Ethernet Academy Forum Syndication http://www.ethernetacademy.net Sat, 25 May 2013 04:29:20 -0500 Kunena 1.0.8 http://www.ethernetacademy.net/index.php/MEF-Professional-Certification/1102-MEF-Certification.html#1102 Hi i am based in New Delhi and i want to do MEF certification course. Do you have any course program or any of your franchise running a course program over here in New Delhi, India. Fri, 03 May 2013 15:09:41 -0500 http://www.ethernetacademy.net/index.php/Mobile-Backhaul/1093-Single-CoS-versus-Multi-CoS.html#1101 Hi, which attribute as MEF 10.2, would be needed to setup a service as Single CoS verus Multi CoS. I checked the list of attributes but did not find one that clearly defines single verus mult Cos option. Thu, 25 Apr 2013 23:42:04 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/1088-Status-of-Ethernet-Service-Latching-Loopback.html#1100 This group defines loopback per port per VLAN (and per specific Source Address). The work is on going and the plan is to reach letter ballot at Q4 2013. Tue, 23 Apr 2013 06:15:23 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/1089-MEF-Requirements-for-Reflection-Support.html#1099 There is a project in the MEF that is called Latching Loopback which defines exactly such loopbacks per VLAN. If you are a MEF member than you can see drafts in the Latching Loopback folder. Dudu Tue, 23 Apr 2013 06:11:34 -0500 http://www.ethernetacademy.net/index.php/Cloud-Networks/874-Cloud-Networks-Forum/Page-2.html#1098 I think for "cloud stuff" the most interesting technologies in the future could be some of those: * DMVPN * PBB-TE I'm not sure if TRILL will ever take hold. Many "upper end" providers in fact use infiniband, which is not used for cross-datacenter connections, but has high reliability. These providers must be looking for something equal on the ethernet - metro/wide area - side of things. What I'm thinking they might want is something that can also replace the current use of DWDM datacenter links. What always puzzles me is the horribly lagging adoption of most related technologies in Linux upstream. Providers are tied to unsupported, custom patches for many core aspects of their networks because the community seems to dislike advanced networking protocols. (example: s-vlan, qinq with different ethertype, rstp, mpls, ethernet/ip over infiniband natting, gvrp) Greetings, Florian Sun, 21 Apr 2013 13:15:22 -0500 http://www.ethernetacademy.net/index.php/Mobile-Backhaul/1093-Single-CoS-versus-Multi-CoS.html#1097 A single-CoS EVC means that the Operator / Service provider sees only a single CoS, hence the question is irrelevant. Any customer marking is not relevant. You can't eat the cake and have it full  If you want congestion management, go for multiple-CoS EVC. My 2 cents, Ayal Wed, 17 Apr 2013 14:44:28 -0500 http://www.ethernetacademy.net/index.php/Mobile-Backhaul/1093-Single-CoS-versus-Multi-CoS.html#1096 Thanks, what is the default behaviour of SP to single CoS traffic that has mix of traffic with high to low priority having different P bit values. At the time of congestion ? is low priority traffic dropped. Does this behaviour need to be set by SP or this is the defualt treatment? Wed, 17 Apr 2013 13:04:54 -0500 http://www.ethernetacademy.net/index.php/Mobile-Backhaul/1093-Single-CoS-versus-Multi-CoS.html#1095 Hi, If the SP offers a single-CoS EVC and this EVC served the customer's traffic based on CE-VLAN IDs (without looking at PCP bits or DSCP bits) then the customer's Pbits are not relevant to the CEN. They may have significance to the CE "on the other side of the CEN", but this is outside the scope of Carrier Ethernet. My 2 cents, Ayal Wed, 17 Apr 2013 06:54:09 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/1091-VPLS-Loop-Protection-Requirement-Internet-Draft.html#1094 Thank you for taking the time to read the draft and comment. I appreciate the feedback very much. You are correct that in scenario 2 the customers own deployment of a traditional layer 2 loop prevention protocol should take action to stop the loop however what if it doesn't. What if there is a miss-configuration or a code bug that prevents this? The answer to this is often something along the lines of "well that's the customers fault so it's on them!" - while I agree with this to some extent, as a service provider I'd like to limit the affect that such a problem causes not only for the customer but also for my own network as the broadcast storm that sweeps the whole VPLS instance is traffic on my routers and switches - process and memory load too. Also there's am OPEX cost here too as the customer see's a meltdown of his/her's entire service they inevitably call their service provider and this kind of thing tends to get directors called at 3am. So really I agree with you 100% - but having a protocol that actually works independently of the customers RSTP and helps to confine such issues to the local segment is the goal - this isn't a replacement for STP on the local-segment - it's protection for the Service Provider that they can retain control of. Check out my proposed solution draft here: http://tools.ietf.org/html/draft-l2vpn-vlpm-01 This is going to be replaced soon with a more consolidated version. Sat, 13 Apr 2013 18:09:35 -0500 http://www.ethernetacademy.net/index.php/Mobile-Backhaul/1093-Single-CoS-versus-Multi-CoS.html#1093 Hi, I read the white paper by MEF on advantages of Multi CoS but a couple of things are not clear. I would like to know if a service provider gives a single CoS to customer with highest priority, then what is the advantage of using PCP bits by the customer. Is it right to say that the service provider uses PCP bits for priority queing only when there is Multi CoS implemented ? Faisal Fri, 12 Apr 2013 09:22:54 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/1091-VPLS-Loop-Protection-Requirement-Internet-Draft.html#1092 Hello Rich. I have read draft. I think that it is nessesary to separate 3 scenarios. 1. Layer-1 loop at the PE interface or within the same ethernet collision domain. 2. Layer-2 loop external to the PE within a broadcast domain that includes only one PE interface. 3. Layer-2 loop external to the PE within a broadcast domain that includes more than one PE interface. For 1 scenario. I suppose that Ethernet loop frame must do its work here. For scenario 2. It depends where the loop actually exist and how many devices we have ( CS ). If we have only one Customer switch and loop between CS and PE - this scenario merged with 1. If we have loop behind the switch. What is the first comes to the mind - STP/RSTP inside Customer side must resolve this issue. On the PE side - MAC flapping between CE facing interface and LSP. For scenario 3. More interesting scenario. As I remember we actually do not have standart mechanisms here ( not including STP/RSTP ), but maybe my knowledge is old and I need refresh them. Best regards Konstantin. Mon, 08 Apr 2013 09:17:28 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/1091-VPLS-Loop-Protection-Requirement-Internet-Draft.html#1091 Hi All, I'd like to draw the communities attention to a requirement internet draft submitted to the l2vpn IETF working group for VPLS loop protection. https://datatracker.ietf.org/doc/draft-boldy-l2vpn-vplsloop-req/ This draft needs support from service-providers to communicate to the IETF that this is a known, important and real-world issue. If you wish to do so or provide any input please respond to the author (me) from the the link above and be sure to cc the working group at l2vpn@ietf.org Any help from the MEF community here would be appreciated. Please feel free to contact me about this proposal. We also have a solution draft here for this issue that will be submitted once the requirement draft is either adopted by the L2VPN WG or gains sufficient peer-review. You can also sign-up for the working-group mailing list here: https://www.ietf.org/mailman/listinfo/l2vpn Kind Rgds Rich./ Wed, 27 Mar 2013 12:49:15 -0500 http://www.ethernetacademy.net/index.php/Operations-Administration-Maintenance/1090-Hard-or-soft-implementation-of-reflection.html#1090 I'm wondering about how RFC2544/Y.1564 reflection (loopback) is typically implemented in deployed products. Obviously whether this is implemented in hardware or software will affect how accurately the measurements are. For example with respect to latency and IFDV ("jitter"), an implementation in Linux might add hundreds of microseconds of latency to the data path for reflected packets and some amount of jitter. Given that typical LTE backhaul SLA requirements are in the 5-10ms latency and 1-2ms "jitter" this seems like a problem. But I don't know what's common. Does a lot of existing equipment implement reflection functions in software and thus this kind of variation from real world performance is expected? Or what? Tue, 26 Feb 2013 12:34:52 -0600 http://www.ethernetacademy.net/index.php/Standards-Technologies/1089-MEF-Requirements-for-Reflection-Support.html#1089 I'm trying to figure out if any of the MEF 2.0 specs (or any other standard for that matter) require implementation of Y.1564 (or even RFC2544) reflection, meaning are Carrier Ethernet devices REQUIRED to support either port or VLAN-based configurable loopback? Obviously some customers/service providers may require such support. I'm wondering if there are any standards that dictate this? Tue, 26 Feb 2013 12:27:50 -0600 http://www.ethernetacademy.net/index.php/Standards-Technologies/1088-Status-of-Ethernet-Service-Latching-Loopback.html#1088 I don't know exactly what this group is doing, but I'm assuming it's related to RFC2544/Y.1564 reflection in a device, either a NID or a deployed carrier switch. Does anybody know what the status of this work is? Looks like the work is still in progress as I can't find anything discussing it at this point... Tue, 26 Feb 2013 12:20:20 -0600 http://www.ethernetacademy.net/index.php/Standards-Technologies/1085-Color-modes-inside-operator-network.html#1087 Hi Alex. Thank you for the answer. I have understood what I wanted from you reply. Konstantin Thu, 21 Feb 2013 08:28:58 -0600 http://www.ethernetacademy.net/index.php/Standards-Technologies/1085-Color-modes-inside-operator-network.html#1086 Hi Konstantin, The metering typically happens at the demarcation line between customers and operators (at the edge of the network) and not inside the network. From this perspective, the metering/marking is performed by ENNI or UNI-N (the side of UNI that belongs to the operator) and the Color Mode (CM) is applicable to the equipment which belongs to the operator. After the incoming frame receives a color at ENNI/UNI-N, the network uses this color to resolve possible congestions without changing the color. The next chance to change the color will be at the edge when the frame exits the operator's domain (either ENNI or UNI-N). This is where an egress BW profile may be applied and the CM may be relevant again. However, egress BW profiles are rarely applied, so in practice the CM mode mostly applied when customer frames enter the operator's domain. Getting back to your original question, the decision of a customer to use its own metering scheme within customer premises does not affect operator's decisions regarding the CM at UNI-N/ENNIs. From operator's perspective these decisions are independent. Regards, Alex Wed, 20 Feb 2013 12:41:04 -0600 http://www.ethernetacademy.net/index.php/Standards-Technologies/1085-Color-modes-inside-operator-network.html#1085 Hi, Does operator network must always has color-aware mode ? For example, if we have color blind mode on UNI and metering customer traffic ( green/yellow/red) operator must comply with SLS and account that colors inside it's network. The same correct if we already have color-aware mode on the UNI. So situations where Customer | Operator blind blind aware blind Impossible Is this correct ? Konstantin Wed, 20 Feb 2013 09:50:37 -0600 http://www.ethernetacademy.net/index.php/Cloud-Networks/874-Cloud-Networks-Forum.html#1084 Hi Roman, I agree with your observation about the Cloud Services which seem to be unrelated to the Carrier Ethernet. However, the Ethernet layer within Data Centera is essentially different from a typical Ethernet Layer which you'll find within an Enterprise environment. You may want to have a look at the following technologies appicable for the Data Center Ethernet: * Priority based flow control - 802.1Qbb, 802.3bd * Congestion Notification - 802.1Qau * Enhanced Transmission Selection - 802.1Qaz * Data Center Bridging Exchange - 802.1Qaz * Shortest Path Bridging - 802.1aq * EITF TRILL * Fiber channel over Ethernet None of these are natively supported by Carrier Ethernet Networks, which in turn creates some issues for Ce Service providers. Additionally, from the CE Services perspective, the requirements of a typical service requested by Cloud Services operators from CE Service providers does not fit the classic business models that are applicable for Triple Play, Mobile or business services over CEN. So this is a very complex issue that deserves well to be in a separate forum thread. I guess we could benefit from a contribution on this subject. /Alex Wed, 20 Feb 2013 06:35:49 -0600 http://www.ethernetacademy.net/index.php/Cloud-Networks/874-Cloud-Networks-Forum.html#1083 Yes, of course, the ethernet services is essential for cloud networks. But they don't intersect. Most of cloud services (Software on Demand, for example) work at upper layers than Ethernet services. Wed, 20 Feb 2013 03:45:41 -0600 http://www.ethernetacademy.net/index.php/Transport-Technologies/642-Carrier-Ethernet-Transport-Technologies-definitions/Page-2/Page-3.html#1082 The beauty of the layer model is the way it keeps things simple. The other nicety is that each layer can have layers. It was proposed earlier that Layer 2 has two major groups of protocols: L2-Transport (PDH, SDH, etc.) and L2-Data (HDLC, Ethernet etc.). When trying to apply this classification to new protocols I find that a good test is how the protocol multiplexes its different users. When the bitstream is divided up into channels - it is a good bet that the protocol is L2-T. If the multiplexing allows dynamic allocation of the bitstream as needed, it is probably L2-D. L1 also is in the business of providing channels (FDM, WDM, Polarization Mode etc) and this is maybe why L1 and L2-T are thought of as one. Tue, 19 Feb 2013 16:57:04 -0600 http://www.ethernetacademy.net/index.php/Cloud-Networks/874-Cloud-Networks-Forum.html#1081 Hi Roman, CEN can be/are used to interconnect between different cloud networks. This in turn creates a unique combination of requirements for CEN which are typically not seen in all other cases. The planning and management of CEN for interconnecting such Cloud Services is less then trivial task for modern providers, therefore MEF currently considers further actions regarding the Cloud Services. /Alex Tue, 19 Feb 2013 16:31:55 -0600 http://www.ethernetacademy.net/index.php/Global-Interconnect/443-ENNI-and-MultiCast.html#1080 Anuradha Udunuwara wrote: Dear Daniel, When 2 EoIP/MPLS (Ethernet over IP/MPLS) based C/ME (Carrier/Metro Ethernet) networks are connected using ASBRs (Autonomous System Border Routers), keeping the 2 networks in 2 different ASs, the ASBR link becomes an ENNI. You can think the 2 ASs as 2 operators. In this case to distribute Protocol Independent Multicast-Sparse Mode (PIM-SM) traffic across the AS boundary we use the protocols like MP-BGP (Multi Protocol extensions of BGP) and MSDP (Multicast Source Discovery Protocol). According to the above, the ENNI supports multicast traffic distribution. Hope this helps. Anuradha Hi Anuradha, The technologies you've mentioned belong to Layer 3. To the contrary, the word Ethernet (as in Carrier Ethernet) implies Layer 2. Therefore, I do not see how your reply relates to the original question. Anyway, ENNI can support Multicast with either IGMP and in some rare cases with MMRP protocols. Alternatively, ENNI can simply treat the multicast frames similarly to broadcast/unknown unicast. So the answer is indeed, yes, ENNIs can support multicast. Alex Tue, 19 Feb 2013 16:25:21 -0600 http://www.ethernetacademy.net/index.php/Cloud-Networks/874-Cloud-Networks-Forum.html#1079 How Cloud Networks relate to Carrier Ethernet technologies ? Tue, 19 Feb 2013 07:41:21 -0600 http://www.ethernetacademy.net/index.php/Transport-Technologies/642-Carrier-Ethernet-Transport-Technologies-definitions/Page-2/Page-3.html#1078 I would like to add some other classification of Carrier Ethernet Transport, I would call it "physical environment" for Ethernet: - Ethernet over coaxial media - Ethernet over copper media - Ethernet over fiber/optical media Regarding to first message: there is Ethernet over ATM technology, it is widely used in ADSL connections: Ethernet frames between DSLAN and CE (ADSL router) encapsulated in ATM cels. Tue, 19 Feb 2013 07:33:08 -0600 http://www.ethernetacademy.net/index.php/Transport-Technologies/642-Carrier-Ethernet-Transport-Technologies-definitions/Page-2/Page-3.html#1077 Hi Anuradha, The WDM is a modulation, not a layer. Perhaps you want to say EoOTN. In this case, it is ETH over CO-CS. The fiber is a media, not a layer. You want to say, perhaps, ETH over ETY. In this case, it is CL-PS, except in PBB-TE networks where it is CO-CS. Best regards, Michel Mon, 18 Feb 2013 06:13:48 -0600 http://www.ethernetacademy.net/index.php/Global-Interconnect/443-ENNI-and-MultiCast.html#1076 Dear Daniel, When 2 EoIP/MPLS (Ethernet over IP/MPLS) based C/ME (Carrier/Metro Ethernet) networks are connected using ASBRs (Autonomous System Border Routers), keeping the 2 networks in 2 different ASs, the ASBR link becomes an ENNI. You can think the 2 ASs as 2 operators. In this case to distribute Protocol Independent Multicast-Sparse Mode (PIM-SM) traffic across the AS boundary we use the protocols like MP-BGP (Multi Protocol extensions of BGP) and MSDP (Multicast Source Discovery Protocol). According to the above, the ENNI supports multicast traffic distribution. Hope this helps. Anuradha Mon, 18 Feb 2013 04:20:35 -0600 http://www.ethernetacademy.net/index.php/Cloud-Networks/874-Cloud-Networks-Forum.html#1075 Any updates? Anuradha Mon, 18 Feb 2013 03:52:22 -0600 http://www.ethernetacademy.net/index.php/Transport-Technologies/642-Carrier-Ethernet-Transport-Technologies-definitions/Page-2/Page-2.html#1074 Dear Michel, Where would you place EoxWDM and EoFiber in the below categories? ETH over CO-CS (as SDH) ETH over CO-PS (as MPLS-TP) ETH over CL-PS (as IP) Anuradha Mon, 18 Feb 2013 03:44:18 -0600 http://www.ethernetacademy.net/index.php/Transport-Technologies/559-Comparison-of-performance-analysis-tests-for-CE-or-ATM-natively-or-over-SDH.html#1073 Hi Martin, Hope you completed the project and the degree. Can you share us your findings? Anuradha Mon, 18 Feb 2013 03:26:30 -0600 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/91-Question-about-design-of-Carrier-Ethernet-networks.html#1072 Dear Jerry, Pls check some of my slideshares in http://www.slideshare.net/udunuwara might help. Anuradha Mon, 18 Feb 2013 02:44:45 -0600 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/199-Provider-Backbone-Bridge-PBB-Primer/Tutorial/Page-2/Page-2.html#1071 Dear All, Pls check http://www.slideshare.net/udunuwara/pbbte might help. Anuradha Mon, 18 Feb 2013 02:00:28 -0600 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/197-MPLS-TP/Page-2.html#1070 Dear All, http://www.slideshare.net/udunuwara/optical-transport-network might help. Anuradha Mon, 18 Feb 2013 01:50:46 -0600 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/737-Doubts-about-PBB-technology.-Please-help-to-understand.html#1069 Dear Juan, Pls find the answers below; 1) How is PBB a more cost effective solution compared to L2VPN solutions? Only because of QnQ feature. Answer: PBB also can provide L2VPNs. PBB could be cost effective, because it does everything on the Ethernet world rather than using the support of other L2 technologies (SONET/SDH, IP/MPLS, MPLS-TP). 2) We need to run PBB over VPLS so the tunnels on core are going to be same as VPLS solution. Answer: This is alright. 3) In VPLS also routers only learns 2 mac(smac and dmac) so why PBB? Answer: In VPLS, VSI (Virtual Switching Instance)learns all the MACs coming from the participating UNIs. Ex:-If the devices attached through UNIs to the VPLS are routers then no. of MACs/VSI is the no. of routers. If the end devices are L2 switches then PBB can help you in hiding the customer MACs using a SP (Service Provider) MAC, making the scenario similar to if customer has a router as the end device. 4) PBB restrict mac learning on access switches is the only advantage over other L2VPN services? Answer: The main advantage of PBB is the MAC hiding or addressing the MAC learning scalability. 5) Is their any other cost effective solution on carrier Ethernet which we could suggest to customer instead of PBB? Answer: Other possible technologies are PB, PBB-TE, EoIP/MPLS and EoMPLS-TP with their own pros and cons including cost. 6) If we extend PE functionalities up to access using 1U switch(VFR on access switch) will do better convergence and restrict mac till access? whether this solution will be more effective that PBB? Answer: I think you mean VRF. Taking the VRF to access level requires taking IGP (Ex: OSPF) and label distribution protocols (LDP/RSVP-TE/BGP), which eventually makes the entire network flat (otherwise, the network has a hierarchy with IP/MPLS core, IP/MPLS aggregation and non-IP/MPLS access) and complex in terms of O & M. It is advisable to keep the access networks as simple as possible as they deal with many access Network Elements, many links and direct customer termination. In that sense it will not be cost effective than PBB. 7) Kindly provide some case studied on PBB and more documents. Answer: Pls check the following for more info; http://www.slideshare.net/udunuwara/pbbte http://www.slideshare.net/udunuwara/metro-ethernet-concepts http://www.slideshare.net/udunuwara/why-eompls-for-ce Anuradha Mon, 18 Feb 2013 01:37:04 -0600 http://www.ethernetacademy.net/index.php/Standards-Technologies/921-E-Tree-Question.html#1068 Ayal Lior wrote: Hi, Nothing preclude you from doing so. AFAIK, NodeB always communicated with the same RNC However, this RNC may be connected to 2 switches/routers who very well could connect over 2 links to PEs thus have 2 roots. These 2 roots are active at the same time. But again just one service scenario. My 2 cents, Ayal Hi Ayal, I agree with your observation about BTS/NodeB communicating with a single BSC/RNC. Just to clarify, this single RNC can actually comprise two physically different RNCs (in some cases, at different locations) connected to two different roots and running either HSRP or VRRP between them. This is to eliminate a single point of failure on RNC side. But you are right, they still appear as a single RNC device to NodeB. In case of a failure, HSRP/VRRP reconnects a backup RNC instead, so NodeB is not aware of the change (almost). Cheers, Alex Fri, 15 Feb 2013 11:19:29 -0600 http://www.ethernetacademy.net/index.php/Standards-Technologies/921-E-Tree-Question.html#1067 Hi, Nothing preclude you from doing so. AFAIK, NodeB always communicated with the same RNC However, this RNC may be connected to 2 switches/routers who very well could connect over 2 links to PEs thus have 2 roots. These 2 roots are active at the same time. But again just one service scenario. My 2 cents, Ayal Fri, 15 Feb 2013 09:09:11 -0600 http://www.ethernetacademy.net/index.php/IP-/-VPN-Services/168-Ethernet-Private-Line-for-IP/VPN-vs.-ATM-etc/Page-3/Page-4.html#1066 Anuradha Udunuwara wrote: Hi Alex, With H-VPLS you can have a hub-spoke type of architecture where the hub VSI learns the MACs and not the spokes. H-VPLS can also be used to connect to VPLSs. But in either case, you loose true any-to-any (mesh) connectivity (as compare to an IP VPN). Anuradha Hi Anuradha, There is a good reason why H-VPLS is called Hierarchical Private LAN service. I can assure you from my first hand experience that it is widely deployed for MP2MP services (Tier-3 to Tier-1 customers). Since getting into details will be off-topic for this thread, I recommend you to start with some research on the differences between deploying H-VPLS at Network Edge and at Network Core. If you have more questions, feel free to ask them in a separate thread, I will be more than happy to help. /Alex Fri, 15 Feb 2013 05:32:21 -0600 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/296-Question-about-perform-layer-2-ping-between-network-to-network-interface/Page-3.html#1065 Dear All, Similar to IP ping; MAC ping, MAC trace-route, LSP ping etc. are now available to to the same connectivity checking at L2. Anuradha Fri, 15 Feb 2013 04:34:36 -0600 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/317-Question-about-type-of-MPLS-VPNs/Page-2.html#1064 Dear All, IP/MPLS uses 2 labels; inner label and outer label. The inner label can be distributed through LDP(Targeted LDP)using the help of the IGP within a single AS. For the outer label signaling we can use LDP, RSVP-TE or BGP. Most of the vendors support all 3. In the case of VPLS, Cisco and others used to advocate LDP based VPLS while, Juniper advocated BGB based VPLS. The claim from the Juniper was that why use another protocol while iBGP is already been used for VRF connectivity in IP VPN. But now most of the vendors support both versions of the implementation. Some vendors support LDP and use BGP only for end discovery. Hope this helps. Anuradha Fri, 15 Feb 2013 04:29:57 -0600 http://www.ethernetacademy.net/index.php/Standards-Technologies/614-40G-UNI-E-NNI.html#1063 Hi Ayal, Can you update us with the current status? Anuradha Fri, 15 Feb 2013 03:11:46 -0600 http://www.ethernetacademy.net/index.php/Standards-Technologies/921-E-Tree-Question.html#1062 Why can't we use 2 roots for redundancy? ex:- leaves: BTS/NB/eNB, Roots: 2 BSC/RNC Anuradha Fri, 15 Feb 2013 03:06:50 -0600 http://www.ethernetacademy.net/index.php/Standards-Technologies/955-Need-info-on-MPLS-vs-MEF-for-regional-networks.html#1061 Hi Edwardo, One possible transport technology for MEF defined Ethernet services is MPLS. MEF is transport technology agnostic. Anuradha Fri, 15 Feb 2013 03:02:48 -0600 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/388-Enabling-RSVP-on-Cisco-Catalyst-6500.html#1060 Hi Agata, I hope you completed your testing. Can you share your experience? Anuradha Fri, 15 Feb 2013 02:54:34 -0600 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/122-Changing-our-existing-services-to-ethernet.html#1059 Hi Conner, Did you migrate the services to Ethernet? Can you share your experience? Anuradha Fri, 15 Feb 2013 02:51:59 -0600 http://www.ethernetacademy.net/index.php/Application-Implementation-Guidelines/846-Mobile-Backhaul-in-global-markets.html#1058 Hi Matt, I don't see any reason why one could not do that, except the challenge of meeting the delay/latency and IFDV required by different mobile technologies. Mainly the longer distances will affect clocking/synchronization. Anuradha Fri, 15 Feb 2013 02:46:09 -0600 http://www.ethernetacademy.net/index.php/Mobile-Backhaul/741-Planning-Tools.html#1057 Pls check http://www.opnet.com/solutions etwork_management/ Fri, 15 Feb 2013 01:10:57 -0600 http://www.ethernetacademy.net/index.php/Mobile-Backhaul/676-Live-deployment-of-MEF22-compliant-mobile-backhaul-networks.html#1056 Hi Ayal, It depends. A mobile operator who wants to have his own MBH network will select Ethernet or other (TDM) depending on the type of mobile technology he uses (2G, 3G or 4G) and the type of services he intends to offer. Anuradha Fri, 15 Feb 2013 01:07:46 -0600 http://www.ethernetacademy.net/index.php/Mobile-Backhaul/496-Mobile-Backhaul-market-drivers-and-challenges.html#1055 Hi Doron, Majority of the eNBs will have an Ethernet unlink. Anuradha Fri, 15 Feb 2013 00:59:49 -0600 http://www.ethernetacademy.net/index.php/Operations-Administration-Maintenance/993-Frame-Delay-Variation-or-Jitter.html#1054 Thanx Ayal for updating us on this highly important, yet least understood terms. Hope everyone will now use the correct term in correct place (IFDV vs. jitter). Anuradha Fri, 15 Feb 2013 00:42:48 -0600 http://www.ethernetacademy.net/index.php/Operations-Administration-Maintenance/844-Network-Visualization-Tools.html#1053 Dear Ayal, I have used ALU's 5620 SAM (Service Aware Manager), which supports Cisco and Juniper CE equipment in addition to ALU with some special MIBs. I have also used Huawei N/U2000 and it supports only Huawei equipment. Both of them can auto discover topologies and with a GUI can display the info. Anuradha Fri, 15 Feb 2013 00:39:33 -0600 http://www.ethernetacademy.net/index.php/Operations-Administration-Maintenance/279-Overprovisioning-for-Ethernet-OAM.html#1052 Dear All, If the Communications Service Provider (CSP) is reaching the customer with high bandwidth links (ex:- fiber), then they can sell the service BW excluding the OAM BW. The services with OAM can be charged higher (as they provide improved SLA) than services without OAM. My 2 cents. Anuradha Fri, 15 Feb 2013 00:16:56 -0600 http://www.ethernetacademy.net/index.php/Carrier-Ethernet-Equipment/688-Outdoor-CEN-Equipment.html#1051 Hi Mohd, Few vendors support these with/without outdoor cabinets. Ciena, Huawei, Cisco and ALU are some examples. Anuradha Fri, 15 Feb 2013 00:05:57 -0600 http://www.ethernetacademy.net/index.php/Carrier-Ethernet-Equipment/84-Cisco-ASR-9000.html#1050 Hi Bjőrn, Juniper's MX series, ALU's 7750-SR series and Huawei CX series are some examples. Anuradha Thu, 14 Feb 2013 23:06:56 -0600 http://www.ethernetacademy.net/index.php/Ethernet-Services/28-Ethernet-Services-Defined.html#1049 Now we have E-Access (Access EPL and Access EVPL) also. Thu, 14 Feb 2013 21:48:36 -0600 http://www.ethernetacademy.net/index.php/Ethernet-Services/757-Growing-Pains-Having-10Gig-Ethernet-Now-vs-Upgrading-to-40+GigE-Later.html#1048 Hi Jason, Though your post is 2 years old, I still thought of sharing my view for the benefit of the others. If you are at 1G, the next step would be n x 1G, then 10G, then n x 10G and then 100G. Anuradha Thu, 14 Feb 2013 21:19:57 -0600 http://www.ethernetacademy.net/index.php/Ethernet-Services/753-Limits-on-number-of-MAC-addresses-per-port.html#1047 Hi Chuck, In my view, you need to find the value of a MAC in your network and add some margin to it. No. of MACs supported per box is a constant. No. of VSIs per box is also a constant. You have to consider both of these when finding the real value and then set the price. PS: VLANs, LSPs, PWs etc are also limited resources per box. Anuradha Thu, 14 Feb 2013 21:14:51 -0600 http://www.ethernetacademy.net/index.php/Ethernet-Services/750-Application-of-Universal-Service-Fund.html#1046 Hi Chuck, CES (Carrier Ethernet Service) can be considered as a telecommunication service. But it is more appropriate to consider it as a data service. Anuradha Thu, 14 Feb 2013 21:08:52 -0600 http://www.ethernetacademy.net/index.php/Ethernet-Services/560-Ethernet-Services-performance-data-for-CE-to-compare-with-ATM-natively-or-over-SDH.html#1045 Dear Martin, Did you complete the project/degree? Anuradha Thu, 14 Feb 2013 21:03:47 -0600 http://www.ethernetacademy.net/index.php/IP-/-VPN-Services/168-Ethernet-Private-Line-for-IP/VPN-vs.-ATM-etc/Page-3/Page-3.html#1044 Hi Alex, With H-VPLS you can have a hub-spoke type of architecture where the hub VSI learns the MACs and not the spokes. H-VPLS can also be used to connect to VPLSs. But in either case, you loose true any-to-any (mesh) connectivity (as compare to an IP VPN). However, the main purpose of H-VPLS is to reduces the no. of LSPs (as compared to a fully meshed architecture) and hence reduce the complexity and increase the manageability. What I have highlighted is the normal industry practice of selecting and implementing L2 and L3 VPNs. Anuradha Thu, 14 Feb 2013 20:59:00 -0600 http://www.ethernetacademy.net/index.php/IP-/-VPN-Services/168-Ethernet-Private-Line-for-IP/VPN-vs.-ATM-etc/Page-3/Page-3.html#1043 Anuradha Udunuwara wrote: Dear Alex, I have not mentioned any specific problem, nor have I proposed OSPF. ... This relates to the size of your broadcast domain. If it a VPLS you are on a single broadcast domain. This is something you need to watch other than the IGP scalability. If the no. of branches are above 25 and the bandwidth requirement is low, use IP VPN. The problem of uncontrolled broadcasts in L2VPN with VPLS is typically addressed by H-VPLS without changing the service type to IP VPN. /Alex Thu, 14 Feb 2013 05:07:00 -0600 http://www.ethernetacademy.net/index.php/IP-/-VPN-Services/168-Ethernet-Private-Line-for-IP/VPN-vs.-ATM-etc/Page-3/Page-3.html#1042 Dear Alex, I have not mentioned any specific problem, nor have I proposed OSPF. Anuradha Thu, 14 Feb 2013 03:05:56 -0600 http://www.ethernetacademy.net/index.php/Ethernet-Services/53-Carrier-Ethernet-Services.html#1041 When this was posted over 3 years ago E-Access wasn't there  Ayal Wed, 13 Feb 2013 23:50:17 -0600 http://www.ethernetacademy.net/index.php/IP-/-VPN-Services/168-Ethernet-Private-Line-for-IP/VPN-vs.-ATM-etc/Page-3/Page-3.html#1040 Hi Anhurada, Since Shahram happens to be an editor of a few MPLS and MEF standards/drafts, his understanding is obviously the correct one here. Anyway, for the problem you mention the standard solution is H-VPLS, not OSPF. /Alex Wed, 13 Feb 2013 05:57:25 -0600 http://www.ethernetacademy.net/index.php/Transport-Technologies/1016-PBB-TE-vs-MPLS-TP/Page-2.html#1039 Anuradha Udunuwara wrote: Hi Alex, However, all the technologies mentioned there are CE transport technologies. I doubt what you mean by "Service layer technologies". For me, service layer technologies are E-Line, E-LAN, E-Tree and E-Access (as defined by MEF) and VPWS, VPLS (LDP/BGP), CES, PWE3, etc. defined by IETF and other. I guess we are getting off topic. Anyway, E-Line, E-LAN, E-TREE, E-ACCESS can be implemented with PBB, PBB-TE, MPLS-TP, IP/MPLS, EoSDH, POS, etc. There are still some more or less significant differences between these technologies for practical use, but from the perspective of your Table 5 all transport technologies should be equally fit the same purpose. I don't understand how you support MP2MP with MAC learning on SDH. Supporting MP2MP on top of SDH requires MPLS (IP/MPLS or MPLS-TP) with VPLS or PB/PBB/PBB-TE. Here is a Cisco white paper about their E-LAN services over EoS/POS. http://www.cisco.com/en/US/prod/collateral/optical/ps5724/ps2006/prod_white_paper0900aecd805fad97.html You should consider the EoSDH to be nothing more than just a transport "pipe" that connect two endpoints. The 802.1ad L2 bridge can happily run on top of any Transport layer "pipes" including 1000BaseT, EoSDH, POS, radio-links, OTN, MPLS-TP PWs, PBT trunks, etc. /Alex Wed, 13 Feb 2013 04:45:30 -0600 http://www.ethernetacademy.net/index.php/Ethernet-Services/123-Price-comparison.html#1038 Dear All, The normal practice: reduce price/bit as the bandwidth increase. Anuradha Wed, 13 Feb 2013 04:30:56 -0600 http://www.ethernetacademy.net/index.php/Ethernet-Services/53-Carrier-Ethernet-Services.html#1037 Dear Ayal, Your last paragraph should start as E-Tree, not E-Line. Other than the above 3 services where UNIs are involved, you also have E-Access , where the service is from UNI to E-NNI, which defines a OVC, instead of EVC. Anuradha Wed, 13 Feb 2013 03:20:36 -0600 http://www.ethernetacademy.net/index.php/IP-/-VPN-Services/168-Ethernet-Private-Line-for-IP/VPN-vs.-ATM-etc/Page-3/Page-3.html#1036 Hi Shahram, What David has said is correct. When you have few no.of sites with high bandwidth requirement go for VPLS. As long as your end device is a router you are talking about learning only the router WAN Ethernet MAC. So if you have 50 branches, it's 50 MACs. Anyway, I would suggest you limit it to max of 25 (no. of branches). If the end devise is not a router, but a L2 switch, then you need to learn all the host MACs in all branches and you need to have a good idea of what you are doing. This relates to the size of your broadcast domain. If it a VPLS you are on a single broadcast domain. This is something you need to watch other than the IGP scalability. If the no. of branches are above 25 and the bandwidth requirement is low, use IP VPN. If the no. of branches are above 25 and the bandwidth requirement per branch is also high, then you can still use IP VPN with Ethernet (fiber) access (Metro Ethernet). Anuradha Wed, 13 Feb 2013 03:12:13 -0600 http://www.ethernetacademy.net/index.php/Other-Services/493-VPN-vs-PPPOE.html#1035 Dear All, In the case of PPP being used in an ADSL environment, from the ADSL router/modem to the DSALM it is PPPoEoA. From the DSALM to the BRAS it is PPPoE. Therefore, from ADSL router/modem to BRAS it is PPPoE. Here I assume the network from DSALM to BRAS is based on Ethernet and not ATM. If it is ATM, then it could be PPPoEoA from ADSL router/modem to BRAS. Anuradha Wed, 13 Feb 2013 02:50:19 -0600 http://www.ethernetacademy.net/index.php/The-Functional-Applications-of-Carrier-Ethernet/479-Carrier-Ethernet-in-multi-site-multi-performance-environment.html#1034 Hi Tno, Pls remember that for a point-to-point communication (L1 bits/s, L2 frames/s, L3 packets/s, L4 segments/s) the effective rate of information/data transfer will be the rate that of the slowest link in the path. Anuradha Wed, 13 Feb 2013 01:45:42 -0600 http://www.ethernetacademy.net/index.php/The-Functional-Applications-of-Carrier-Ethernet/478-Carrier-Ethernet-and-data-replication-over-WAN.html#1033 Dear Dmitri, I think most of the questions you have asked are related to latency/delay. Delay/latency increases with distances and no. of hops the Ethernet farnme has to traverse. In general, as long as the physical interface associated with the UNI is Ethernet (electrical/optical), we can have storage-to-storage, DB-to-DB or PC-to-PC connectivity for DR/replication across a CE network. As different storage/DB/Computer systems require different delay/latency, frame drop and delay variation requirements, one need to be careful with the distance. Anuradha Wed, 13 Feb 2013 01:40:10 -0600 http://www.ethernetacademy.net/index.php/The-Functional-Applications-of-Carrier-Ethernet/906-Symbol-Error-event.html#1032 Dear Ramesh, Pls check http://www-rohan.sdsu.edu/doc/matlab/toolbox/comm/tutor14.html Anuradha Wed, 13 Feb 2013 01:16:02 -0600 http://www.ethernetacademy.net/index.php/General/683-Carrier-Ethernet-Network-performance-tests-for-commissioning-bring-in-to-service.html#1031 Hi LUCIO, MEF focus is "Carrier/Metro Ethernet Service" and MEF harnesses the testing (OAM) capabilities available in IEEE and ITU to test the services. It's not in MEFs scope to define any underlying "Carrier Ethernet Transport Technologies". Depending on the transport technology you chose (IEEE technologies: B/PB/PBB/PBB-TE, MPLS technologies: IP/MPLS, MPLS-TP and transparent technologies: SONET/SDH, OTN, xWDM), you need to use differnt testing tools at Layer 1, 2 and 3 of the OSI stack. These are not specific to the service and been around for some time. Ex:- SDH BIR testing, OTDR, MAC ping, MAC trace-route, LSP ping, IP ping. Hope this helps. Anuradha Wed, 13 Feb 2013 00:59:22 -0600 http://www.ethernetacademy.net/index.php/Transport-Technologies/1016-PBB-TE-vs-MPLS-TP/Page-2.html#1030 Hi Alex, Thanx for pointing out the mistake. I agree that with VPLS you can learn MACs in MPLS-TP and support multi-point. I have also changed EoMPLS to IP/MPLS to make the things clarify more. I'll upload the corrected version to slide-share soon. However, all the technologies mentioned there are CE transport technologies. I doubt what you mean by "Service layer technologies". For me, service layer technologies are E-Line, E-LAN, E-Tree and E-Access (as defined by MEF) and VPWS, VPLS (LDP/BGP), CES, PWE3, etc. defined by IETF and other. I don't understand how you support MP2MP with MAC learning on SDH. Supporting MP2MP on top of SDH requires MPLS (IP/MPLS or MPLS-TP) with VPLS or PB/PBB/PBB-TE. With PBB-TE, there's no automatic MAC learning. The forwarding tables are made using the EMS (I've updated that as well in the table). What I compare in the slide, subjected to the corrections mentioned above, are some features (service specific or not) supported by each transport technology. Thanx again for the clarification. Anuradha Tue, 12 Feb 2013 22:33:51 -0600 http://www.ethernetacademy.net/index.php/Transport-Technologies/1016-PBB-TE-vs-MPLS-TP/Page-2.html#1029 Anuradha Udunuwara wrote: Hi Ishkhan, Pls check http://www.slideshare.net/udunuwara/why-eompls-for-ce for a comparison between various CE technologies. Anuradha Hi Anuradha, The summary table (Slide 5) is very misleading due to the fact that it compares Service Layer technologies with Transport Layer technologies. For example, while the fact that MPLS-TP does not support the MAC learning may be somewhat correct, implementing MP2MP services with MPLS-TP is pretty common today: http://www.cisco.com/en/US/docs/wireless/asr_900/feature/guides/VPLS-MPLS-TP.html Just to give you an idea why we shouldn't compare Services with Transport, consider the following example. A 1000BaseT electric link is by definition a point to point connection which does not support MAC learning. It would be obviously incorrect to jump into conclusion that 1GE links can not be used with MP2MP deployments. To summarise, from the Services perspective (this is what operators most care about), operator can deploy MP2MP services with MAC learning over all the technologies listed in the Table 5 - over PBB-TE, over MPLS-TP, over EoSDH, etc. Anyway, all these technologies are solving the same set of network problems and therefore they provide a similar set of tools to tackle them. I hope this helps. Best Alex Tue, 12 Feb 2013 05:55:05 -0600 http://www.ethernetacademy.net/index.php/General/863-Questionaire-on-Metro-Ethernet.html#1028 Dear Derrick, Pls check http://www.slideshare.net/udunuwara/metro-ethernet-concepts Anuradha Mon, 11 Feb 2013 23:53:29 -0600 http://www.ethernetacademy.net/index.php/Optical-Equipment/169-I-need-some-recommendations.html#1027 Dear Eric, The following slide-shares might help; http://www.slideshare.net/udunuwara/optical-transport-network http://www.slideshare.net/udunuwara ext-generation-otn Anuradha Mon, 11 Feb 2013 23:50:37 -0600 http://www.ethernetacademy.net/index.php/Other-Equipment/391-CES-Interworking-between-different-brand.html#1026 Hi Mohd, As Ayal mentioned, if both vendors comply MEF 18, then it should work. In more practical terms, not only for CES, even for any other standard, plug-fest/hot-stage IOT is the best way. EANTC IOTs happen with Carrier Ethernet World Congress is a very good example. Anyway, before you purchase, you should also check them yourself. Anuradha Mon, 11 Feb 2013 23:44:31 -0600 http://www.ethernetacademy.net/index.php/Carrier-Ethernet-Equipment/847-different-equipment-vendors-to-support-Carrier-Ethernet.html#1025 Hi Matt, Your observations area right. International data connectivity services have predominantly and historically are based on SONET/SDH. "International Private Leased Circuits (IPLC)" are Point-to-Point links of N x 64 kbps, E1/T1, E3/T3, STM-1/4/16/64 dedicated speeds. This was ok for TDM payloads, but we are sometimes using them today even for data. With EoSDH with VCAT and GFP we could use N x VC12 to achieve different Ethernet speeds (2, 4, 6 ..Mbps). The main reason to keep this model, as you have observed, is because of the Interoperability. SONET/SDH systems of different countries have proven to interoperable over the years. IP peering between international operators have also proven to interoperable over the years. But for Ethernet, it has not yet happened, though the MEF expects the Global Ethernet Interconnect to be poised. It'll take sometimes and as the legacy/TDM to IP/Ethernet migration completes the international connectivity also will migrate to Ethernet, truly end-to-end. Anuradha Mon, 11 Feb 2013 23:34:59 -0600 http://www.ethernetacademy.net/index.php/The-Functional-Applications-of-Carrier-Ethernet/1020-MEF-in-INDIA.html#1024 Hi Ashlesha, I would rather say that you should check the Carrier/Metro Ethernet adaptation in India, in which the MEF defines "Carrier Ethernet". Not only in India, worldwide we see operators transform their networks to IP/Ethernet from legacy TDM. The value/business propositions of Ethernet and that of Carrier/Metro Ethernet are well understood by the industry (operators, SDOs, manufacturers, etc.) The 4 services defined by MEF, namely, E-Line, E-LAN, E-Tree and E-Access are used by many operators as services. This should be same for India as well. The future of Carrier/Metro Ethernet and the services worldwide is very clear and many research firms have predicted huge revenues in the coming years from services and equipment. Anuradha Mon, 11 Feb 2013 23:14:08 -0600 http://www.ethernetacademy.net/index.php/Transport-Technologies/1016-PBB-TE-vs-MPLS-TP.html#1023 Hi Ishkhan, Pls check http://www.slideshare.net/udunuwara/why-eompls-for-ce for a comparison between various CE technologies. Anuradha Mon, 11 Feb 2013 23:03:15 -0600 http://www.ethernetacademy.net/index.php/Transport-Technologies/1016-PBB-TE-vs-MPLS-TP.html#1022 Hello, I would really love to see comparisons showing negatives of PBB/PBB-TE vs. MPLS/MPLS-TP. Scalability, Security, OAM, Flexibility, etc. So far have not seen that clear “read line”. I'd would think of it's more of a platform which effectively suites the purpose. If multi protocol process to be considered (TDM, FR, etc.) then probably MPLS/MPLS-TP would be in favor. If it's about Ethernet/IP processing then I assume Ethernet centric approach (PBB/PBB-TE) would be more effective. And even here CESs are available. In regards to cost studies. I'm not quite sure if I have an objective info available. But can bring our own example. Our company is in process of deployment of PBB/PBB-TE network. What we've seen is cost effectiveness is just amazing. Thank you Ishkhan Sun, 10 Feb 2013 10:08:06 -0600 http://www.ethernetacademy.net/index.php/The-Functional-Applications-of-Carrier-Ethernet/1020-MEF-in-INDIA.html#1021 Hi, You can see a list of MEF members at http://metroethernetforum.org/members.php Ayal Sun, 27 Jan 2013 23:49:51 -0600 http://www.ethernetacademy.net/index.php/The-Functional-Applications-of-Carrier-Ethernet/1020-MEF-in-INDIA.html#1020 Hi, Can anyone please enlighten me with the progress of MEF adaptation in India? How many companies are MEF certified? What's the future of MEF services in India? How are the current job opportunities in MEF certified companies? I am a novice in this field, I would appreciate more informative and detailed answers. Thanks in advance. Sun, 27 Jan 2013 17:29:29 -0600 http://www.ethernetacademy.net/index.php/Transport-Technologies/1016-PBB-TE-vs-MPLS-TP.html#1019 I would agree that MPLS-TP is becoming more of a standard and most vendors are pushing this technology. Cisco has a turn-key solution using Cisco CPT platform which uses MPLS-TP. Advantasge of having MPLS-TP is that it integrates with IP-MPLS (MPLS-TE) very seamlessly without any further devices for translation of protocols unlike PBB-TE. Wed, 23 Jan 2013 14:45:52 -0600 http://www.ethernetacademy.net/index.php/Transport-Technologies/1016-PBB-TE-vs-MPLS-TP.html#1018 Hi...According to some of our (Infonetics Research) past gloabal service provider studies, not many operators are considering PBB-TE any longer, but a big majority are planning or "using" MPLS-TP in Access and some in Aggregation. Not many manufacturers even offer PBB-TE, but those include Cyan, Tejas, and I'm not remembering any others, but there may still be 2-3 more with older equipment/software. I don't recollect ever seeing any cost studies, since MPLS-TP is not available even yet on widespread basis. Pretty much PBB-TE nearly died while MPLS-TP was getting standardized, so any "cost studies" would have to have been theoretical--certainly not based on any side-by-side live network implementations. Thanks, Michael Fri, 04 Jan 2013 12:01:05 -0600 http://www.ethernetacademy.net/index.php/Transport-Technologies/1016-PBB-TE-vs-MPLS-TP.html#1017 Zsigmond P.Toth wrote: ... Some studies claim that PBB-TE has lower TCO, others claim it has low industry acceptance ... Hi Zsigmond , I wonder if you can provide a reference to the studies that find PBB-TE to be cheaper than MPLS-TP? Since one of the MPLS-TP goals was to reduce the TCO compared to the TCO of the good old MPLS-TE, I am surprised by this claim. Thanks Alex Fri, 04 Jan 2013 11:06:15 -0600 http://www.ethernetacademy.net/index.php/Transport-Technologies/1016-PBB-TE-vs-MPLS-TP.html#1016 Hey, I'm writing my diploma thesis on the topic of Carrier Ethernet. I'm trying to create a comparison of the most promising CE transport technologies but cannot decide which one is 'better' or 'more cost effective': PBB-TE or MPLS-TP Some studies claim that PBB-TE has lower TCO, others claim it has low industry acceptance. Some claim that MPLS-TP is not mature enough, others claim everybody to go with MPLS-TP when CE transport has to be chosen. So what's the truth? Please shoot with pros and cons for both! Thanks in advance Fri, 28 Dec 2012 20:52:55 -0600 http://www.ethernetacademy.net/index.php/The-Functional-Applications-of-Carrier-Ethernet/419-What-why-and-how/Page-2.html#1015 thx Ayal Thu, 20 Dec 2012 11:05:48 -0600 http://www.ethernetacademy.net/index.php/Standards-Technologies/1011-About-Ingress-BWP-per-UNI-attribute.html#1014 Hi, I checked with the MEF technical committee. The intent is that for private services, the ingress BWP be specified either on the single EVC or per CoS ID for that EVC. Therefore, also EP-LAN and EP-Tree MUST not have per UNI BWP. MEF 6.2 is expected to reflect this. Ayal Thu, 20 Dec 2012 09:17:47 -0600 http://www.ethernetacademy.net/index.php/The-Functional-Applications-of-Carrier-Ethernet/419-What-why-and-how/Page-2.html#1013 Hi, I checked with the MEF technical committee. The intent is that for private services, the ingress BWP be specified either on the single EVC or per CoS ID for that EVC. Therefore, also EP-LAN and EP-Tree MUST not have per UNI BWP. MEF 6.2 is expected to reflect this. Ayal Thu, 20 Dec 2012 09:17:12 -0600 http://www.ethernetacademy.net/index.php/The-Functional-Applications-of-Carrier-Ethernet/419-What-why-and-how/Page-2.html#1012 ..form MEF6.1 Tables 10 and 11 ..Must not Specify at UNI but is OPTIONAL at EVC per UNI attribute... My guess is there is no use case for them because it is a point-to-point service and can be handled at the EVC per UNI attribute level ..but like you said, they are optional for the EP-Lan and EP-Tree multipoint/root - multipoint services Wed, 19 Dec 2012 10:38:49 -0600 http://www.ethernetacademy.net/index.php/Standards-Technologies/1011-About-Ingress-BWP-per-UNI-attribute.html#1011 Does anybody know why ingress BWP per UNI attribute on EPL service must not be spesified? but it is supported on the other port-based services? Thanks. Mon, 17 Dec 2012 04:06:45 -0600 http://www.ethernetacademy.net/index.php/The-Functional-Applications-of-Carrier-Ethernet/419-What-why-and-how/Page-2.html#1010 Does anybody know why ingress BWP per UNI attribute on EPL service must not be spesified? But it is supported on the other port-based services? Thanks. Mon, 17 Dec 2012 03:11:28 -0600 http://www.ethernetacademy.net/index.php/The-Functional-Applications-of-Carrier-Ethernet/419-What-why-and-how.html#1009 Replying to self: http://www.ethernetacademy.net/index.php/Contributed-Articles/why-should-you-join-an-ethernet-exchange.html Mon, 10 Dec 2012 04:53:06 -0600 http://www.ethernetacademy.net/index.php/The-Functional-Applications-of-Carrier-Ethernet/419-What-why-and-how.html#1008 Are there any MEF suggested or standard templates for provisioning / ordering that tabulates the service details ? Do carriers have their own templates for collecting customer required parameters? Dimitri's article is on-target with this as there can be alot of information / details that needs to be exchanged between customer to providers and operators. Sun, 09 Dec 2012 04:06:21 -0600 http://www.ethernetacademy.net/index.php/The-Functional-Applications-of-Carrier-Ethernet/419-What-why-and-how.html#1007 Sun, 09 Dec 2012 03:56:27 -0600 http://www.ethernetacademy.net/index.php/The-Functional-Applications-of-Carrier-Ethernet/419-What-why-and-how.html#1006 The basic service offering is clearly explained in MEF 6.1 You see the attributes, their possible values and even have few examples in the appendix. This is a fine place to start. Then,. When you think about global interconnect, MEF 26.1 provides the answer. Ayal Sun, 09 Dec 2012 01:24:53 -0600 http://www.ethernetacademy.net/index.php/MEF-Certification-Equipment-Services/1002-L2CP-processing-Option-1-and-Option-2.html#1005 Hi, The MEF 9 & MEF 14 certification (CE 1.0) do not cover these options. These certifications are based on MEF 6, while these 2 options were introduced later on in MEF 6.1. Please refer to mEF9 test cases 24 through 26 to see what is tested. As to the new CE 2.0, I will try to get an answer. Note to Alex's explanation. MPLS transport can provide transparent L2CP handling only when there is no bridge component (e.g. not using VPLS MPLS) My 2 cents, Ayal Sat, 08 Dec 2012 08:36:39 -0600 http://www.ethernetacademy.net/index.php/The-Functional-Applications-of-Carrier-Ethernet/419-What-why-and-how.html#1004 With major carriers promoting and offering carrier ethernet services, the brochures just give a brief overview of E-Line E-LAN E-Tree services maybe some hierarchical MPLS drawing. My question is with all the features available on the MEF framework how is the provisioning of VLANs, CoS, Bandwidth Profiles, service attributes, etc exchanged. This may vary by carrier but is there a plain vanilla offering or where when and who exchanges (any granular attributes and values) with who ? Fri, 07 Dec 2012 02:13:37 -0600 http://www.ethernetacademy.net/index.php/MEF-Certification-Equipment-Services/1002-L2CP-processing-Option-1-and-Option-2.html#1003 1. The EPL, option 2 can/may be deployed with transport technologies other than SONET/SDH. It can be deployed, for example, over MPLS transport networks as well. 2. The certification for MEF-9 is performed by Iometrix (www.iometrix.com). I recommend you to contact them directly. They are very helpful and they will provide you with full information about tests, test beds, etc. This information is disclosed under the NDA, so theoretically, without singing the NDA, you are restricted to use only the publicly available information of the original MEF-9/14/etc. documents. /Alex Thu, 06 Dec 2012 11:57:53 -0600 http://www.ethernetacademy.net/index.php/MEF-Certification-Equipment-Services/1002-L2CP-processing-Option-1-and-Option-2.html#1002 Hi experts, MEF 6.1.1 describes the L2CP processing rules in the UNI. It gives two different paths for EPL: option 1 and option 2. Option 2 seems to be more transparent. My question is: when a equipement/SP is MEF9 certified, are both options tested or it depends on the core technology the customer will use? I am asking this because the Study Guide seems to imply that option 2 is used only with SDH/SONET/OTN core. "Transparent EPL Services For the very transparent EPL option (which can be carried over non packet-aware transport like SONET/SDH/OTN), the 2-step logic does not apply" http://www.carrierethernetstudyguide.org/MEF%20SG/pages/5attributes/studyguide_5-4-1.html I have a customer that will certify a product with MPLS core and is wondering wether EPL option 2 will be also tested. thxs Guillermo Wed, 05 Dec 2012 02:52:45 -0600 http://www.ethernetacademy.net/index.php/Ethernet-Services/998-Bandwidth-Delivered-on-EoSDH.html#1001 Hi Toni, I agree with Alex and would like to add some more details. Each VC-12 has a nominal rate of 2 Mbps which delivers 1.6 Mbps of Ethernet traffic. I believe you should measure the later as you sell BW to the user who doesn't care that your SONET infrastructure consumes 20% overhead. Using VCAS you can get bundle VC-12 usually up to 63, so you can get to ~ 100 Mbps in 1.6 Mbps BW steps. My 2 cents, Ayal Mon, 08 Oct 2012 01:45:07 -0500 http://www.ethernetacademy.net/index.php/Ethernet-Services/998-Bandwidth-Delivered-on-EoSDH.html#1000 You may want to have a quick look on VCAT and LCAS technologies. I would also recommend you to check the differences between STM-1, STM-4, STM16, STM-64, etc. As a brief answer, you can deliver EoSDH for Ethernet Services up to 10Gbps. For higher bit-rates vendors/operators typically switch to OTN. Sun, 07 Oct 2012 17:36:47 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/764-EtherSAM-vs-RFC-2544/Page-2.html#999 Thanks, Bruno Sun, 07 Oct 2012 12:10:24 -0500 http://www.ethernetacademy.net/index.php/Ethernet-Services/998-Bandwidth-Delivered-on-EoSDH.html#998 [1] I have ongoing thought about 'real' bandwidth delivered through EoSDH, should it be 2 Mbps as ethernet in common, or 1.6 Mbps as SDH bandwidth cause overhead issue. as common customer, EoSDH with ethernet in use, will only know that he have to get 2 Mbps for this service while the reality is only get 1.6 Mbps (using E-1 on SDH). some thought that we better put more E-1 (in this case 2 E-1) for EoSDH with consecuence of higher pricing (double reguler SDH price), so the customer will get 3.2 Mbps (we put bandwidth cap on CPE to hold only 2 Mbps). or the other way around. You just state EoSDH will only get 1.6 Mbps in real throughput. [2] How do you delivered 50, 60, 70 Mbps EoSDH. what with 2x SDH VC-3 or else? thanks for reply Sat, 06 Oct 2012 00:16:42 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/764-EtherSAM-vs-RFC-2544/Page-2.html#997 Hi Derrick, This is a very interesting question with regards to threshold settings in RFC 2544. Actually there are no standardized settings. You could argue if you are testing a network element that the Throughput should be around 100% of line rate at L1 (smaller at L2) and latency below a predefined value, but in a service testing environment, this is not possible. If you have a 10M business service between NYC and LA, it will have different service attributes then a 20M service from a Switch Site to a cell tower. To provide guidance with regards to threshold, I would refer you to MEF 23.1 (CoS IA Phase 2). MEF 23.1 will provide information for different services from and end-to-end point of view. From there, you can determine threshold that make sense in service activation (either RFC or Y.1564). Hope this helps! Best Regards Bruno Wed, 03 Oct 2012 14:00:09 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/764-EtherSAM-vs-RFC-2544/Page-2.html#996 Hi Bruno, there's a THRESHOLD setting in RFC2544 test; is there a standard setting? Eg 64byte (throughput >70%; latency Wed, 03 Oct 2012 03:56:28 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/994-Regarding-Carrier-Ethernet-prerequisite-skills.html#995 Hi, Basic understanding of carrier Ethernet can be obtained by reading some of the papers available on the "Papers" section. Then you read the MEF-CECP study guide at http://www.carrierethernetstudyguide.org/MEF%20SG/pages/general/studyguide-toc.html Good luck, Ayal Tue, 18 Sep 2012 23:51:44 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/994-Regarding-Carrier-Ethernet-prerequisite-skills.html#994 Hello everyone, I have good knowledge about TDM system (SDH,PDH,NGSDH,Ethernet over SDH,VCG,etc). Reading the basic information about carrier ethernet on various group under linkedin and other web materials, I am getting interested on it. As of the information i gather, EoSDH is the beginning for carrier ethernet. And VLAN tagging/MPLS/VPLS skill are necessary for configuring carrier ethernet services. Therefore, if you may, help me with the basic skill I need before trying out carrier ethernet field and also tools like (GNS3 and carrier ethernet support routers image) that i can self-learn it. Please suggest to start into carrier ethernet field/service. Highly hopeful to get response. Best Regards, Ashish Batajoo Mon, 17 Sep 2012 09:56:56 -0500 http://www.ethernetacademy.net/index.php/Operations-Administration-Maintenance/993-Frame-Delay-Variation-or-Jitter.html#993 One of the MEF-defined performance metric for an EVC is the Inter-Frame Delay Variation (IFDV). Some sources refer to this performance metric as jitter, so are they the same? The answer is NO. The term Jitter comes from signal processing and was later used as a performance metric of TDM traffic. Jitter measures the variation from a fixed known minimum (which corresponds to the frequency of the "signal"). In packet networks there is no such minimum and thus MEF 10.2 uses a method defined in RFC 3393 to compute the inter-frame delay variation. That is, measure the difference of one-way delay of two (consecutive) frames. Note that this measurement depends on the frame size amongst others. Indeed, both metrics try to solve the same problem of measuring the changes in delay between frames but they use different methods and of course would yields different results. Please note that the following specifications which define the performance metrics of Ethernet services never use the term jitter: • MEF 10.2 • MEF 35 • ITU-T Y.1731 Hope this helps. Ayal Mon, 17 Sep 2012 00:24:40 -0500 http://www.ethernetacademy.net/index.php/Application-Implementation-Guidelines/564-l2cp-handling-with-EVPL/Page-2/Page-3.html#992 zhu yao wrote: So why MEF 9 case 13 said that? MSTP and PTP have tag, I know any other protocols? This is not correct. There are no tags in MSTP BPDU encapsulation. MSTP uses a single untagged frame that encapsulates all the information about VLANs and STP instances inside. As for proprietary Tagged STP implementations (PVST), they use proprietary MAC aadresses and as such are not part of L2CP protocols defined by MEF/IEEE. I also assume that L2CP part of the PTP ( Peer-Delay messages ) uses untagged frames as well. To summarize, there are no tagged L2CP frames within the DA range of 01-80-C2-00-00-00 to 01-80-C2-00-00-0F. L2CP protocols that use VLANs use the DA range of 01-80-C2-00-00-10 to 01-80-C2-00-00-2F. MVRP is one of such protocols, for example. It looks like MEF-9 performs tests for the more common L2CPs in the 00 to 0F MAC addresses block, and therefore it correctly requires to use the proper EVC to handle untagged traffic. I hope this helps. Alex Thu, 13 Sep 2012 07:02:11 -0500 http://www.ethernetacademy.net/index.php/Application-Implementation-Guidelines/564-l2cp-handling-with-EVPL/Page-2/Page-3.html#991 OK, sounds reasonable Thanks for your help Wed, 12 Sep 2012 03:15:29 -0500 http://www.ethernetacademy.net/index.php/Application-Implementation-Guidelines/564-l2cp-handling-with-EVPL/Page-2/Page-2.html#990 Hi, Let's step back and think of L2CP handling. You first need to pass the decision at the UNI, now say you decided to tunnel a certain L2cP you now map it to EVC. If it is tagged, then it goes according to the CE-VLAN ID / EVC map. Case 13 tests that untagged case and now the requirement is to map these L2CP along side untagged service frames to the EVC that untagged frames map to (per UNI's attribute "CE-VLAN ID for untagged and priority tagged Service Frames"). I hope this clarifies your concern. Ayal Wed, 12 Sep 2012 03:08:44 -0500 http://www.ethernetacademy.net/index.php/Application-Implementation-Guidelines/564-l2cp-handling-with-EVPL/Page-2/Page-2.html#989 So why MEF 9 case 13 said that? MSTP and PTP have tag, I know any other protocols? Wed, 12 Sep 2012 03:00:00 -0500 http://www.ethernetacademy.net/index.php/Application-Implementation-Guidelines/564-l2cp-handling-with-EVPL/Page-2/Page-2.html#988 Hi, Not quite so. Indeed, some L2CP are link layer and have no tag (e.g. PAUSE, LLDP, Link OAM). However other L2CPs may very well be tagged. For example: 1. MSTP is clearly VLAN tagged. 2. PTP (1588 ) can be tagged My 2 cents, Ayal Wed, 12 Sep 2012 00:57:52 -0500 http://www.ethernetacademy.net/index.php/Application-Implementation-Guidelines/564-l2cp-handling-with-EVPL/Page-2/Page-2.html#987 MEF9 Case 13 Verify an EVC configured with the EVC Layer 2 Control Processing Attribute associates two or more UNIs and when specific Layer 2 Control Protocols are configured for tunneling, the MEN tunnels the specific Layer 2 Control Protocols on theEVC that is mapped to the untagged CE-VLAN ID in the CE-VLAN ID/EVC Mapand deliver the Service Frames carrying the specific Layer 2 protocols at all egress UNIs in the EVC identical to the corresponding ingress Service Frames. Wed, 12 Sep 2012 00:45:25 -0500 http://www.ethernetacademy.net/index.php/Application-Implementation-Guidelines/564-l2cp-handling-with-EVPL/Page-2/Page-2.html#986 In MEF 9 ATS, it said that only frames untagged or the vlan which untagged frames mapped to are recognized as L2CP,other are recognized as user traffic I think pratically all L2CP is untagged Wed, 12 Sep 2012 00:21:57 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/764-EtherSAM-vs-RFC-2544/Page-2.html#985 Hi Toni, Sorry for the long time to reply. The concept of iteration comes directly from RFC 2544. If you look in section 26.2 of RFC 2544 you will see the description of the methodology. The last paragraph states that "The test MUST be repeated at least 20 times with the reported value being the average of the recorded values." This is where my iteration concept is coming from. So each test is 2min, and must be repeated 20 times for each frame size (7 of them). The total is a minimum of 280 min or 4.66 hours. For your point 2. Acceptance of Y.1564 is spreading. Smaller service providers can easily convert their methods and procedures to Y.1564 as larger ones require more time. I will leave it to operators to comment more on this. For your point 3. You are right in stating that RFC 2544 and Y.1564 are methodologies. They are used to measure the different attributes of the service being tested (CIR/CBS/EIR/EBS, FD, FDV, Frame Loss) and compared to Service Acceptance Criteria. The attributes can differ from a customer to another depending on what is being delivered Hope this helps Bruno Thu, 06 Sep 2012 15:32:40 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/976-Reg-ETH-AIS.html#984 Hi, 1. The case when ETH-CC is disabled applies only for the propagation of AIS when an AIS is received. The client level MEP must process the received AIS and suppress alarms at its own level. But, whether the MEP generates AIS to propagate this to its client level depends on whether ETH-CC is enabled or disabled. 2. The MEP transmits AIS to only one higher level (as configured by the administrator, which is usually the level at which the Client MEG is configured) and not to all of its higher levels. 3. You can download it from here http://www.itu.int/ITU-T/recommendations/rec.aspx?rec=11512 Regards, Sriram Mon, 03 Sep 2012 02:10:18 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/981-RSTP-STP-Diameter.html#983 The spanning tree protocol takes some time to converge into the final spanning tree. When the topology is too complex, the spanning tree can't converge and may create loops. While the standard requires the diameter to be less or equal to 7 it is possible to have larger networks as well. The trick is to update the values of Hello time = 1 sec, Max Age and Forwarding delay: max_age = (4 x hello) + (2 x diameter) – 2 forward_delay = ((4 x hello) + (3 x diameter)) / 2 As far as I remember, you should be able to get to working networks with diameter up to 16, however I remember others mentioning diameters up-to 30+ (depends on the actual network topology). /Alex Tue, 28 Aug 2012 20:05:02 -0500 http://www.ethernetacademy.net/index.php/MEF-Certification-Equipment-Services/980-MEF-Compliance-for-Vendors.html#982 Hi Asad, Thanks for these important questions. To date, there have only been certifications for compliance with MEF 9, MEF 14 and MEF 18. As of February this year, we refer to MEF 9, MEF 14 and MEF 18 certification as Carrier Ethernet 1.0 certifications - although the use of the terms MEF 9, MEF 14 and MEF 18 certification is still valid as well. Your question asks why we only refer to certification according to MEF 9, MEF 14 and MEF 18, and not other specifications. The reason for this is that there are a range of specifications (service definitions, attributes, implementation agreements) In order to determine what should be tested, the MEF Technical Committee has defined documents called Abstract Test Suites (ATS) that describe what to test in specific areas. For example, MEF 9 describes what to test for functionality of an EPL, EVPL and ELAN. MEF 14 describes what to test for performance of an EPL, EVPL and ELAN. The ATS may draw on multiple specifications such as service definitions and attributes. By the way, not all ATSs result in a MEF certification. That is a function of market demand. In short, when complying with MEF 9 and MEF 14 ATSs, the equipment vendor (and the SP) are complying with a wider set of underlying specifications. Regarding your second question - E-Tree is not mentioned in the Certification Registry because MEF 9 and MEF 14 only cover EPL, EVPL and ELAN (CE 1.0) E-Tree belongs in CE 2.0 together with E-Line (EPL, EVPL), E-LAN (EP-LAN and EVP-LAN) and E-Access (Access EPL, Access EVPL) CE 2.0 certification (including for E-Tree compliance) is now underway as part of Cert Wave 1 and the first certified companies under Cert Wave 1 will be announced in January or February 2013 and will appear in the updated MEF Certification Registry. Sun, 26 Aug 2012 01:11:09 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/981-RSTP-STP-Diameter.html#981 Hello, I would like to submit you a question related to STP (IEEE802.1D 1998) and RSTP (IEEE 802.1D 2004). IEEE802.1D 1998 §8.10 specifies the 'Maximum Bridge Diameter" as the maximum number of bridges between any two points of attachement of end stations. The diameter is fixed to 7 bridges for the delay values indicated in the IEEE802.1D 1998 specifications. Does this notion of Diameter also apply to the RSTP IEEE802.1D 2004? The IEEE802.1D 1998 never refers to Diameter. Is there in a RSTP network some constraints in terms of number of bridges that a frame can go through? thanks Philippe Fri, 24 Aug 2012 07:40:55 -0500 http://www.ethernetacademy.net/index.php/MEF-Certification-Equipment-Services/980-MEF-Compliance-for-Vendors.html#980 MEF has a registry section where it lists the Vendors and their equipments, who are are compliant to MEF Specs. However, most of the equipment is certified for MEF 9 and 14, while a few are for MEF 18 (CESoETH) as well. My question to the experts is that why other MEF Specs are not complied by any Vendor? Is it sufficient for a buyer to get MEF 9/14 certified equipment? Or it is understood that if the Vendor is complying to Abstract Test Suites (MEF 9/14) it is also complying with MEF Definitions, Services and their Parameters? Further, only a few of the Services (eg EPL. EVPL. etc) are listed in the certified section. In fact I did not find any E-TREE service listed. Any explanation? Thu, 23 Aug 2012 23:41:30 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/764-EtherSAM-vs-RFC-2544/Page-2.html#979 dear all, thanks for your sharing. it helps me a lot to understand the ethernet service test. but there's some question that i would ask : 1. to bruno : what do you mean by iteration ?? and why you stated 20 iteration? does it refer to bandwidth load? you are looking at a 4.66 hour test (2min per iteration x 20 iterations x 7 frame size) 2. to all : which operator are using RFC 2544 and which Y 1564. caused these thread is 1 year ago, how is recent udpate. is all Operator already using Y 1564?? 3. to all : RFC 2544 and Y 1564 is a methodology not parameter standard. so how should i know what customer want. and it woould be relative between one customer to other. CMIIW thanks for your reply Thu, 23 Aug 2012 05:15:46 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/976-Reg-ETH-AIS.html#978 Thanks Sriram ! Follow up of previous questions and 2 more. Please throw some light if you are aware. 1) Still i am not clear with the case if ETH-CC is disabled? What is deployment case and the condition to trigger AIS ? Do you mean to say at level n CC is enabled, which encountered CC loss but at its immediate higher level, CC is disabled but still it has to receive and propagate AIS? 2) Standard does not mandate the way of transmission? Is it that a level 'n' upon detecting CC loss has to transmit AIS to most immediate higher level and this level has to propagate it to its most immediate higher level? or a level encountered failure has to transmit AIS at one-shot to all of its higher levels? 3) Unfortunately, i am not having latest G.8021 document. Could you please share it across? Tue, 21 Aug 2012 00:29:46 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/976-Reg-ETH-AIS.html#977 Hi, It is not necessary that AIS has to be transmitted to level n+1. The client level (which is configurable) must be > MEP's level. 1. AIS is generated on receiving AIS frame (AIS condition) and on receiving LCK frame (LCK condition) when ETH-CC is disabled. 2. AIS is not generated for RDI defect. It is generated for LOC, Unexpected MEL, Unexpeted MEP, Unexpected MEG, AIS condition and LCK condition. Refer to G.8021 section 9.2.1.2. 3. CCMs follow the same path as data frames. When one of the pseudowire path goes down, the data takes the backup path. Similarly CCMs should take backup path. If there is no signal fail condition, I don't think AIS need to be generated. 4. Yes, I think so. Logically speaking, the AIS is generated for the VPLS instance corresponding to the service which is then forwarded out on all Attachment Circuits connected to the VPLS instance. Note: MEF 30.1 doesn't recommend AIS for multi-point services. Regards, Sriram Thu, 09 Aug 2012 09:00:41 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/976-Reg-ETH-AIS.html#976 Hi Experts/Friends, I am trying to understand Y.1731 ETH-AIS. As i understand, whenever there is CCM adjacency loss between two MEPs at level 'n', AIS frames has to be transmitted to level 'n+1', so that MEPs at higher level will suppress the alarms. Few questions here. Please clarify me. 1) I thought AIS is always coupled with CCM(loss/UP). But ITU-T Y.1731 statements resulted in confusion. What is the case if ETH-CC is disabled? • Signal fail conditions in the case that ETH-CC is enabled. • AIS condition or LCK condition in the case that ETH-CC is disabled. 2) Do we need to generate/propagate AIS even for RDI ? Is the CCM adjacency loss alone has to be considered for AIS or any defect condition should be considered ? Example: RDI, MAC_STATUS_ERR, ERRONEOUS_CCM, CROSS_CONNECT_CCM etc. 3) As per MEF compliant services, we have pseudowire redundancy, which means, at provider level, we have CCM protection aka service protection. So, do we need to still generate/propagate AIS to the higherlevels even if one service is not DOWN? 4) For ELAN services, do we need to just flood the AIS frames to all the interfaces part of bridge/VPLS? Because, we don't which other UNIs getting service via this CFM path at lower level are facing the problem. Thanks, http://www.ethernetacademy.net/images/fbfiles/files/T_REC_Y.pdf Wed, 08 Aug 2012 07:54:58 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/973-Layer-2-control-protocol-question.html#975 thank you Tue, 31 Jul 2012 09:28:16 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/973-Layer-2-control-protocol-question.html#974 L2CP and SOAM activities are done by different MEF working groups. MEF has decided to remove all the SOAM definitions from L2CP scope and to let the SOAM related people to produce all SOAM specific requirements within SOAM FM and SOAM PM related documentation. /Alex Tue, 31 Jul 2012 05:01:45 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/973-Layer-2-control-protocol-question.html#973 I don't understand why CFM packets(0180.c200.0030-0180.c200.003F) are not defineded as L2CP in MEF 6.1.1? thank you, Fri, 27 Jul 2012 22:07:24 -0500 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/963-About-SOAM-EVC-Level/Page-2.html#972 Thank for your help! Tue, 17 Jul 2012 09:14:08 -0500 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/963-About-SOAM-EVC-Level/Page-2.html#971 Ayal Lior wrote: Jason, The EVC level is NOT MPLS-TP OAM. The MPLS-TP OAM is considered the TRAN layer for the Ethernet service. It will be based on the LSP and will detect faults on that layer. If you are performing fault management or Performance measurement for an EVC, then you need to send the OAM frames through the EVC to the remote MEP. It will be then use the VC label of the PW that leads you to the remote MEP. As Alex explained, a single VPLS instance would have more than one VC label even if we assume that there is no redundancy. Hope this helps. Ayal Tue, 17 Jul 2012 09:13:14 -0500 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/963-About-SOAM-EVC-Level.html#970 Jason, The EVC level is NOT MPLS-TP OAM. The MPLS-TP OAM is considered the TRAN layer for the Ethernet service. It will be based on the LSP and will detect faults on that layer. If you are performing fault management or Performance measurement for an EVC, then you need to send the OAM frames through the EVC to the remote MEP. It will be then use the VC label of the PW that leads you to the remote MEP. As Alex explained, a single VPLS instance would have more than one VC label even if we assume that there is no redundancy. Hope this helps. Ayal Mon, 16 Jul 2012 03:44:22 -0500 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/968-WHAT-IS-CE-VLAN-ID-PRESERVATION.html#969 Hi Ahmed, The MEF-CECP study guide explains it quite well. Please refer to http://www.carrierethernetstudyguide.org/MEF%20SG/pages/5attributes/studyguide_5-4.html Ayal Mon, 16 Jul 2012 03:35:09 -0500 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/968-WHAT-IS-CE-VLAN-ID-PRESERVATION.html#968 Please tell me WHAT IS CE-VLAN ID PRESERVATION? Mon, 16 Jul 2012 02:47:41 -0500 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/963-About-SOAM-EVC-Level.html#967 Hi Alex, Thank you for your classify and cofirm. While on VPLS/VPWS mode , the SOAM for EVC level should be base VC label(MPLS-TP OAM). Is this correct? Sat, 14 Jul 2012 11:04:10 -0500 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/963-About-SOAM-EVC-Level.html#966 Jason Young wrote: As i known , In QinQ mode , cvlan -> svlan mapping, i think evc is map to one svlan; In MPLS mode, cvlan-> VPLS(vsi) or cvlan -> VPWS . Correct. i think evc is particular VC label. Is this correct? The EVC part of the VSI can be seen as an association of all PWs attached to the VSI. There can be more than one VC label in that case - VPLS, PW redundancy schemes, etc. Fri, 13 Jul 2012 05:30:05 -0500 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/963-About-SOAM-EVC-Level.html#965 Hi Alex, Thank you for your reply. Can you classfy the EVC layer on MPLS-TP mode. As i known , In QinQ mode , cvlan -> svlan mapping, i think evc is map to one svlan; In MPLS mode, cvlan-> VPLS(vsi) or cvlan -> VPWS . i think evc is particular VC label. Is this correct? Thank you again! Thu, 12 Jul 2012 10:03:43 -0500 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/963-About-SOAM-EVC-Level.html#964 Jason Young wrote: While deploy MEF service base MPLS-TP. What's encapsulation for EVC layer OAM. MPLS-TP PWE3 base Y1731 or Ethernet OAM over MPLS-TP PWE3? I mean which type OAM should run on EVC layer, CFM or MPLS-TP PW layer OAM. Thank you! SOAM is by definition the OAM for Ethernet Services (and nto for PWs, LSPs, etc.) Therefore the encapsulation of the SOAM over MPLS-TP is essentially the same as encapsulation of all other Ethernet Service Frames. Please note, that MPLS-TP OAM provides the OAM for network layers other than Ethernet Services layer and can be/will be used independently from SOAM frames. /Alex Thu, 12 Jul 2012 04:18:54 -0500 http://www.ethernetacademy.net/index.php/Network-Designs-and-Solutions/963-About-SOAM-EVC-Level.html#963 While deploy MEF service base MPLS-TP. What's encapsulation for EVC layer OAM. MPLS-TP PWE3 base Y1731 or Ethernet OAM over MPLS-TP PWE3? I mean which type OAM should run on EVC layer, CFM or MPLS-TP PW layer OAM. Thank you! Thu, 12 Jul 2012 03:02:44 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/961-Multi-point-EVC-with-ENNI.html#962 Hi Jason, You've managed to ask a simple question in a very complicated way. I would suggest you to study the examples of the MEF26, including the section 4.6 which describes the relationship between OVC and EVC. Best Alex Fri, 22 Jun 2012 04:07:59 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/961-Multi-point-EVC-with-ENNI.html#961 Hi all: One E-LAN service include three UNI. UNIA and UNIB in the same operator named as MENA. UNIC in the operator named as MENB. UNIA we need add the OVC mapping . Example: CVID-10 ---> OVC1 UNIA also need connection with UNIB, so we need configure EVC mapping . Example : CVID-10 ----> EVC1 I confused , UNIA/UNIB/UNIC is the same E-LAN service. The connection should be the same Multi-point to Multi-point EVC. But the configration include 2 EVCs. One EVC connect by OVC1 and other is connect directly. please help and correct my understand. Thank you. Thu, 14 Jun 2012 21:20:24 -0500 http://www.ethernetacademy.net/index.php/Organizations/574-MPLS-TP.html#960 Thanks Ayal, MPLS TP does it not work at the transport layer i.e ATM layer. MPLS TE is this at layer 2/3 Why two separate networks? Thu, 14 Jun 2012 10:00:28 -0500 http://www.ethernetacademy.net/index.php/Organizations/574-MPLS-TP.html#959 Hi Michael, I assume that you are asking about a scenario where you have 2 networks, one operated using MPLS-TE and the other operated using MPLS-TP. From data-plane perspective, these would look the same. Of course, each network and each link can have different label values, but the packets look exactly the same. They behave differently with regards to provisioning, OAM, etc. My 2 cents, Ayal Thu, 14 Jun 2012 06:45:31 -0500 http://www.ethernetacademy.net/index.php/Organizations/574-MPLS-TP.html#958 Hi, I am interested in the relationship between MPLS TP and MPLS TE. Are there any changes that take place between TP and TE. When a frame goes from TP to TE and maybe back to TP is there anything added or remove from the label or does it stay the same. Thanks Michael Wed, 13 Jun 2012 15:54:49 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/956-How-can-i-understand-OVC.html#957 Hi Jason, May I refer you to the MEF CECP study guide at http://www.carrierethernetstudyguide.org/MEF%20SG/pages/4components/studyguide_4-2.html An OVC could be thought of as EVC within a single MEN. OVCs are concatenated in order to create EVCs. An OVC may map multiple OVC end-points based on S-VLAN values. However, the OVC is NOT the VLAN. Hope this helps Ayal Sun, 10 Jun 2012 04:33:49 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/956-How-can-i-understand-OVC.html#956 Can i seems OVC as a S-VLAN? Sun, 10 Jun 2012 00:50:23 -0500 http://www.ethernetacademy.net/index.php/Standards-Technologies/955-Need-info-on-MPLS-vs-MEF-for-regional-networks.html#955 We have a client that has been indoctrinated by the MPLS players and we are trying to move them to an MEF type design. The client has limited MPLS knowledge so is being persuaded by the MPLS players and the client doesn't understand what the MEF provides. So we're trying to gather information that will help us educate our client. Is there something available regarding MPLS v MEF for regional networks? Thank you in advance for any assistance. EF Thu, 31 May 2012 12:38:23 -0500