology 94 Bridge: IEEE802.1Q compliant device. In this draft, Bridge is used 95 interchangeably with Layer 2 switch. 97 DC: Data Center 99 EOR: End of Row switches in data center. 101 FDB: Filtering Database for Bridge or Layer 2 switch 103 ToR: Top of Rack Switch. It is also known as access switch. 105 VM: Virtual Machines 107 3. Impact to TRILL by massive number of hosts 109 In a virtualized data center, a VM may be placed on any physical 110 server. A variety of algorithms can be applied to select the location 111 of a VM. Resource aware algorithms (e.g. energy, bandwidth, etc,) 112 will use a placement that satisfies the processing requirements of 113 each VM but require the minimal number of physical servers and 114 switching devices. 115 With this, and similar types of assignment algorithm, subnets tend to 116 extend throughout the network. When this happens, the broadcast 117 messages within each subnet will be flooded across the TRILL domain, 118 which not only consumes a lot of bandwidth on links in TRILL domain, 119 but also causes a TRILL edge port to learn all the hosts belonging to 120 all the subnets which are enabled on the port. Even though a TRILL 121 edge port is only supposed to learn the entries which communicate 122 with hosts underneath, the frequent ARP/ND from all hosts within each 123 subnet will always refresh the TRILL edge node's MAC<->TRILL-Edge 124 mapping table. 125 Consider a data center with 80 rows, 8 racks per row and 40 servers 126 per rack. There can be 80*8*40=25600 servers. Suppose each server is 127 virtualized to 20 VMs, there could be 25600*20=512000 hosts in this 128 data center. 129 Let's consider a case that the TRILL edge starts at an Ingress port 130 of a TOR switch. Assuming there are 5 different VLANs enabled on the 131 TRILL Ingress port (i.e. the 20 VMs in one server belong to 5 132 different VLANs) and each VLAN has 200 hosts, then the TRILL edge 133 port has to learn 5*200=1000 MAC&VLAN entries. Since there are 40 134 ports on the TOR, the total number of MAC&VLAN entries for the TOR 135 switch is 1000*40= 40000. Under this scenario, there will be 25600 136 entries in the TRILL routing domain if protection is not considered. 137 When protection is considered, the number of ports in TRILL domain 138 will double. That may be too many nodes for the IS/IS routing domain. 139 Let's consider another case of TRILL edge starting at the End of Row 140 switches. With the same assumption as before, there are 40*20 = 800 141 hosts to attached to each port of an EoR switch and 8*800=6400 hosts 142 attached to an EoR switch. If all those 6400 hosts belong to 640 143 VLANs and each VLAN has 200 hosts, then the total number of MAC&VLAN 144 entries to be learned by the TRILL edge (i.e. EoR) = 640*200=128000. 145 Under this scenario, there will be 80*8 = 640 EoR ports in the TRILL 146 routing domain when protection is not considered and 1280 EoR ports 147 when protection is considered. However, the number of MAC&VLAN 148 entries to be learnt by the TRILL edge node is very large. 150 4. Directory Server for TRILL in Data Center environment. 152 As described in the Section 1, the VM placement to server/rack is 153 orchestrated by Server (or VM) Management System(s). Therefore, there 154 is a central location with the information on where each VM is 155 placed. So it is relatively reliable to build a centralized (or 156 distributed) directory server(s) who has the knowledge on where each 157 VM is placed. 159 Here can be a procedure for TRILL edge node to utilize a Directory 160 Server 162 TRILL edge node can simply intercept all ARP requests and 163 forward them to the Directory Server, 165 The reply from the Directory Server can be the standard ARP 166 reply with an extra field showing the TRILL egress node address 168 TRILL ingress node can cache the mapping 170 If TRILL edge node receives an unknown MAC-DA, it simply 171 forwards the packet to the directory server. The directory 172 server can simply drop the frame if it doesn't have the 173 information, or forward the frame to the correct egress node and 174 send down a new mapping to the ingress Trill edge node. 176 Another approach is for Directory Server to pass down the MAC&VLAN 177 mapping for all the hosts belonging to all the VLANs enabled on the 178 TRILL edge port. 180 5. Conclusion and Recommendation 182 The traditional TRILL learning approach of observing data plane can 183 no longer keep pace with the ever growing number of hosts in Data 184 center. 186 Therefore, we suggest TRILL to consider directory assisted 187 approach(es). This draft only introduces the basic concept of using 188 directory assisted approach for TRILL edge nodes to learn the MAC to 189 TRILL mapping. We want to get some working group consensus before 190 drilling down to detailed steps required for the approach. 192 6. Manageability Considerations 194 This document does not add additional manageability considerations. 196 7. Security Considerations 198 This document has no additional requirement for security. 200 8. IANA Considerations 202 9. Acknowledgments 204 This document was prepared using 2-Word-v2.0.template.dot. 206 10. References 208 [ARMD-Problem] Dunbar, et,al, "Address Resolution for Large Data 209 Center Problem Statement", Oct 2010. 211 [ARP reduction] Shah, et. al., "ARP Broadcast Reduction for Large Data 212 Centers", Oct 2010 214 Authors' Addresses 216 Linda Dunbar 217 Huawei Technologies 218 1700 Alma Drive, Suite 500 219 Plano, TX 75075, USA 220 Phone: (972) 543 5849 221 Email: ldunbar@huawei.com 223 Intellectual Property Statement 225 The IETF Trust takes no position regarding the validity or scope of 226 any Intellectual Property Rights or other rights that might be 227 claimed to pertain to the implementation or use of the technology 228 described in any IETF Document or the extent to which any license 229 under such rights might or might not be available; nor does it 230 represent that it has made any independent effort to identify any 231 such rights. 233 Copies of Intellectual Property disclosures made to the IETF 234 Secretariat and any assurances of licenses to be made available, or 235 the result of an attempt made to obtain a general license or 236 permission for the use of such proprietary rights by implementers or 237 users of this specification can be obtained from the IETF on-line IPR 238 repository at http://www.ietf.org/ipr 240 The IETF invites any interested party to bring to its attention any 241 copyrights, patents or patent applications, or other proprietary 242 rights that may cover technology that may be required to implement 243 any standard or specification contained in an IETF Document. Please 244 address the information to the IETF at ietf-ipr@ietf.org. 246 Disclaimer of Validity 248 All IETF Documents and the information contained therein are provided 249 on an "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE 250 REPRESENTS OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE 251 IETF TRUST AND THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL 252 WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY 253 WARRANTY THAT THE USE OF THE INFORMATION THEREIN WILL NOT INFRINGE 254 ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS 255 FOR A PARTICULAR PURPOSE. 257 Acknowledgment 259 Funding for the RFC Editor function is currently provided by the 260 Internet Society.