WO2009116026A1

WO2009116026A1 - Method of forwarding ethernet packets in telecommunication networks

Info

Publication number: WO2009116026A1
Application number: PCT/IL2009/000237
Authority: WO
Inventors: Ido Ouzieli; Lior Yeheskiel
Original assignee: Eci Telecom Ltd.
Priority date: 2008-03-19
Filing date: 2009-03-02
Publication date: 2009-09-24
Also published as: IL190304A0

Abstract

A technique for forwarding Ethernet layer-2 frames in access and/or metro networks, wherein each frame carrying an overhead field OHF intended for VLAN ID. The OHF has total length of N bits and comprises a first VID sub-field (tag) and second VID sub-field (tag). The technique assumes that the two sub-fields of the OHF may have unequal sizes and therefore suggests to analyze the whole N-bit length of the OHF of each Ethernet layer-2 frame before the forwarding.

Description

Method of forwarding

Ethernet packets in telecommunication networks

Field of the invention The present invention deals with forwarding Ethernet packets for providing

Ethernet based services in various telecommunication networks, for example - in access and metro Ethernet networks.

Background of the invention Ethernet has grown from its roots in LANs to contend in previously unchartered territory of MANs and WANs. A group of projects underway in the IEEE 802 standards bodies plan to groom Ethernet with earner grade features like high availability, fault management, and resiliency thus far found only in other circuit-switched technologies. When forwarding services based on Ethernet -layer 2 in access or metro networks, two VLAN tags of an Ethernet frame are commonly used: an S- VID field (also known as S-VLAN, S-Tag) and a C-VID field (also known as C-VLAN, C-Tag). Each of the mentioned fields is used for its own purpose. According to one of the Ethernet relevant standards presently accepted in the telecommunication community (IEEE Standard for Local and metropolitan area networks Virtual Bridged Local Area Networks Amendment 4: Provider Bridges - 802.1 ad), S-VID field is intended to serve as a Service -identification VLAN tag and C-VID field serves as a Customer-identification VLAN tag (wherein VLAN is a virtual Local Area

Network). Presently, each of the two (S-VID and C-VID) fields is 12 bits long.

It has been recognized by the Inventors that the limitations, imposed by the currently standardized sizes of the S-VID and C-VID fields, cause scalability problems both in access networks (where the number of clients which might be served is limited by the C-VID field size), and in metro Ethernet and aggregation (combined) networks. These problems have been felt for quite a long time. In access networks they hamper and sometimes simply prevent the growth of network facilities, while in metro networks they also bring difficulties in proper forwarding (routing) of Ethernet packets to destination.

The difficulties in the metro networks are explained by the following facts: a) the number of services, which might be provided, is limited by the S-VID field size, and b) the number of subscribers is also limited. Presently, as defined in the standard IEEE 802. lad, recognition of the required destination of an Ethernet packet in metro networks is based on analyzing its S-VID field and its conventional MAC field, while does not include analyzing the C-VID field. The Ethernet forwarding mechanism is built on the assumption that each

Ethernet MAC address is globally unique. The MAC address uniqueness allows an Ethernet bridge/switch to automatically select its relevant port to use it for forwarding a frame to the unique destination MAC address. Unfortunately, in real world, some types of equipment do not comply with the assumption of globally unique MAC addresses. The undesired duplication of MAC addresses may arise for many reasons. One of the reasons may stem just from the fact that some equipment is manufactured in large quantities, where each specific sample is identical to another one, including the engraved MAC address of the equipment sample. Manufacturers of the equipment assume that the probability of using such samples in one networks domain is very low. There may be other reasons for duplication of MAC addresses. Changing of MAC addresses within the network may cause it; sometimes users may also contribute to creating end devices with identical MAC addresses. In case a duplicate MAC address appears in one Ethernet Broadcast Domain, forwarding of Ethernet frames becomes corrupted because some of the packets sent by the source to such an address will be forwarded to a first copy of the address, and other packets - to the second copy. Theoretically, more than two copies of MAC addresses may occur in one

Ethernet network domain.

Several technologies have been developed to resolve these problems, for example: a) MPLS has been proposed as an additional tunneling technology to be used on top of Ethernet, for resolving the scalability issue with the S-VID number. MPLS is able to tunnel 802. lad Ethernet frames from source to destination, so that two Ethernet 802. lad frames having, for example, the same S-VID, can be tunneled to their two different destinations by two different MPLS tunnels and without causing a problem. b) Mac in Mac (also known as Provider Backbone Bridge PBB) has also been proposed as an additional tunneling technology on top of Ethernet, thus resolving the scalability problem with the S-VID number. The IEEE 802.1 ah (Provider Backbone Bridges) addresses the service and MAC address scalability of provider backbone bridges. By tunneling the 802. lad frames via 802.1 ah (PBB) tunnels, a separation similar to that in the MPLS technique is achieved, and frames with identical S-VID values can reside in two different PBB tunnels.

Both of these solutions add complexity to the networking equipment and the network operation, since they require adding an additional protocol on top of the Ethernet layer 2 protocol. Object and summary of the invention

The Inventors have recognized that the problems of scalability and double addressing, when forwarding Ethernet frames in access and metro networks, can be solved in a much simple way than the approaches being presently used.

The Inventors point out that both the problem of scalability in access and metro networks, and the problem of ambiguous addressing in metro networks (when more than one identical MAC addresses exist in one network domain) can be solved by effectively utilizing and properly analyzing overhead fields of standard Ethernet packets.

For example, the Inventors point out that the growing number of customers in access networks can be properly assigned using the standard Ethernet packets, and then all the customers can be recognized when forwarding data to them. Surprisingly, it has not been provided by now though it presently seems simple to implement just by utilizing the standard

Ethernet packet overhead (as will be explained below).

The Inventors also point out that surprisingly, the problem of non-exact (ambiguous) routing of Ethernet packets in metro networks, caused by duplication of MAC addresses, has not been solved by now at the Ethernet layer, since according to the presently used techniques, the information in the C-VID field is not analyzed while sometimes it is critical for correctly selecting the required destination.

It is therefore the object of the present invention to find a simple alternative solution to the long felt problem of limited scalability and ambiguous forwarding of data, when providing Ethernet services in telecommunication networks. More particularly, the object of the invention is to resolve the problem at the layer 2 of Ethernet, without applying any additional protocol to the Ethernet layer. The above object can be achieved by providing a method of forwarding Ethernet layer-2 frames in access and/or metro networks, each frame carrying an overhead field OHF intended for VLAN ID (for Identification of Virtual Local Access Networks VLANs), wherein the OHF has a total length of N bits and comprises a first VLAN Identification (VID) subfield and a second VLAN Identification (VID) subfield; the method comprises performing at least one of the following steps before the forwarding: a) flexibly assigning said OHF, i.e. so as to comprise two said subfields having unequal sizes each being from O to N bits, while the total N-bit length of the OHF remains the same, and b) analyzing the whole N-bits length of the OHF of each Ethernet frame.

In the preferred version of the method, the first VID subfield is a Service-identification VLAN tag/subfield (known, for example, as S-VID,

S-VLAN, S-Tag) and the second VID subfield is a Customer-identification

VLAN tag/subfield (known, for example, as C-VID, C-VLAN, C-Tag). It should be kept in mind that further in the description and in the claims, the Service-identification VLAN tag/subfield may be intermittently called S-VID, S-VLAN or S-Tag which terms should be considered synonyms. Similarly, the Customer-identification VLAN tag/subfield may be intermittently called C-VID, C-VLAN or C-Tag ( to be considered synonyms).

According to the presently accepted standard IEEE 802. lad, the existing frame format comprises the OHF field of a standardized length being N=24 bits. By now, the 24 bits of the OHF are divided into a standardized C-VID subfield of 12 bits and a standardized S-VID subfield of 12 bits.

For access networks with the growing number of end customers, it is proposed to make the size of the C-VID sub-field variable; the C-VID subfield is preferably enlarged at the expense of the standardized size of the S-VID sub-field. Both of the sub-fields are analyzed, as required, at the suitable pre-programmed Ethernet enabled equipment.

For metro networks, the new method always comprises analyzing the entire OHF field including both the S-VID and the C-VID sub-fields. The

OHF may optionally comprise the S-VID sub-field of the enlarged size to cover more services required in the MAN/WAN network, or vice versa

(may comprise an enlarged C-VID subfield).

Actually, the invention allows reducing the probability of a conflict of MAC addresses, by reducing the size of the broadcast domain "formed"

(defined) by a combination of one specific value of S-VID and one specific value of C-VID.

Indeed, in the prior art situation, if two or more identical MAC addresses existed in one S-VID domain, the forwarding of Ethernet frames would be corrupted. With the proposed innovation, if these identical MAC addresses have the same S-VID value, but not the same C-VID/S-VID combination, there will be no problem in forwarding the frames to their respective addresses.

The method allows analyzing the entire field of N bits, as well as allows expanding the number of bits either in the S-VID sub-field or in the C-VID sub-field as desired, and thus to resolve the scalability and the forwarding issue in a simple and elegant manner, just with a minor modification/adjustment to the Ethernet capable equipment.

The method also comprises analyzing a field of destination MAC address DA before the forwarding.

The second aspect of the invention is providing Ethernet capable equipment pre-programmed in a manner to analyze and forward Ethernet layer-2 frames according to the proposed method.

One (first) type of the equipment is designed for assigning Ethernet layer-2 frames in access and/or metro networks, wherein each frame carries an overhead field OHF intended for VLAN ID, the OHF having a total standardized length of N bits and comprising an S-VID sub-field and a C- VID sub-field; the Ethernet capable equipment being capable of flexibly assigning said OHF field, by introducing information into the S-VID sub-field and the C-

VID sub-field in such a manner that, for example, a portion of the information, intended to be recognized as C-VID information, be allocated in the C-VID sub-field enlarged at the expense of the S-VID sub-field.

However, the reverse manner is possible where the S-VID sub-field is enlarged at the expense of the C-VID sub-field.

Another (second) type of the Ethernet capable equipment is designed for forwarding Ethernet layer-2 frames in access and/or metro networks, wherein each frame carries an overhead field OHF intended for VLAN ID, the OHF having a total standardized length of N bits and comprising a first VID sub-field and a second VID sub-field; the Ethernet capable equipment of that type being capable of analyzing the whole N-bits length of the OHF of each Ethernet layer-2 frame both in case said two sub-fields have equal sizes and in case they have unequal sizes, and capable of forwarding said frame using results of the analysis. The first and second types of the Ethernet enabled equipment are preprogrammed so as to mutually suit to one another and, preferably, are in control communication with one another directly or via a central management entity.

The proposed solution makes unnecessary such complex measures as adding an additional tunneling protocol to the Ethernet one.

The invention will be described in details as the description proceeds. Brief description of the drawings

The invention will be. further described with the aid of the following non-limiting drawings, in which: Figs. Ia and Ib illustrate one example of a standard Ethernet packet shown in the standard IEEE 802. lad as a double-tagged (Q-in-Q tagged) frame format, and the overhead fields of interest in that packet.

Figs. 2a and 2b illustrate one arrangement of an Ethernet frame for the proposed technique, presented in comparison with a conventional Ethernet frame (packet) structure that comprises two equal overhead fields S-VID and C-VID of 12 bits each. In the arrangement of Fig. 2b, the C-VID sub- field is enlarged at the cost of the S-VID sub-field.

Figs. 3a and 3b illustrate another arrangement of an Ethernet frame to be used in the proposed technique, also given in comparison with the conventional arrangement. According to Fig. 3b, the total OHF field of 24 bits is supposed to be analyzed by the inventive Ethernet enabled equipment for unambiguously forwarding the Ethernet packets.

Detailed description of the invention The solution proposed and described below is applicable in various forwarding scenarios for Ethernet environment, for example as defined by the DSLF Architecture & Transport WG TR-IOl (Work Group, Technical Recommendation- 101).

Fig. Ia illustrates an example of an Ethernet frame (marked 10) of a so-called Q-in-Q format mentioned in the IEEE-802.1ad. The following overhead fields can be seen in the standard frame 10:

DA-Destination address field; SA- Source address field; VID-VLAN ID field, Ethertype - Ethernet type/length field; FCS - Frame check sequence. In more detail: field VID 30 (marked 12) is a field carrying indication of VLAN ID according to IEEE- 802. lad; field "Type 88a8" carries 802. lad Ethertype; it is placed after the VID 30 to indicate the length and type of the preceding field ( i.e., of the VID 30); field VID 20 (marked 14) is a field carrying indication of 802. IQ VLAN ID; field Type 8100 is 802. IQ Ethertype; it is placed after the VID 20 to indicate the length and type of the preceding field (i.e., of the VID 20); Fig. Ib illustrates that the two VID fields (12 and 14) virtually form, as two sub-fields, a common VID Overhead field which is called OHF in the present patent application (VID-OHF is symbolically marked 16). Usually, the sub- field 12 is called S-VID, and the sub-field 14 is called C-VID. In the schematic template of Fig. Ib, these sub-fields are shown adjacent, though it does not have to be like that in practice. Fig. Ib is used as a reference for comparing the conventional arrangement and the arrangements proposed in Figures 2b and 3b. It should be noted that the Q-in-Q frame format is selected for the purposes of exemplary illustration and should not be understood as the only one possible Ethernet format limiting the scope of the invention. Other types of Ethernet layer-2 standard frames may comprise similar two VID overhead sub-fields, which can be assigned and analyzed according to the proposed invention.

Figs. 2a, 2b schematically show fragments of two Ethernet frames, positioned one under another for illustration. The upper illustration (Fig. 2 a) is a portion of a conventional, presently accepted Ethernet frame comprising an overhead field symbolically called OHF and comprising an S-VID (S-VLAN, S-Tag) subfield of 12 bits and a C-VID (C-VLAN, S-Tag) subfield also having length of 12 bits. The upper illustration 2a corresponds to the model (shown in Fig. Ib) of the VID-OHF of a conventional frame. The lower frame (Fig. 2b) presents a fragment of a modified Ethernet frame (having a modified assignment); it should be noted that the modified frame must be then analyzed accordingly. It is suggested that the OHF field, having the unchanged length of 24 bits and comprising the S-VID and C-VID tags (subfields), will be considered as two flexible entities, where X bits are regarded as a New- S-VID subfield, and 24-X bits are regarded as a New-C- VID subfield.

The New-S-VID, which can be longer or shorter than the 12 bits- long S-VID, will be used for the same puiposes as the old S-VID, but will support flexible combinations/values range. The same applies to the New-C- VID subfield. For the puiposes of formal definition, the new C-VID subfield may be considered equal to 0 bits, thereby the New S-VID subfield will be equal to the total OHF field (one example of a new S-VID subfield will be shown in Fig. 3). One of the forwarding functions in Ethernet environments is known as 1 : 1 Forwarding, where the one-to-one mapping correspondence exists between each specific user port and a dedicated and unique combination of two VLAN tags/subfields S-VID and C-VID (see the upper conventional frame in Fig. 2a). The S-VID tag (S-VLAN, S-Tag) is used for indicating the service-ID that is carried by the Ethernet frame, and the C-VID tag (C-VLAN, C-Tag) is used for indicating the customer- ID or ID of the end user for which the Ethernet frame is intended. Since the same service can be shared between several end-users, the S-VID is not unique in the network, and this fact simplifies addressing in a WAN network.

Since in Fig. 2a both S-VID and C-VID are 12 bits long, up to 4096 services and up to 4096 end-users can be supported by this arrangement. In the existing access networks, such a limitation (4096 end-users per DSL access multiplexer DSLAM) is acceptable. However, with the new rising technologies, such as GPON, a much higher number of end-users is expected per network-interface and per DSLAM. For example, a GPON DSLAM is expected to serve up to 40,000 end users.

While the number of end-users grows, the number of services remain fairly constant, e.g., a typical DSLAM will support 3-6 services (such as voice, data, video) However, these services can be provided by more than one application provider, and thus similar services will require different identifying S-VID values to distinguish the services of different providers.

Taking into account the above situation, the Inventors have proposed dividing the OHF field into two non-equal sub-fields: a New- S -VID and a New-C-VID. In this case, the New-S-VID will use less than 12 bits and the New-C-VID will use more than 12 bits. This way, a much higher number than 4096 end-users can be supported even by adding a single bit to the conventional C-VID subfield. Still, a great number of services provided by various providers can be identified and accurately forwarded. Fig. 2b illustrates one possible modified arrangement in the OHF (please note that the assigned numbers are used as an example only):

Six first bits of the conventional S-VID subfield will be used as the New-S- VID subfield for service indication. Using these six bits, up to 64 services can be supported (assigned, indicated and further identified for forwarding). The six remaining bits of the conventional S-VID tag will be now used as an extension of the C-VID subfield and, together with 12 bits of the conventional C-VID subfield, will result in the New-C-VID subfield intended for customer indication. Since 18 bits will be now available for the customers' indication, up to 260K customers can be supported. As can be seen, the limitation on the number of end-users is easily removed by the proposed solution, and the 1 : 1 mapping can still be used in an expanded access network.

According to Fig. 2b, the total OHF field of 24 bits is supposed to be analyzed by a suitable Ethernet enabled equipment (not shown), taking into account a pre-programmed known in advance pattern (lengths' proportion) of S-VID and C-VID sub-fields. Based on the analysis, the Ethernet packets will be then unambiguously forwarded by the equipment to clients of a communication network. The inventive Ethernet enabled forwarding equipment, of course, is accompanied with a piece of equipment responsible for suitably assigning the S-VID and C-VID sub-fields in the Ethernet frame according to the predetermined pattern. This pattern should be somehow agreed between the assigning equipment and the forwarding equipment (in advance or dynamically, by control communication).

Figs. 3a and 3b illustrate a method for another forwarding function for layer-2 Ethernet frames. Fig. 3a schematically shows a conventional Ethernet frame arrangement ( as a reference), and Fig. 3b illustrates a second modification of the Ethernet frame arrangement.

A forwarding function known as "N:l Forwarding", where many user ports are mapped to one S-VID, is known in Ethernet environments (IEEE 802. lad). In this case the forwarding is based on a learning process where the S-VID and the MAC DA (Destination Address) are analyzed to define together the User port.

However one of the problems, that the modern network topologies face, is duplicate MAC addresses, i.e., when the same MAC DA might be repeated with the same S-VID value, but with different C-VID values (or vice versa, the same MAC DA might be repeated with the same C-VID, but with different S- VID). In the approach proposed in the Provider Bridge Ethernet (802. lad), the broadcast domain is defined by either C-VID or S-VID, therefore, only one of them (according to the forwarding domain - customer or provider) is learned, and the others are ignored. In the case when duplicate MAC Addresses exist, the conventional Ethernet forwarding mechanism is broken and packets are either wrongly forwarded or lost. The Inventors propose that instead of learning the MAC DA addresses with either a conventional S-VID or a conventional C-VID, the forwarding equipment be adapted to learn the MAC DA addresses together with a New-S- VID (see Fig. 3b) which will be a combination of the conventional S-VID and C-VID (the entire 24 bits of the old S-VID and C-VID are regarded as the New-S-VID subfield).

By using the above solution, the learning will be applied to the total N bits of the OHF (called a New S-VID field), which is a combination of bits from the "old" S-VID and C-VID subfields. Therefore, even if the S-VID and MAC DA combination is repeated, the New-S-VID will be unique (as it contains the old C-VID subfield, in addition to the old S-VID).

The proposed arrangement/assignment of the Ethernet frame and, more important, the suitable analysis thereof will thus reduce probability of the wrong forwarding of Ethernet frames in case of existence of duplicate MAC addresses within one VLAN broadcast domain. With the proposed method, the probability of traffic corruption due to the duplicate addresses (which may happen although forbidden by the Ethernet protocol) will be drastically reduced. Such a reduction will almost eliminate urgent maintenance operations which are usually required when a duplicate MAC address occurs in real life networks.

According to Fig. 3b, the total OHF field of 24 bits is supposed to be analyzed, together with the DA field (see Fig. Ia), for unambiguously forwarding the Ethernet packets. This operation can be performed by a suitable piece of inventive Ethernet enabled equipment (not shown). It should be appreciated that other formats of an Ethernet layer-2 frame comprising two or more different VILAN tags/fields can be assigned and analyzed according to the present invention, and such techniques are to be considered part of the invention.

Claims

Claims:

1. A method of forwarding Ethernet layer-2 frames in access and/or metro networks, wherein each frame carrying an overhead field OHF intended for VLAN Identification (VID), the OHF having a total length of N bits and comprising a first VID sub-field and a second VID sub-field; the method comprises performing at least one of the following steps before the forwarding: a) assigning said OHF to comprise two said sub-fields having unequal sizes each being from O to N bits, while the total N-bit length of the OHF remains the same, and b) analyzing the whole N-bit length of the OHF of each Ethernet layer-2 frame.

2. The method according to Claim 1, wherein said total length is a standardized length of N=24 bits, wherein the first VID sub-field is a

Service-identification S-VID sub-field, the second VID sub-field is a Customer-identification C-VID sub-field, and wherein a standardized size of both the S-VID and C-VID sub-fields is 12 bits.

3. The method according to Claim 1 or 2, adapted for access networks, wherein at step a) size of the second VID sub-field is enlarged at the expense of a standardized size of the first VID sub-field, and wherein both of the steps a) and b) are performed.

4. The method according to Claim 1 or 2, adapted for metro networks, wherein only step b) is performed.

5. The method according to Claim I₅ comprising analyzing a field of destination MAC address DA of each Ethernet frame before the forwarding.

6. An Ethernet capable equipment pre-programmed in a manner to assign, and/or analyze and forward Ethernet layer-2 frames according to the method of Claim 1.

7. An Ethernet capable equipment for assigning Ethernet layer-2 frames to be forwarded in access and/or metro networks, wherein each frame carries an overhead field OHF intended for VLAN Identification (VID), the OHF having a total length of N bits and comprising a Service-identification S- VID sub-field and a Customer-identification C-VID sub-field; the Ethernet capable equipment being capable of flexibly assigning said OHF field by introducing information into the S-VID sub-field and the C- VID sub-field in such a manner that a portion of the information, intended to be recognized as C-VID information, be allocated in the C-VID sub-field enlarged at the expense of the S-VID sub-field.

8. An Ethernet capable equipment for forwarding Ethernet layer-2 frames in access and/or metro networks, wherein each frame carries an overhead field OHF intended for VLAN ID, the OHF having a total length of N bits and comprising a first VID sub-field and a second VID sub-field; the Ethernet capable equipment being adapted to analyze the Whole N-bits length of the OHF of each Ethernet layer-2 frame both in case said two sub- fields have equal sizes and in case they have unequal sizes, and capable of forwarding said frame using results of the analysis.