Copyright ©Shanghai Baud Data Communication Co.

,
LTD. 2022. All Rights Reserved.

BCNE – BDCOM Certified Network Engineer

PPPOE Technical Service Dept.

Course Objectives

1. Internet technology
2. Key Technologies of Network Interconnection
3. Network Interconnection Case Analysis
4. Exercises
CONTENTS

1. Overview

2. PPPOE principle

3. PPPOE application

4. Exercises
PPPOE Overview
Overview

Development of PPPoE
PPPoE － Point to Point Protocol over Ethernet
PPP is a typical point-to-point link layer protocol, usually used on serial links; Ethernet is a typical
broadcast network. Why mix the two?
This is related to the broadband network development of ISP operators:
In the early days, users mostly used PPP dialing for Internet access, which mainly came from two aspects:
1. The hardware is carried on the PSTN telephone line. These are ready-made network resources does
not need repeated construction;
2. The PPP itself has authentication protocols such as PAP and CHAP, which is convenient for the ISP to
perform user authentication and billing functions
However, this method also has disadvantages:
1. The upper limit of bandwidth is 64k
2. Voice and data cannot be performed simultaneously
5
Overview
Development of PPPoE (continued)
Later, the product, ISDN, called "Integrated Services Digital Network", was introduced to improve the Internet bandwidth
and the simultaneous operation of data and voice

However, the popularization of ISDN in China is not good for many reasons:

1. The upper bandwidth limit is upgraded to 64k * 2, but it does not fundamentally solve
the problem

2. When to use ISDN, they must reapply for lines. The original PSTN line resources cannot
be fully utilized

3. The cost is relatively high

With the development of Ethernet (especially transmission) technology, large-scale Ethernet MAN networking
becomes possible, and the construction cost is getting cheaper and cheaper

Ethernet is characterized by high bandwidth and broadcast network: The former helps to meet the increasing
bandwidth requirements; The latter is also conducive to ISP networking. When an end user needs to be added, the
local end network does not need to change
6

Ethernet also has disadvantages: The layer 2 protocol ARP is too simple, and there is no necessary user
Overview

Development of PPPoE (continued)

In the current network development process, Ethernet and PPP come together naturally.

This is the ADSL of China Telecom, whose core technology is PPPoE

ADSL has two modes on network hardware:

1. Transmission on PSTN line through AD modulation and demodulation; But for user
terminal equipment, the network is Ethernet with bandwidth of no more than 2M, which
is widely used at that time

2. Use full Ethernet and optical fiber for networking with PPPoE protocol, the bandwidth
can reach FE, GE or even higher; This mode does not require Modem, but the networking
cost is high. 7
PPPOE principle
PPPOE Principle introduction

Composition of PPPoE protocol

Generally speaking, ADSL or PPPoE protocols consist of two stages:
One is PPPoE, mainly responsible for establishing a virtual point-to-point connection between the client and
the server in a broadcast network; The other is the PPP, which has no difference from the traditional PPP in
terms of functions, just its protocol stack.

Protocol stack
The PPPoE protocol stack is divided into two parts. The first part is about the establishment of point-to-point
virtual connection, which is the Discover discovery stage; The second is about the data stage of carrying PPP
messages. No matter which type, they are carried in the Ethernet standard message structure.
Disc o ve r DATA

Pa ylo a d
Ta g
PPP
PPPo E 9
Eth e rn e t
 PPPOE Principle introduction
PPPoE Packet Header
PPPoE header
Regardless of the stage, the content carried by the upper layer of PPPoE is different. The
standard header structure is:

Both Ver and Type must be forced to 0x01

Code determines the specific content in the payload, which has different meanings at different stages
Session_ID represents the identity of each virtual point-to-point connection and is a unique identifier
Length refers to the effective length of the payload
Payload is the load. For the Discover stage, all kinds of TAGs are carried here; for the DATA stage, all PPP frames are
carried here.
10
PPPOE Principle introduction
PPPoE Discover message
For the Discover phase, the Code values are as follows:
Code=09 ， message is PADI ， that is, PPPoE Active Discover Initiation
Code=07 ， message is PADO ， that is, PPPoE Active Discover Offer
Code=19 ， message is PADR ， that is, PPPoE Active Discover Request
Code=65 ， message is PADS ， that is, PPPoE Active Discover Session-Confirm
Code=a7 ， message is PADT ， that is, PPPoE Active Discover Terminate

 PPPoE common TAG

 In the control message in the PPPoE Discover stage, the Payload field is composed of several TAGs,
and each TAG conforms to the TLV structure, that is, Type-Length-Value. The common TAGs are:
 Service-Name
Type=0101, indicating the name of the PPPoE service; the service name provided by the server
must be the same as that of the Client.
But if the Length of Service-Name is 0, it means that any provided PPPoE service can be accepted.

11
PPPOE principle

PPPoE common TAG (continued)

 AC-Name
AC ， Access Concentrator ， generally called a concentrator or server.
Type=0102 ， this name is used to uniquely identify each PPPoE server, usually
composed of ISP's trademark
Host-Uniq
Type=0103, used to establish an association between the Client host and the AC server.
Specifically, the Client randomly generates the Host-Uniq value in the message initiated by itself,
and requires the Host-Uniq parameter to be the same when the AC responds
AC-Cookie
Type=0104, AC-Cookie is also a string of random values, different from Host-Uniq, it is
randomly generated by AC, requires Client to respond to the same content; if Client does not
respond, it will be judged as a DoS attacker, and PPPoE will be terminated forcibly negotiate.

12
PPPOE principle

Stage-I ： PPPoE Discover Negotiation

Establish
All Discover negotiation packets are carried on Ethernet, and the type of ETH is 8863

AC 1 c lie n t AC 2

PADI PADI Disc o ve r

PADO PADO Pa ylo a d

Ta g
PPP
PADR
PPPo E
PADS Eth e rn e t (typ e =8863)

13
PPPOE principle
Stage-I ： PPPoE Discover Negotiation
PADI, PPPoE Active Discovery Initiation, always initiated by the client

Since the MAC address of the AC concentrator/server is not known, the destination MAC address of the underlying Ethernet
frame is broadcast; the Session_ID must be set to 0, waiting for the other AC to assign

(Why Client cannot be specified?)

In Payload, common TAGs are Service-Name and Host-Uniq; among them, Service-Name is a mandatory TAG, and the
Length is 0, indicating that any PPPoE service is acceptable

PADO ， PPPoE Active Discovery Offer

After AC receives the PADI request message, if it has available services, it will send a response to CLient through PADO.
Note that the message is unicast, and the Session_ID remains 0.

(Why Session_ID is still 0 in PADO)

In PADO's Payload, common TAGs are Service-Name and AC-Name, both of which are mandatory

Note:

1. AC-Name is provided to the Client to "select"; 14

2. 2. In PADO, multiple Service-Name TAGs are usually included to improve the "hit rate" of the Offer.
PPPOE principle
Stage-I ： PPPoE Discover Negotiation
PADR, PPPoe Active Discovery Seesion-confirmation.

Since PADI messages are broadcasted, the Client may receive multiple Offers at the same time; since PPPoE needs to establish a virtual point-to-
point connection, one of the multiple Offers must be selected for subsequent negotiation.

The client is selected based on the AC-Name or the Service provided by the AC, and the implementation of each manufacturer may be different.

PADR will be unicast to the selected AC. Note: The Session_ID is still 0, which is designated.

In the Tag of PADR, at least one Service-Name must be included to notify the service selected by the Client; only one Service-Name TAG can be
included.

PADS ， PPPoe Active Discovery Seesion-confirmation 。

After AC receives the PADR, it will confirm the PPPoE connection, and prepare for the subsequent PPP connection

AC generates a globally unique Session_ID for the Client, and unicasts it to the Client in the PADS message; it must include the Service-Name
selected by the Bearer Client in the TAG for confirmation.

After that, AC will enable a logical port and trigger PPP negotiation

If the requested Service-Name is abnormal in the received PADR, the AC still responds to the PADS, but the Session_ID is forced to be 0, and the
TAG is Service-Name-error
15
PPPOE principle
Stage-I ： PPPoE Discover Negotiation
Among the four packets in the PPPoE discovery phase, establish a virtual point-to-point connection through negotiation and interaction of at least four packets. Among them,
the first two messages are mainly used to select the AC (server), and the last two messages are for application.

The above process is very similar to the DHCP protocol.

Thinking: What are the main differences between the discovery phase of PPPoE and the discovery of the DHCP protocol?

Unlink

The process of unlinking is relatively simple. When the connection needs to be terminated actively or an abnormal situation occurs, the Client or AC can initiate a PADT
message to forcefully terminate the connection.

PADT unicasts to the other host through Ethernet, and the Session_ID must be specified, that is, the number of the PPPoE virtual point-to-point that needs to be
terminated

There is no need to fill any TAGs in the PADT message.

The initiator of disconnecting the link needs to clear the relevant content of PPPoE after sending the PADT message, and at the same time terminate the transmission of
the upper layer PPP message

After the other party receives the PADT message, it also needs to terminate the transmission of any PPP message, and set the corresponding PPPoE Session (virtual
interface) to down.

AC c lie n t
PADT

16
PPPOE principle

Stage-II ： PPP Negotiation

After the PPPoE Discover negotiation, a point-to-point link is established on the broadcast
Ethernet, which is determined by two MAC addresses and a unique Session_ID.
Stage-II is the negotiation of PPP. The PPP is no different from the PPP in the actual serial
communication.
PPP message is carried on the PPPoE header, and the PPPoE is carried on the Ethernet header. At
this time, the ETH type is 8864.
DATA

Pa ylo a d
Ta g
PPP
PPPo E
Eth e rn e t (typ e =8864)
17
PPPOE principle

Stage-II ： PPP Negotiation

Brainstorming:
we did not find that PPPoE has a link maintenance mechanism, such as the sending of Hello
packets. So in PPPoE applications, how to ensure that the Client and AC can know whether there
is a link fault?

18
PPPOE application
PPPOE application

PPPoE networking topology

Topology
In a PPPoE network, at least one AC and one client are required, but there are usually
many clients. Sometimes there will be multiple ACs for traffic load sharing.

internet

AC

…
PC Router Router 20
Exercises
Exercises

PPPoE configuration
Command:
vpdn enable //Enable PPPoE function
vpdn-group <number> // Create PPPPoE group
protocol pppoe //Specify the protocol
{request-dialin | accept-dialin} //Specify as Client or AC
port ｛ vt |vn <number> ｝ //Call the upper layer virtual PPP interface
pppoe bind interface <type>/<number>//bind physical Ethernet interface
Command:
show pppoe session
debug pppoe event
debug pppoe packet
debug pppoe error
Usually, the first two commands are used more.

22
Exercises
PPPoE configuration (continued)
Note:
VN is virtual-tunnel, which is a virtual PPPoE interface created by the client to serve for PPP
negotiation;
VT is virtual-template, not a logical interface, but just a template; called by the AC server. Every time
the Client successfully dials in, a virtual interface is generated according to this template, called
Virtual-access, referred to as VA.
When the client unlinks, the corresponding va interface is also deleted immediately
For the AC side of PPPoE, its virtual template VT is not an "interface", and the real virtual interface
VA is unstable, so it cannot be directly used for routing;
In view of the above problems, for the PPPoE server side, when setting the route, either next-hop is
a determined IP address, or use dynamic routes such as RIP and OSPF to learn dynamically.

23
PPPOE experiment

PPPoE experiment
Connect the devices according to the topolog shown in the figure. Two routers act as the AC and Client of
PPPoE respectively. After the dial-up is completed, the Client obtain the IP in the AC address pool. On this
basis, it is required to realize intercommunication between the PC and the Server ( two methods)
Change the topology, in which PC acts as a PPPoE client to dial up directly. What is the difference in the
PPPoE negotiation process? How is the routing setup different?

Server
Client AC

PC 24
Thanks
Copyright ©Shanghai Baud Data Communication Co., LTD.