Some lab notes for Dynamic Trunking Protocol

Dynamic Trunking Protocol (DTP) Notes

Effect of 'switchport mode access' on DTP

After disabling DTP on all other ports, using 'switchport nonegotiate' and enabling 'debug dtp packets' I started investigating the effect of different port settings on DTP. I had been reading some discussion about whether an access port would still send out some DTP packets even after being turned into an access port using the 'switchport mode access' command.

So I put the port into dynamic desirable mode on both ends, successfully established a trunk, and then set one end as an access port.

Here are the results:

DLS2(config-if)#switchport mode access

DLS2(config-if)#

00:43:43: DTP-pkt:Fa0/5:Sending packet ../dyntrk/dyntrk_process.c:1241

00:43:43: DTP-pkt:Fa0/5: TOS/TAS = ACCESS/OFF ../dyntrk/dyntrk_process.c:1244

00:43:43: DTP-pkt:Fa0/5: TOT/TAT = ISL/NEGOTIATE ../dyntrk/dyntrk_process.c:1247

00:43:43: DTP-pkt:Fa0/5:datagramout ../dyntrk/dyntrkprocess.c:1279

00:43:43: DTP-pkt:Fa0/5:Invalid TLV (type 0, len 0) in received packet. ../dyntrk/dyntrk_core.c:1334

00:43:43: DTP-pkt:Fa0/5:Good DTP packet received: ../dyntrk/dyntrk_core.c:1500

00:43:43: DTP-pkt:Fa0/5: Domain: ../dyntrk/dyntrk_core.c:1503

00:43:43: DTP-pkt:Fa0/5: Status: TOS/TAS = ACCESS/DESIRABLE ../dyntrk/dyntrk_core.c:1506

00:43:43: DTP-pkt:Fa0/5: Type: TOT/TAT = ISL/NEGOTIATED ../dyntrk/dyntrk_core.c:1508

00:43:43: DTP-pkt:Fa0/5: ID: 000F90236585 ../dyntrk/dyntrk_core.c:1511

So we can see that only one final DTP packet is sent and received to advise that the port has been placed in Access mode. It then ignores any further DTP packets, even though I can see them still being sent from the other end if I disable and enable DTP by putting the port on the other end into access mode, then back to dynamic desirable.

'switchport nonegotiate' limitations

'switchport nonegotiate' cannot be configured on a port already configured as a DTP trunk i.e. dynamic desirable or dynamic auto. It doesn't just switch DTP off on the port; you would have to place the port into 'switchport mode access' or 'switchport mode trunk' first.

Trunk encapsulation negotiation

Manually setting encapsulation on one end of the link

When DTP is used to negotiate encapsulation ('switchport trunk encapsulation negotiate'), which is default, then the trunk will be negotiated, if both switches support it, as

1. ISL, then
2. 802.1q, if ISL is not supported by both switches.

However, even between two switches that support ISL, if encapsulation is set manually, using 'switchport trunk encapsulation isl | dot1q', at only one end, then DTP will negotiate that encapsulation on the link.

Limitations on the 'switchport mode trunk' command

The 'switchport mode trunk' command is used to manually set a link to always be a trunk. DTP packets are still sent out of the interface, so a trunk could still be formed with an 'active' DTP port.

However, the 'switchport mode trunk' command cannot be applied if encapsulation is negotiated. The encapsulation must be set manually.

DLS1(config-if)#switchport mode trunk

Command rejected: An interface whose trunk encapsulation is "Auto" can not be configured to "trunk" mode.

The error message is slightly misleading, referring to "Auto" encapsulation. This confused me the first time I saw it, until I realised it was referring to 'switchport trunk encapsulation negotiate' i.e. negotiated encapsulation. It would be great if Cisco kept their error messages consistent with their command syntax!

Multi-Layer Switch: routed port, switchport and SVIs

'switchport'

The 'switchport' command tells the switch (usually a Multi-Layer Switch or MLS) to treat the port as a layer 2 port, i.e. as a member of a VLAN and to allow it to switch frames and learn MAC addresses etc., as well as participating in all other layer 2 processes such as spanning-tree.

'no switchport'

The 'no switchport' command tells the switch to treat the port as a layer 3 interface, so that you can run a routing protocol, add an interface IP address (or other layer 3 address) and create sub-interfaces, none of which is possible on a layer 2 interface. If you try running this command on a layer 2 only switch (e.g. a 2950) it will not understand it and reject it as 'incomplete', as shown below:

ALS1#conf t
Enter configuration commands, one per line. End with CNTL/Z.
ALS1(config-if)#no switchport
% Incomplete command.

A routed port does not belong to a VLAN as far as the MLS is concerned because it has no concept of VLANs at layer 3, just a like a port on a router. However, on a MLS each VLAN also has a layer 3 interface: the VLAN interface, also known as an SVI. This is created on an MLS when the VLAN itself is created.

On a pure layer 2 switch, such as the 2950, there is only one layer 3 interface: this is the 'VLAN1' interface (an SVI) that you configure to allow management connectivity.

ALS1#show run int vlan 1
Building configuration...
Current configuration : 67 bytes
!
interface Vlan1
no ip address
no ip route-cache
shutdown
end

Private VLAN summary

Private VLANs

Allows for the separation of ports into private port groups, while still making use of the same subnet. This is more efficient in terms of IP addressing usage and STP and ACL complexity and of particular use in some shared environments such as Service Provider (SP) data centres where access to common resources on a subnet are required in a secure way.
There are essentially three different port classifications in terms of function. Ports that need to communicate with:

• all devices
• each other and with shared devices (e.g. router or web server)
• ONLY shared devices

Private VLANs are constructed so that there exists a primary VLAN, and one or more secondary VLANs. Each secondary VLAN is mapped to a primary VLAN.

Private VLANs are only supported by VTPv3, so "VTP transparent" mode should be configured if not using VTPv3.

Primary VLANs

Contains promiscuous ports i.e. can send and receive to any other port in the PVLAN including those assigned to secondary VLANs. Devices in this VLAN are likely to include the router L3 gateway, web servers, database servers etc.

Secondary VLANs

Are one of two types Community or Isolated

Community VLANs

• Ports can talk to other ports in the community and to primary VLAN (promiscuous) ports
• Each PVLAN has zero or more community VLANs associated with it.

Isolated VLANs

• Ports can ONLY talk to primary VLAN (promiscuous) ports
• Each PVLAN has AT MOST ONE isolated VLAN, since only one is required

Private VLAN trunks

Extending Private VLANs across multiple switches is a simple matter; simply use the same VLAN IDs and trunk the VLANs as you would normally. Frames arriving from a port within a Private VLAN (primary or secondary) are tagged with the primary or secondary VLAN tag for transport between switches.

However, there are two special trunk types that are used with Private VLANs:

Promiscuous PVLAN Trunk

This is used when a trunk is carrying traffic for a Primary VLAN, as well as its associated secondaries, and needs to be considered a promiscuous port. It may also be carrying other normal VLANs. In this case, the device on the other end of the trunk is unaware of the relationship between the Private VLANs, and traffic from all secondary VLANs associated with a Primary VLAN is tagged with the Primary VLAN ID. A use case for this scenarios is a "router on a stick" configuration where the gateway interface of a Primary VLAN (on he router) is considered promiscuous and allowed to be communicated with my all associated Secondary VLANs.

The Promiscuous PVLAN Trunk port re-writes secondary VLAN IDs of sent frames into the corresponding primary VLAN ID so that the external device always sees only the primary VLAN. It does not manipulate tags of incoming frames.

Isolated PVLAN Trunk

This is used to extend the isolated VLAN over a trunk carrying multiple VLANs to a switch that does not support Private VLANs but is capable of isolating its own ports e.g. with the port protection feature on entry-level Catalyst switches.

The Isolated PVLAN Trunk re-writes a primary VLAN ID of a sent frame to the ID of the isolated VLAN that is associated with the primary VLAN. It does not manipulate tags of incoming frames.

Dangers of Tunnel vision

I gave myself a real fright today.

I was trying to troubleshoot an issue with a configuration that I've never done before; IPv6 ISIS. It was while I was on a training course, I was trying to do some route summarisation. and I struggled with the lab for about 45 minutes, taking out configuration, checking the routing table, putting it back in, checking it again. The topology was about as simple as you can get; two routers. And I checked and double-checked my configuration and found a little error and that made a difference but not the one I wanted.

Over and over and over again I looked at the routing table.

Except...

That is wasn't the full routing table. I was just looking at the ISIS routes. And as soon as the instructor made one little suggestion I looked at the WHOLE routing table, and there was the route... advertised by another protocol with a lower AD! Hidden in the trees, was the wood!

It turned my blood to ice to realise how easily I fell into the tunnel-vision trap - forgetting to take a pause and a step back; to think about alternatives. So that's why I'm writing about it today. If I read this again in 6-9 months' time I can remember today, and the fear I felt when I thought about how easily that could have happened to me in a lab exam, and how I would have totally blown it.

A long way to go, but food for thought.

RJ45 pinout diagram and copper Ethernet cabling

One of the most fundamental things in networking is the construction of a straight-through or cross-over copper cable.

Copper Ethernet cabling is typically constructed with Category-5 (a.k.a. Cat-5), Cat 5e or Cat 6, each of which are rated for higher bandwidth capability than the previous. They are made up of four pairs of copper wires, each insulated then tightly and precisely twisted together (leading to the term Unshielded Twisted Pair or UTP) to make the impedance characteristics (and hence the behaviour at very high frequencies) of the cable very precisely known.

The pin positions of the pairs in the T568A and T568B connectors (RJ-45) are shown below, as well as 'crossover' cable connections. I found it useful to remember that pairs 1 and 4 never change position, no matter what cable type, and that pairs 2 and 3 swap position in a cross-over. Regardless of the connector used at either end, this is easy to remember, and leads to the well-known connection pattern, of a crossover cable, of 1-3, 2-6, 3-1, 6-2.

Straight-through is self-explanatory,where the same connector type is used on both ends (either T568A or T568B), and the pin positions of pairs 1-4 do not change.

In terms of pair numbering within the connector, I try to remember that pairs are numbered outward from the centre toward the 'top' (i.e. the lower numbered pins) for T568A i.e. 1,2,3,3 and from the centre to the 'bottom' 1,2,4,4.

Ethernet framing summary (taken from the 802.3-2012 standard)

I realised that I didn't really understand the construction and transmission of the Ethernet frame very well, even after reading the Cisco texts, so I thought I would dig into the 802.3 standard. It's actually quite readable! All emphasis is mine.

ETHERNET FRAMING SUMMARY

Taken from 802.3-2012 standard.

Preamble

Section 4.2.5 Preamble generation
Upon request by TransmitLinkMgmt to transmit the first bit of a new frame, PhysicalSignalEncap shall first transmit the preamble, a bit sequence used for physical medium stabilization and synchronization, followed by the Start Frame Delimiter. If, while transmitting the preamble or Start Frame Delimiter, the collision detect variable becomes true, any remaining preamble and Start Frame Delimiter bits shall be sent. The preamble pattern is:

10101010 10101010 10101010 10101010 10101010 10101010 10101010

The bits are transmitted in order, from left to right. The nature of the pattern is such that, for Manchester encoding, it appears as a periodic waveform on the medium that enables bit synchronization. It should be noted that the preamble ends with a “0.”

Start Frame Delimiter

4.2.6 Start frame sequence
The receiveDataValid signal is the indication to the MAC that the frame reception process should begin. Upon reception of the sequence 10101011 following the assertion of receiveDataValid, PhysicalSignalDecap shall begin passing successive bits to ReceiveLinkMgmt for passing to the MAC client.

Address fields

3.2.3 Address fields
Each MAC frame shall contain two address fields: the Destination Address field and the Source Address field, in that order.
The Destination Address field shall specify the destination addressee(s) for which the MAC frame is intended.
The Source Address field shall identify the station from which the MAC frame was initiated.
The representation of each address field shall be as follows:
a) Each address field shall be 48 bits in length.
b) The first bit (LSB) shall be used in the Destination Address field as an address type designation bit [I/G bit] to identify the Destination Address either as an individual or as a group address. If this bit is 0, it shall indicate that the address field contains an individual address. If this bit is 1, it shall indicate that the address field contains a group address that identifies none, one or more, or all of the stations connected to the LAN. In the Source Address field, the first bit is reserved and set to 0.
c) The second bit shall be used to distinguish between locally or globally administered addresses [U/L bit]. For globally administered (or U, universal) addresses, the bit is set to 0. If an address is to be assigned locally, this bit shall be set to 1. Note that for the broadcast address, this bit is also a 1.
d) Each octet of each address field shall be transmitted least significant bit first.
3.2.3.1 Address designation
A MAC sublayer address is one of two types:
a) Individual Address. The address associated with a particular station on the network.
b) Group Address. A multidestination address, associated with one or more stations on a given network. There are two kinds of multicast addresses:

1. Multicast-Group Address. An address associated by higher-level convention with a group of logically related stations.
2. Broadcast Address. A distinguished, predefined multicast address that always denotes the set of all stations on a given LAN.

All 1’s in the Destination Address field shall be predefined to be the Broadcast Address.
This group shall be predefined for each communication medium to consist of all stations actively connected to that medium; it shall be used to broadcast to all the active stations on that medium. All stations shall be able to recognize the Broadcast Address. It is not necessary that a station be capable of generating the Broadcast Address.
The address space shall also be partitioned into locally administered and globally administered addresses. The nature of a body and the procedures by which it administers these global (U) addresses is beyond the scope of this standard. [IEEE]

Destination Address field

3.2.4 Destination Address field
The Destination Address field specifies the station(s) for which the MAC frame is intended. It may be an individual or multicast (including broadcast) address.

Source Address field

3.2.5 Source Address field
The Source Address field specifies the station sending the MAC frame. The Source Address field is not interpreted by the MAC sublayer.

Length / Type field

3.2.6 Length/Type field
This two-octet field takes one of two meanings, depending on its numeric value. For numerical evaluation, the first octet is the most significant octet of this field.
a) If the value of this field is less than or equal to 1500 decimal (05DC hexadecimal), then the Length/Type field indicates the number of MAC client data octets contained in the subsequent MAC Client Data field of the basic frame (Length interpretation).
b) If the value of this field is greater than or equal to 1536 decimal (0600 hexadecimal), then the Length/Type field indicates the Ethertype of the MAC client protocol (Type interpretation).[IEEE]
The Length and Type interpretations of this field are mutually exclusive.
When used as a Type field, it is the responsibility of the MAC client to ensure that the MAC client operates properly when the MAC sublayer pads the supplied MAC Client data, as discussed in 3.2.7. Regardless of the interpretation of the Length/Type field, if the length of the MAC Client Data field is less than the minimum required for proper operation of the protocol, a Pad field (a sequence of octets) will be added after the MAC Client Data field but prior to the FCS field, specified below. The procedure that determines the size of the Pad field is specified in 4.2.8.
The Length/Type field is transmitted and received with the high order octet first.

MAC Client Data field

3.2.7 MAC Client Data field
The MAC Client Data field contains a sequence of octets. Full data transparency is provided in the sense that any arbitrary sequence of octet values may appear in the MAC Client Data field up to a maximum field length determined by the particular implementation.
Ethernet implementations shall support at least one of three maximum MAC Client Data field sizes defined as follows:
a) 1500 decimal—basic frames (see 1.4.102)
b) 1504 decimal—Q-tagged frames (see 1.4.334)
c) 1982 decimal—envelope frames (see 1.4.184)
If layer management is implemented, frames with a MAC Client Data field larger than the supported maximum MAC Client Data field size are counted. It is recommended that new implementations support the transmission and reception of envelope frames, item c) above.
NOTE 1—The envelope frame is intended to allow inclusion of additional prefixes and suffixes required by higher layer encapsulation protocols (see 1.4.180) such as those defined by the IEEE 802.1 working group (such as Provider Bridges and MAC Security), ITU-T or IETF (such as MPLS). The original MAC Client Data field maximum remains 1500 octets while the encapsulation protocols may add up to an additional 482 octets. Use of these extra octets for other purposes is not recommended, and may result in MAC frames being dropped or corrupted as they may violate maximum MAC frame size restrictions if encapsulation protocols are required to operate on them.
NOTE 2—All IEEE 802.3 MAC frames share a common format. The processing of the three types of MAC frames is not differentiated within the IEEE 802.3 MAC, except for management. However, they may be distinguished within the MAC client.
NOTE 3—All Q-tagged frames are envelope frames, but not all envelope frames are Q-tagged frames.
See 4.4 for a discussion of MAC parameters; see 4.2.3.3 for a discussion of the minimum frame size and minFrameSize.

Pad field

3.2.8 Pad field
A minimum MAC frame size is required for correct CSMA/CD protocol operation (see 4.2.3.3 and 4.4). If necessary, a Pad field (in units of octets) is appended after the MAC Client Data field prior to calculating and appending the FCS field. The size of the Pad, if any, is determined by the size of the MAC Client Data field supplied by the MAC client and the minimum MAC frame size and address size MAC parameters (see 4.4).

FCS field

3.2.9 Frame Check Sequence (FCS) field
A cyclic redundancy check (CRC) is used by the transmit and receive algorithms to generate a CRC value for the FCS field. The FCS field contains a 4-octet (32-bit) CRC value. This value is computed as a function of the contents of the protected fields of the MAC frame: the Destination Address, Source Address, Length/ Type field, MAC Client Data, and Pad (that is, all fields except FCS).

Preparing for the CCIE Routing and Switching Written Exam v5 (400-101)

After about 10 years or so in Cisco networking, the last three of which have been dedicated to network design and implementation, and having earned my CCDA and CCNA certifications, lapsed my CCDA, and re-earned, and now lapsed again(!) my CCNA, I decided that I should just skip my CCNP certification path and jump straight into CCIE.

This was prompted by the experience and advice of many of my colleagues who have already earned the coveted CCIE certification in either R&S or Service Provider. They said that the effort (and expense) involved in studying and passing the three CCNP exams would be better spent studying for my CCIE which, strangely, requires no pre-requisite qualifications to sit. When I began my networking journey I had told myself that I wasn't really interested in the time and effort it would take to get my CCIE, but the longer I am in this field, the more I realise I want to prove to myself that I still have it in me to work really hard at a very difficult task and attempt to overcome the inevitable issues that will arise in order to succeed. In a way, I am viewing this as my PhD, or at least a pinnacle of achievement in my engineering career of which I can be proud and which will help me in terms of better knowledge and understanding of the details of my chosen field, and for the kudos such an achievement will bring.

I started last year, but was seriously derailed by the death of my father and the need to travel to my family's home in New Zealand for the funeral and to help out my mum. Now, with that substantially taken care of, and with my own emotional energy returning, I decided once again to set myself the task of climbing the "Everest" of Internet Protocol networking; the CCIE.

My goal is to be prepared for the written exam in February of 2016. That gives me 6 months to study and prepare, and then one month to revise and cram exam technique for the exam itself. After that, of course, it's on to the infamous lab exam, which has left many a broken engineer in its wake! It is, frankly, terrifying, and I don't like to fail. But this journey will teach me, I'm certain, that failure is just one of the stones on the path to success.

I am based in and around Reading, UK, and would welcome any local study groups to get in touch.

19 Jul 2015: Updated with reminders of why I'm doing this.

Why become a CCIE?

Rarity: Fewer than 1% of all networking professionals hold a CCIE

Knowledge: Passing the exam is a by-product of being an expert.

CCIE is a challenge: It's about the journey, not about passing the exam.