bitmindframes

Wide Area Networks and FRAME RELAY
Frame Realy is the #1 WAN technology worldwide.
Used by all sizes of customers because the Total Cost of Ownership is flexible.
Frame Relay reduces costs by using less equipment, less complexity, & easier implementation than leased lines. It also provides greater reliability, & resilience.

COST EFFECTIVENESS
Frame Relay solutions provide cost effectiveness & flexibility.
With dedicated lines, customers pay for an end to end connection including the local loop & network link.
1.With frame relay you pay for the local loop & for bandwidth.
The Distance between nodes isn’t important.
Lease lines - increments of 64kbs Frame Relay VC(Virtual Connection) are defined with greater granularity, increments as small as 4kbs can be used.
2.Frame Relay is also cost effective in that it shares bandwidth across a larger base of customers.
Using dedicated lines require more DSU/CSU & more complicated routing & switching

Frame Relay flexibility
VC’s provide considerable flex in a network design.
For any site to communicate to any other site , they connect to a VC leading to another site.
The end of each VC is identified by a DLCI (Data Link Connection Identifier).
Leased lines are much more complicated & require much more equipment.
Frame Relay customers pay only for the bandwidth they need.
ISDN calls are metered & can result in unexpectedly high monthly charges.

THE Frame Relay WAN
X.25 provided reliability over unreliable infrastructure.
Frame Relay has lower overhead because it has fewer capabilities.
Frame Relay combines some functions of L2 & L3 into simple protocols.
Frame Relay provides multi logical connections over a single physical circuit.
Frame Relay operates between router & a network.
The network can us Frame Relay, digital circuit switching or ATM.

Frame Relay OPERATION
The connection between a DTE & DCE consists of both L1 & L2 components.
The L2 component defines the protocol used between router and CO(Central Office)
WAN switches move frames from one DTE across network to another DTE.
Computing equipment not on a lan may send data across a Frame Relay network using FRAD(FRAME RELAY ACCESS DEVICE) as the DTE.

VC’s
With VC’s Frame Relay shares the bandwidth & any site can communicate with any other site with multiple lines.
There are 2 ways to establish VC’s :
SVC’s - establish dynamically with signaling
CALL SETUP > DATA TRANSFER > IDLE > CALL TERMINATION

PVC’s - preconfigured by the carrier.
VC stores input-port to output-port mapping in the memory of each switch & forms a continuous path.

A VC can pass through any number of intermediate devices located within the Frame Relay network.
VC are identified by DLCI’s which are assigned by ISP.
DLCI’s have LOCAL SIGNIFICANCE - not unique in the Frame Relay WAN.
DLCI have no significance beyond a single link.
The 2 ends of a VC may use different DLCI.
DLCI may be duplicated throughout the ISPs network.
Frame Relay labels each frame with a DLCI in the address field.
10 bit field = 1024 possible DLCIs.
DLCIs 0 - 15 * 1008 - 1023 are reserved.
Frames will leave the router using the assigned DLCI, but at each Frame Relay switch, the DLCI value wil be changed.
DLCI are mapped to ports on each Frame Relay switch.
Thus each VC consists of a end to end DLCI port mapping.

MULTIPLE VC’s
A router can have multiple VC’s connecting it to various endpoints.
Multiple VC’s on a single physical line are distinguished because each VC has its own DLCI.
This capability reduces equipment & complexity, making it a very cost effective replacement for a mesh of access lines.
Each endpoint needs only a single access line & interface.
The capacity of the access line is based on the average bandwidth requirement of the VCs, rather than the max.
AT L3 you might have 1 ip address but many DLCIs.
With Frame Relay, customers pay for the bandwidth they use.
In effect they pay for a Frame Relay port.
Increase the # of ports & you pay for more bandwidth, but don’t pay for more equipment because there is no change to the physical infrastructure.

Frame Relay ENCAPSULATION PROCESS
Frame Relay takes data packets from IP & encapsulates them as the data portion of a Frame Relay frame.

The frame is defined by the LAPF(Link Access Procedure for Frame Mode Services) spec.
Like LAPB, LAPM, LAPD thay are all based on HDLC.

Frame Relay HEADER
DLCI - 10 bits DLCI, locally significant VC id.
Extended address EA - allows longer DLCIs in the future.
C/R not used.
Congestion control - 3 flag bits (FECN, BECN, & DE) that control congestion-notification mechanisms.
Frame Relay is a subset of HDLC so it uses (0×7e) flag fields.
FCS helps test for L2 errors (error detection only).
Error control is left to the upper osi layers.

Frame Relay TOPOLOGIES
Cost-effective Frame Relay networks may link hundereds of sites.
Documenting topologies can be very complicated.
However every network segment can be viewed as being one of 3 topology types:
Star (hub & spoke) - a central router acts as a hub & hosts the primary services.
Each remote site accesses the Frame Relay cloud with a single VC.

Full-mesh topology
Connects every site to every other site.
A full mesh is suitable where services are geographically dispersed & highly reliable access to them is required.
Leased lines require additional interfaces & line costs.
Frame Relay allows multi connections simply by configuring additional VCs on an existing link.
This soft upgrade grows the star topology to a full mesh topology without the additional hardware or lines.
Providers charge for additional bandwidth but way less.

Partial mesh topology
A full mesh is seldom affordable because of the number of links required.
Besides hardware cost, there is a limit of < 1,000 VC’s per link.

Web Hosting : Unlimited Websites, Free Domain, Live Support.

Configuring a basic Frame Relay PVC
With up to date routers (IOS) and Frame Relay switches, it is possible to configure Frame Relay with very few settings.

CONFIGURE A BASIC Frame Relay PVC


Step1. Set the ip address on interface

Step2. Configure encapsulation
R1(config-if)#encapsulation frame-relay [cisco |ietf]
Step3. Set bandwidth
R1(config-if)#bandwidth #

Step4. setting LMI type (optional)
R1(config-if)#frame-relay lmi-type cisco | ansi | q933a

Frame Relay ADDRESS MAPPING
Remember that ethernet maps MAC & IP address.
Frame Relay must map DLCI & IP addresses.
This can be done by static or dynamic mapping.
Inverse ARP (inARP) obtains L3 addresses dynamically for Frame Relay & ATM.
Whereas ARP translates IP to MAC, inARP does the opposite DLCI to IP addresses

DYNAMIC MAPPING
Dynamic Mapping uses inverse ARP to resolve a remote IP address to a local DLCI.
A router sends out INARP requests on its PVC.
The router builds & maintains a mapping table, wich contains resolved inARP requests & static mapping entries.
To view this table use:
R1#show frame-relay map

The interface is up: remote IP = 10.1.1.2; local DLCI = 102
Link is using cisco encapsulation as opposed to IETF

STATIC MAPPING
inARP is enabled by default.
A user can override inARP by statically mapping.
You cannot use inARP & static mapping for the same DLCI.
Use static mapping if a remote site doesn’t support inARP.
Can also be used in hub n spoke, on spoke routers to provide spoke-spoke reachability.
Because the spoke routers do not have direct connectivity with each other, dynamic inARP would not work.
Dynamic inARP relies on a direct point-to-point connection.

CONFIGURE STATIC MAPPING
Ip-addr is remote address dlci is local only.
R1(config-if)#frame-relay map ip-addr dlci [broadcast] [ietf] | [cisco]
Use ietf when connecting to non cisco router.
You can greatly simplify ospf configs by adding the optional keyword broadcast.
Static mapping allows users to select the type of Frame Relay encapsulation on a per-VC basis.

Creating a Frame Relay switch
Frame Relay switches are very large expensive dedicated pieces of hardware.
Here we will transform a router into a Frame Relay switch.
R1(config)#frame-relay switching
R1(config) #interface serial0/0/0
R1(config-if)#clock rate 64000
R1(config-if)#encapsulation frame-relay
R1(config-if)#frame-relay intf-type dce
R1(config-if)#frame-relay route 102 interface serial0/0/1 201

LMI (Local Management Interface)
The original Frame Relay design did not include the ability for DTE’s to find the status of the Frame Relay network.
A consortium of CISCO, DEC,NORTEL & stratacom extended Frame Relay to provide additional capabilities.
These extensions are referred to collectively as the LMI.
LMI is a keepalive mechanism that provides status info about Frame Relay connections between the DTE & the Frame Relay switch.
Every 10 secs the DTE polls the network & if it gets no response may consider the connection to be down.
LMI responses include info on all DLCIs allocated to a line.

LMI FRAME FORMAT
LMI messages are carried in a variant of LAPF frames.
The address field carries one of the reserved dlci’s

LMI EXTENSIONS
It is easy to confuse the LMI and encapsulation.
LMI’s are used between the DTE & the edge of the cloud(frame relay switch).
Encapsulation defines the frames between 2 DTE’s.
The switch & its router need to use the same LMI type.
LMI include some optional extensions.
VC status messages - help synchronize devices.
Multicasting - supports IP multicasting concepts.
Global addressing - lets FR network resemble a LAN.
Simple flow control - for devices that cant use Explicit Congestion Notification.

LMI extensions reserve some dlci’s for messages between the dte & dce
There are several LMI types:
CISCO - original LMI extension DLCI 1023
ANSI DLCI 0
Q933a DLCI 0
LMI type on a router must match DCE.

LMI autosense will usually configure this, but if not, use :
R1(config-if)#frame-relay lmi-type cisco | ansi | q933a

Configuring the LMI type, disables the autosense feature.
Set a keepalive interval to prevent the switch from timing out.
A large mismatch in the keepalive interval on the router & the switch can cause the switch to declare the router dead.
By default the keepalives are 10 sec
R1(config-if)#keepalive interval

LMI & Inverse ARP
LMI together with inARP map L2 & L3 addresses.
When R1 connects to a Frame Relayswitch, it sends an LMI inquiry ,The switch replies with details of every VC on the link.
Periodicly the router repeats the status inquiry.
If the router needs to map the VC’s it sends an Inarp message to the VC.
The inArp message includes the IP address of the router so the remote DTE or router can also perform mapping.

The broadcast keyword
Frame Relay, ATM & x.25 are NBMA (Non Broadcast Multi Access) networks.
NBMA networks do not support multicast/broadcast traffic.
Broadcast functionality is achieved by replicating packets manually to destinations, takes a lot of resources.
Some routing protocols like EIGRP, OSPF require additional configs to support NBMA networks.
The broadcast keyword allows broadcasts & multicasts over the PVC & turn broadcast into many unicasts
To verify FR mapping use :
R1# show frame-relay map

Paying for frame relay
Key terminology
Customers are not exposed to the inner workings of the Frame Relay network wich may be built on T1, T3, SONET or ATM.
Before considering how to pay for the Frame Relay services there are some key terms & concpets to learn :
ACCESS RATE or PORT SPEED - the speed of the physical line connecting the customer to the cloud (access circuit).
The rate at which you access circuits join the Frame Relay network.
These are typically at 56kbs, T1, or fractional T1.
It isn’t possible to send data faster than access rate.

CIR (committed information rate) - Customers & prividers negotiate CIR for each PVC

The amount of data the network receives from access circuit
The provider guarantees the CIR.
A great advantage of Frame Relay is that unused capacity is made available, usually at no extra cost.
This allows customers to burst over their CIR as a bonus.
There are 3 Frame Relay cost components
ACCESS LINE : T1, T3 or fractional T1 connecting to cloud
PVC : setup cost for VC
CIR : usually lower than access rate > allows for bursts

OVERSUBSCRIPTION
Providers sometimes sell more capacity than they can handle.
Assume not everyone demands CIR simultaneously.
This can cause congestion.

BURSTING
If a Frame Relay network has excess capacity inside the cloud customers may use it.
R1 - S0/0/0 access rate = 64k CIR = 32k + 16k =48k
These 2 VC’s could burst up to 16 k extra.
Bursting allows additional bandwidth to be borrowed.
The burst duration should be less than 3 or 4 sec.
Related terms: Committed Burst Info Rate (CBIR) & Excess Burst Size (BE)
The CIBR is a negotiated rate above the CIR.
Can burst up to the CIBR & expect data to get through.
If a long burst persists > purchase higher CIR.
Frames submitted above the CIR are marked as Discard eligeable (DE).
DE frames are dropped if congestion/capacity dictates.
BE = bandwidth above the CIBR.
These frames are usually dropped.

FLOW CONTROL
Frame Relay reduces overhead by implementing simple mechanisms.
These congestion-notification mechanisms are in the Forward Explicit Congestion Notification (FECN) & BECN
Each are controlled by a single bit in the frame header.
The header also contains a Discard Eligibility Bit.
They let a router know if there is congestion & that transmission should slow until the condition is reversed.
BECN is a direct notification FECN is an indirect one.

In periods of congestion, Frame Relay switches apply logic rules to each frame based on whether the CIR is exceeded.
*if incoming frame does not exceed CBIR pass it on
*if incoming frame exceeds CIBR, mark it DE
*if incoming frame exceeds the CIBR + BE discard it
Frames arriving at a switch are queued prior to forwarding.
It is possible for buffers to fill causing delays.
To reduce this the switch notifies DTE’s using ECN.
The FECN bit is set on every frame received.
The BECN bit set on every frame that the switch places onto the congested link.

CONFIGURE Frame Relay SUB-interfaces
Subinterfaces provide a way to subdivide a partial mesh into many smaller fully meshed PtoP subnetworks.
Each subnetwork appears as if it were reachable through a separate interface.
PtoP sub interfaces can be unnumbered for use with IP reducing the addressing burden.
Creating a subinterface, use DLCI as subinterface #
For point-to-point then the DLCI must also be configured
For multipint without an inARP, the DLCI must also be configured.
R1(config-if)#frame-relay interface-dlci #

For multiport w/ static map, no dlci required.
Altering an existing subinterface config may fail to provide the expected results.
Reload may be necessary (after saving).

Tags: 802.1q, router, guides, vlan, wan, ip