Skip to main content

A Framework for Semantic IPv6 Prefix and Gap Analysis
draft-jiang-semantic-prefix-03

The information below is for an old version of the document.
Document Type
This is an older version of an Internet-Draft whose latest revision state is "Expired".
Authors Sheng Jiang , Qiong Sun
Last updated 2012-10-21
RFC stream (None)
Formats
Additional resources
Stream Stream state (No stream defined)
Consensus boilerplate Unknown
RFC Editor Note (None)
IESG IESG state I-D Exists
Telechat date (None)
Responsible AD (None)
Send notices to (None)
draft-jiang-semantic-prefix-03
DHC Working Group                                             S. Jiang 
Internet Draft                            Huawei Technologies Co., Ltd 
Intended status: Informational                                  Q. Sun 
Expires: April 23, 2013                                  China Telecom 
                                                      October 22, 2012 
                                    
          A Framework for Semantic IPv6 Prefix and Gap Analysis 
                   draft-jiang-semantic-prefix-03.txt 

Status of this Memo 

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

   Internet-Drafts are working documents of the Internet Engineering 
   Task Force (IETF), its areas, and its working groups. Note that other 
   groups may also distribute working documents as Internet-Drafts. 

   Internet-Drafts are draft documents valid for a maximum of six months 
   and may be updated, replaced, or obsoleted by other documents at any 
   time. It is inappropriate to use Internet-Drafts as reference 
   material or to cite them other than as "work in progress." 

   The list of current Internet-Drafts can be accessed at 
        http://www.ietf.org/ietf/1id-abstracts.txt 

   The list of Internet-Draft Shadow Directories can be accessed at 
        http://www.ietf.org/shadow.html 

   This Internet-Draft will expire on April 23, 2013. 

    

Copyright Notice 

   Copyright (c) 2012 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. 

 
 
 
S. Jiang & Q. Sun      Expires April 23, 2013                 [Page 1] 


Internet-Draft     draft-jiang-semantic-prefix-03         October 2012 
 
 

Abstract 

   Some Internet Service Providers and enterprises desire to be aware of 
   more information about each packet, so that packets can be treated 
   differently and efficiently. Packet-level differentiating can also 
   enable flow-level and user-level differentiating. 

   IPv6, with a large address space, allows semantics to be embedded 
   into addresses. Routers can easily apply relevant operations 
   accordingly. 

   This document describes a framework that embeds semantics into IPv6 
   prefixes, so that network devices can treat packets based on these 
   explicit semantics. This document also analyzes on the technical gaps 
   for embedding complex semantics. 

   This informational document only discusses usage of semantics in a 
   Semantic Prefix Domain. It does NOT intent or suggest to standardize 
   any common global semantics. 

Table of Contents 

   1. Introduction ................................................ 3 
   2. Terminology ................................................. 4 
   3. Why Prefix .................................................. 4 
   4. The Semantic Prefix Domain .................................. 5 
   5. The Embedded Semantics ...................................... 6 
   6. Applicability ............................................... 6 
      6.1. An ISP semantic prefix example ......................... 6 
      6.2. An enterprise semantic prefix example .................. 7 
   7. Benefits .................................................... 8 
   8. Gaps ........................................................ 8 
      8.1. Semantics management in networks ....................... 8 
      8.2. Semantic relevant interactions with host ............... 9 
   9. Security Considerations .................................... 10 
   10. IANA Considerations........................................ 10 
   11. Change log ................................................ 10 
   12. References ................................................ 10 
      12.1. Normative References ................................. 10 
      12.2. Informative References ............................... 11 
    

 
 
S. Jiang & Q. Sun      Expires April 23, 2013                 [Page 2] 



Internet-Draft     draft-jiang-semantic-prefix-03         October 2012 
 
1. Introduction 

   While the global Internet increases explosively, more and more 
   differentiated requirements are raised for the packet delivery of 
   networks. Internet Service Providers and enterprises desire to be 
   aware of more information about each packet, such as 
   destination/source location, user types, service types, applications, 
   security requirements, traffic identity types, quality requirements, 
   etc. Based on these informations, network operators could treat 
   packets differently and efficiently. Packet-level differentiating can 
   also enable flow-level and user-level differentiating. 

   However, except for destination/source location, almost of 
   abovementioned information is not expressed explicitly. Hence, it is 
   difficult for network operators to identify. 

   Two passive and indirect technologies are already developed to 
   distinguish the packets. Deep Packet Inspection (DPI) has been used 
   by ISPs to learn the characters of packets. But DPI is expensive for 
   both operational costs and process latency. Its time delay is too 
   much to be able to be used for real time traffic control. Overlay 
   networks are constructed in order to permit routing of packets to 
   destinations not specified by IP addresses. But still, the overlay 
   has no control over how packets are routed in the underlying network 
   between two overlay nodes. Although tunnel or label forwarding may 
   operate the traffic path, they introduce extra overhead while they 
   depend on indirect information sources. 

   An initiative solution, Quality of Service (QoS) and DiffServ 
   [RFC2474] was also developed. It specifies a simple, scalable and 
   coarse-grained mechanism for classifying and managing network  
   traffic. However, the DiffServ fields set by the packet senders are 
   not trustable by the network operators. In the real user case, ISPs 
   deploy "remarking" points at the edge network, which classify each 
   received packet and rewrite its DiffServ field according to user 
   information learned from AAA or VLAN. 

   The abovementioned solutions are mainly developed in IPv4 era, in 
   which IP address is only locator, nothing else, giving the limited 
   space. Although DiffServ was developed identically for IPv4 and IPv6, 
   it inherits the same limitation. 

   IPv6 has broken such limitation with its very large address space. It 
   allows certain semantics to be embedded into addresses. ISPs or 
   Enterprises can proactively embed pre-defined information into 
   addresses so that intermediate devices can easily apply relevant 
   operations on packet since addresses are the most explicit element in 
   a packet. It provides an easy access and trustable fundamental for 
   packet differentiated treatment. The technical fact that IPv6 allow 
   multiple addresses on a single interface also provides precondition 
 
 
S. Jiang & Q. Sun      Expires April 23, 2013                 [Page 3] 



Internet-Draft     draft-jiang-semantic-prefix-03         October 2012 
 
   for the approach that user chooses addresses differently for 
   different purposes/usages. 

   This document describes a framework that embeds semantics into IPv6 
   prefixes, so that network devices can treat packets based on these 
   explicit semantics. This approach diverts much network complexity to 
   the planning and management of IPv6 address and IP address based 
   policies. It indeed simplifies the management of ISP networks. 

   Different networks may have very different choose for the most 
   important semantics. Pre-defined semantic definitions are only 
   meaningful locally. Therefore, standardizing a general semantic is 
   almost an impossible job. This informational document only discusses 
   usage of semantics in a Semantic Prefix Domain. It does NOT intent or 
   suggest to standardize any common global semantics. 

   This document also analyzes the technical gaps to maximum the 
   benefits of semantics prefix approach, for which complex semantics 
   may need to be embedded. For now, this document only discusses 
   unicast address within IPv6 Addressing Architecture [RFC4291]. 

2. Terminology 

   Semantic Prefix: a flexible-length IPv6 prefix that was embedded 
   certain semantics. 

   Semantic Prefix Domain: a portion of the Internet over which a 
   consistent set of semantic-prefix-based policies are administered in 
   a coordinated fashion. 

3. Why Prefix 

   Although interface identifier of IPv6 address has arbitrary bits and 
   extension header can carry much more information, they are not 
   trustable by network operators. Selfish users may easily change the 
   setting of interface identifier or extension header in order to 
   obtain undeserved priorities/privileges, while servers or enterprise 
   users may be much more self-restricted since they are charged 
   accordingly. 

   Prefix is almost the only thing a network operator can trust in an IP 
   packet because it is delegated by the network and the network can 
   detect any undesired modifications, then, filter the packet. If one 
   gets the destination address wrong, the packet would not reach; if it 
   gets the source address wrong, the return packet would not arrive. 
   This also would allow enterprise semantics to be able to traverse ISP 
   networks. 

   The prefix concept here refers the most left bits in IP addresses 
   that are delegated by the network management plane. It could be 
 
 
S. Jiang & Q. Sun      Expires April 23, 2013                 [Page 4] 



Internet-Draft     draft-jiang-semantic-prefix-03         October 2012 
 
   longer than 64, if the network operators strictly manage the address 
   assignment by using Dynamic Host Configuration Protocol for IPv6 
   (DHCPv6) [RFC3315] (but in this case standard Stateless Address 
   Autoconfiguration - SLACC [RFC4862] cannot be used). 

   Although IPv6 address space is plentiful, it should not be wasted. 
   This argument can be dealt with by ensuring that only a small number 
   of traffic classes are identified within a given user's traffic, so 
   only a few bits in the prefix are needed.  

4. The Semantic Prefix Domain 

   A Semantic Prefix Domain, analagous to a Differentiated Services 
   Domain [RFC2474], is a portion of the Internet over which a 
   consistent set of semantic-prefix-based policies are administered in 
   a coordinated fashion. A Semantic Prefix Domain can represent 
   different administrative domains or autonomous systems, different 
   trust regions, different network technologies, hosts and routers, 
   different user groups, different services, different traffic groups, 
   different applications, etc. An enterprise Semantic Prefix Domain may 
   span several physical networks, traversing ISP networks. 

   The selections of semantics are various among different Semantic 
   Prefix Domains. Network operators should choose semantics according 
   to their needs for network management and services management. If an 
   ISP has several discontinuous address blocks, it may be organized as 
   a single Semantic Prefix Domain if the same semantic definition 
   shared among these discontinuous address blocks. If these address 
   blocks have different prefix lengths, their Semantic Prefix Domains 
   may be distinguished each other by minimum differences of semantic 
   definition. This indeed introduces the hierarchical Semantic Prefix 
   Domains. A big Semantic Prefix Domain, which has common semantics, 
   may contain several sub Semantic Prefix Domains, each of which has 
   their own additional semantics. The design of the hierarchical 
   Semantic Prefix Domains can also be used to organize and manage logic 
   subnets. 

   A Semantic Prefix Domain has a set of pre-defined semantic 
   definitions, which are only meaningful locally. Without an efficient 
   semantics notification or exchanging mechanism or service agreement, 
   the definitions of semantics are only meaningful within local 
   Semantic Prefix Domain. The semantics notification or exchanging does 
   not have to through protocols. Manual interactions between network 
   operators may also work out. However, this may involve trust models 
   among network operators. 

   Sharing semantic definition among Semantic Prefix Domains enables 
   more semantic based network operations. 

 
 
S. Jiang & Q. Sun      Expires April 23, 2013                 [Page 5] 



Internet-Draft     draft-jiang-semantic-prefix-03         October 2012 
 
5. The Embedded Semantics 

   As mentioned in Section 1, much information regarding to packets is 
   useful for network operators. But, the prefix bits that can be used 
   for embedded semantics are very limited. Therefore, only the  
   selected, most useful semantics can be embedded in the prefix. Note, 
   however, that DiffServ provides a very rich QoS semantic with only 6 
   bits. The available bits increase largely in the strictly managed 
   network by DHCPv6.  

   The following are some semantics may be useful by network operators 
   beside source/destination location: user types, service types, 
   applications, security requirements, traffic identity types, quality 
   requirements, etc. When used, all of them should be restricted in a 
   highly abstracted way. Larger granularity of semantics may provide 
   better aggregation and extensibility. This document does not intend 
   to define or give recommendations on choose of semantics for 
   embedding in prefix. 

   In a given Semantic Prefix Domain, multiple semantics can be used 
   combinatorially. To use the limited bits efficiently, bits semantics 
   should be pre-defined very carefully. The network operators should be 
   very careful to plan and manage the semantic field. The network 
   operators should self-restrict NOT to put too many semantic into 
   prefixes, in order to avoid to be trapped into very complicated 
   management issues. Too many semantics make management for prefix 
   delegation become very complicated and hosts would not be able to 
   handle. 

   An important principle is to avoid semantic overlap for packet though 
   semantic overlap for devices/hosts is fine. Any potential scenarios 
   that a given packet may be mapped two or more semantic prefixes are 
   considered harmful. 

6. Applicability 

6.1. An ISP semantic prefix example 

   The current ISP network is mainly aggregated according to locator. 
   The below ISP semantic prefix example uses the most left bits of 
   prefix for locator function and lower bits for semantics. In other 
   scenarios, if the network operator would like to organize network 
   aggregation by semantic prior, using higher bits for semantics is 
   also possible. Mixed aggregation model can be reached by put 
   semantics or part of semantics bits in the middle of locator bits. 

 
 
S. Jiang & Q. Sun      Expires April 23, 2013                 [Page 6] 



Internet-Draft     draft-jiang-semantic-prefix-03         October 2012 
 
    0                   1                   2                   3 
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 
   |           IANA assigned block         |      locator          | 
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 
   |        locator (Cont.)        | Semantic Field|Subscriber bits| 
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 

                  Figure 1: An ISP semantic prefix example 

   In this example, the Semantic Prefix Domain has a /20 IPv6 address 
   space. The 28 most-left (roughly 26 million of /64 prefixes) bits are 
   allocated as locator. It serves network aggregation of topology  
   based. The 8 most-right bits are subscriber bits. It means /56 prefix 
   is assigned to subscribers. 8 bits (from bit 44 to 51) are assigned 
   as semantic field. 

6.2. An enterprise semantic prefix example 

   The below enterprise semantic prefix example also uses the most left 
   bits of prefix for locator function and lower bits for semantics. 
   However, the locator function of IP address in enterprise networks 
   may not be as important as in ISP networks. The enterprise network 
   operator may prefer to organize network by semantic prior. 

   A multiple-site enterprise may receive several prefixes that have 
   different lengths. The semantic bits should be based on the longest 
   prefix. The shorter prefix can use extra available bits for locators. 
   It is compatible that shorter prefixes serve bigger networks with 
   more users. 

    0                   1                   2                   3 
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 
   |                ISP assigned block                             | 
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 
   |  ISP assigned block   |       Locator         | Semantic Field| 
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 

              Figure 2: An enterprise semantic prefix example 

   In this example, an enterprise have received a 38/ address block for 
   one site A and a /44 for another site B (the semantic prefix shown in 
   Figure 2). They can be organized in a same Semantic Prefix Domain. 
   The most-left 18 (site A) / 12 (site B, as shown in Figure 2) bits 
   are allocated as locator. It serves network aggregation of topology 
   based. The most-right 8 bits (from bit 56 to 63) are assigned as 
   semantic field. 

 
 
S. Jiang & Q. Sun      Expires April 23, 2013                 [Page 7] 



Internet-Draft     draft-jiang-semantic-prefix-03         October 2012 
 
7. Benefits 

   This section presents some, definitely not all, benefits. Depending 
   on embedded semantics, various beneficial scenarios can be expected. 

   - Easy measurement and statistic 

   The semantic prefix provides explicit identifiers for measurement and 
   statistic. They are as simple as checking certain bits of address in 
   each packets. 

   - Easy flow control 

   By applying policies according to certain bit value, it is easy to 
   control packets that have the same semantics. 

   - Policy aggregation 

   The semantic prefix allows many policies to be aggregated according 
   to the same semantics in the policy based routing system [RFC1104]. 

   - Easy dynamic changing for semantic oriented policy  

   The network operators may want to dynamically change the policy 
   actions that are operated on certain semantic packets. The semantic 
   prefix allows such changes be operated easily.  

   - Application-aware routing 

   Embedding application information into IP addresses is the simplest 
   way to realize application aware routing. 

8. Gaps 

   The simplest model of the semantic prefix is only embedded abstracted 
   user type semantic into the prefix. It can be supported with the 
   current network architecture because each subscriber is still 
   assigned one prefix, while they are not notified the semantic within 
   it.  

   In order to maximum the benefits from the semantic prefix design, 
   more functions are needed to allow semantic relevant operations in 
   networks and semantic relevant interactions with hosts. 

8.1. Semantic relevant operations in networks 

   In order to manage semantic prefixes and their relevant network 
   actions, the network should be able to perform semantic relevant 
   operations. 

 
 
S. Jiang & Q. Sun      Expires April 23, 2013                 [Page 8] 



Internet-Draft     draft-jiang-semantic-prefix-03         October 2012 
 
   - Semantics notification within the managed network 

   When DHCPv6-PD [RFC3633] delegates a prefix, the associated semantics 
   should be bounded. This is particular useful for autonomic process 
   when a new device is plugged in. 

   - Policy dynamic changing 

   New functions or protocol extension are needed to enable dynamically 
   changing the policy actions that are operated on certain semantic 
   packets. 

   - Semantics announcement to peer networks 

   Networks may want to announce some of its prefix semantics to their 
   peer networks. Hence, their peer networks may take some correspondent 
   operations. 

8.2. Semantic relevant interactions with hosts 

   The more semantics embedded into prefix, the more complicated 
   functions are needed for semantic relevant interactions between hosts 
   and the network, such as prefix delegation, host notification and 
   address selections, etc. 

   - Notify prefix semantics to hosts 

   When a host connects to network, it should be assign a prefix. The 
   associated semantics should also be notified, also if there are 
   semantic relevant rules. 

   - Address selection according to semantics on hosts 

   In practice, a host may belong to several semantics. It means several 
   IPv6 addresses are available on a single physical interface. A 
   certain packet would only serve a certain semantic. The IPv6 stack on 
   that host must know and understand these semantics and its 
   correspondent bits in order to choose right source address when 
   forming a packet. If the embedded semantic is application relevant, 
   applications on the hosts should also be involved in the address 
   choosing process: the host IPv6 stack reports multiple available 
   addresses to the application through socket API (one example is "IPv6 
   Socket API for Source Address Selection" [RFC5014]). Then the 
   application responses the one it attached. 

   In this architecture, hosts have to be intelligent enough to choose 
   its source address according to its given information. In some 
   complicated scenarios, choosing destination address may also need 
   further supporting functions. 

 
 
S. Jiang & Q. Sun      Expires April 23, 2013                 [Page 9] 



Internet-Draft     draft-jiang-semantic-prefix-03         October 2012 
 
   The current address selection algorithms and address selection API 
   [RFC5014] are too simple to support this architecture. More 
   complicated functions and intelligence are needed.  

9. Security Considerations 

   Embedding semantics in prefix is actually exposing more information 
   of packets explicit. These informations may also provide convenient 
   for malicious attackers to track or attack certain type of packets. 
   When networks announce their local prefix semantics to their peer 
   networks, it may increase the vulnerable risk. 

10. IANA Considerations 

   This document has no IANA considerations. 

11. Change log 

      draft-jiang-semantic-prefix-03: add the concept of hierarchical 
   Semantic Prefix Domain and more gap analysis, 2012-10-22. 

      draft-jiang-semantic-prefix-02: removed detailed examples and 
   recommendations for semantics bits, 2012-10-22. 

      draft-jiang-semantic-prefix-01: added enterprise considerations 
   and scenarios, emphasizing semantics only for local meaning and no 
   intend to standardize any common global semantics, 2012-07-16 

      draft-jiang-semantic-prefix-00: original version, 2012-07-09 

12. References 

12.1. Normative References 

   [RFC1104] H.W. Braun, "Models of policy based routing", RFC 1104, 
             June 1989. 

   [RFC2474] K. Nichols, S. Blake, F. Baker, and D. Black, "Definition 
             of the Differentiated Services Field (DS Field) in the IPv4 
             and IPv6 Headers", RFC 2474, December 1998 

   [RFC3315] R. Droms, et al., "Dynamic Host Configure Protocol for 
             IPv6", RFC 3315, July 2003. 

   [RFC3633] O. Troan, and R. Droms, "IPv6 Prefix Options for Dynamic 
             Host Configuration Protocol (DHCP) version 6", RFC 3633, 
             December 2003. 

   [RFC4862] S. Thomson, T. Narten, and T. Jinmei, "IPv6 Stateless 
             Address Autoconfiguration", RFC 4862, September 2007. 
 
 
S. Jiang & Q. Sun      Expires April 23, 2013                [Page 10] 



Internet-Draft     draft-jiang-semantic-prefix-03         October 2012 
 
   [RFC4291] R. Hinden, and S. Deering, "IP Version 6 Addressing 
             Architecture", RFC4291, February 2006. 

12.2. Informative References 

   [RFC5014] E. Nordmark, S. Chakrabarti, J. Laganier, "IPv6 Socket API 
             for Source Address Selection", RFC 5014, September 2007. 

    

   Author's Addresses 

   Sheng Jiang 
   Huawei Technologies Co., Ltd 
   Q14, Huawei Campus 
   No.156 Beiqing Road 
   Hai-Dian District, Beijing  100095 
   P.R. China 
   EMail: jiangsheng@huawei.com 
    
   Qiong Sun 
   China Telecom 
   Room 708, No.118, Xizhimennei Street 
   Beijing  100084 
   P.R. China 
   EMail: sunqiong@ctbri.com.cn 

 
 
S. Jiang & Q. Sun      Expires April 23, 2013                [Page 11]