Dynamic Router Configuration Protocol (DRCP)
Uploaded by
ijsretDynamic Router Configuration Protocol (DRCP)
Uploaded by
ijsretInternational Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 3, Issue 2, May 2014
Dynamic Router Configuration Protocol (DRCP)
Arif Tanvir
Lecturer, Department of CSE, Asian University of Bangladesh (AUB), Dhaka, Bangladesh
Abstract
Dynamic behavior of the network is all what we want
because that reduces our headache of managing the
network and we would rather expect the network to
manage itself. Here in this paper, I am daring to propose
the idea of DRCP along with the concept of Generalized
Router Discovery Protocol (GRDP) and Auto Subnetting
to automate the whole process of IP Allocation among
the current routers in the world as well as the ones in the
queue to join the network. Five Regional Internet
Registries APNIC, ARIN, RIPE NCC, LACNIC and
AFRINIC will each be associated with five Master Zone
Routers which will be in charge of allocating IPs to the
routers as they enter into the network. This may sound a
little dreamy at the moment, but you know, someone has
to dream. Why cant that be me? I will try to explain all
the concepts and ideas as clearly as possible in this
paper.
Keywords - Dynamic Router Configuration, Router
Discovery Protocol, Automatic IP Allocation, Auto
Subnetting, Regional Internet Registries, Dynamic Host
Configuration Protocol, Cisco Discovery Protocol,
OSPF Routing Protocol, Backbone Area Border Router
(BABR), Master Area Border Router (MABR), Sub Area
Border Router (SABR).
1. Introduction
The concept of DHCP has been used as the backbone for
designing the working mechanism of DRCP. Exchange
of messages between Master Zone Routers (MZR) and
the Client Routers (CR) of that zone controlled by it has also been influenced by the messages exchanged
between DHCP Server & DHCP Client. The concept of
Cisco Neighbor Discovery Protocol has been used, as
well as extended to the level which allows detection of
devices at Layer 3 instead of just Layer 2, with the use
of GRE Tunnel and enabling GRDP on the tunnel
interface. Obviously allocation of IP to the router will
involve subnetting and we will leave that to auto
subnetting algorithm which will deal with the IP Blocks
available at each zone and breaking that down to the
level as requested by the routers.
2. IP & Subnet Configuration
DRCP supports 2 Mechanisms for Allocating IP address:
Manual Allocation: Network Administrator of the
MZRs manually assigns one or more IP Addresses to
Router(s) of that Zone along with appropriate Subnet
Mask considering the direct request of IP Requirements
from the Client.
Dynamic Allocation: DRCP assigns Permanent IP
Address(es) along with Subnet Mask with the use of
Auto Subnetting Algorithm to a Client considering the
Number of IPs requested by the client.
3. Related Terms
GRDP: General Router Discovery Protocol
DRCP MZR: It assigns IP addresses and other network
configuration parameters to the DRCP CRs
DRCP CR: It obtains an IP address and other network
configuration parameters from a DRCP MZR.
DRCP BABR: Backbone Area Border Router.
DRCP MABR: Master Area Border Router.
DRCP SABR: Sub Area Border Router.
4. Dynamic IP Allocation Process
A DRCP CR obtains a valid IP address from a DRCP
MZR in Three Steps:
Discovering: The CR locates a MZR.
Offering: The MZR offers configuration parameters
such as IP Address & Subnet Mask to the CR.
Acknowledging: CR confirms the offered configuration.
www.ijsret.org
271
International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 3, Issue 2, May 2014
will be added to the Address Pool of the MZR of that
particular zone.
6. General
(GRDP)
Fig.1 Dynamic IP Allocation Process
Fig.1 illustrates the dynamic IP address allocation steps
which are described as follows:
Discovering:
The DRCP CR broadcasts a DRCP-DISCOVER
message to locate a DRCP MZR. The DRCP MZR for
that Zone sends reply to the CR upon receiving the
DRCP-DISCOVER message.
Offering:
A DRCP MZR that receives the DRCP-DISCOVER
message offers an appropriate IP address & Subnet Mask
to the CR in a DRCP-OFFER message.
When assigning an IP address & Subnet Mask to a CR,
the MZR needs to confirm that the IP address to be
assigned is not used by any other device by sending
ICMP Echo Request (ping) packets. If the server does
not get any response within a specified period, it will
ping the IP address once again until a specified number
of ping packets are sent. If still no response is received,
the server will assign the IP address to the requesting
CR; otherwise, the IP address will be marked as a
conflicting IP address, and the MZR will choose another
IP address.
Acknowledging:
The DRCP CR returns a DRCP-ACK message to DRCP
MZR, confirming that it is starting to use the allocated
IP address.
5. Releasing an IP Address
When a DRCP CR no more need to use its IP address
anymore, it sends a DRCP-RELEASE message to notify
the DRCP MZR to release the IP address. That address
Router
Discovery
Protocol
GRDP is a generalized version of CDP where the routers
are not meant to be vendor specific and it runs on all
media that supports Subnetwork Access Protocol
(SNAP).
Use of SNMP (Simple Network Management
Protocol) with GRDP MIB (Management Information
Database) allows network management applications to
learn - Device type & SNMP Agent Address of
neighboring devices and to send SNMP queries to those
devices.
But unlike CDP (Cisco Discovery Protocol) which
runs over Data Link Layer only, GRDP can run over
Network Layer as well. Therefore, two systems that
support different network-layer protocols can learn about
each other. As with CDP each device configured sends
periodic messages (Known as Advertisements) to a
multicast address. In case of GRDP this period should be
much longer to reduce the network traffic. And the
router that is supposed to go offline must react instantly
rather than waiting for the timer for the advertisement to
expire. But there must be different periodic calculation
for Zonal Routers.
Each device advertises at least one address - At
which it can receive SNMP messages. The
Advertisements also contain time-to-live (holdtime),
which indicates the length of time a receiving device
should hold GRDP information before discarding it.
Each device also listens to the periodic GRDP messages
sent by other devices in that particular zone in order to
learn about neighboring devices and determine when
their interfaces to the media go up or down.
Here we are borrowing the concept of GRE (Generic
Routing Encapsulation) Tunneling protocol developed
by Cisco that can encapsulate a wide variety of protocol
packet types inside IP Tunnels, creating a virtual pointto-point link to Routers at remote points over an IP
internetwork.
Normally CDP requires a layer 2 connection to
establish the neighbor adjacency. But there are some
tricks to do that across a WAN using a GRE tunnel and
enabling CDP on the tunnel interface. But generally a
WAN link breaks a CDP neighbor chain.
www.ijsret.org
272
International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 3, Issue 2, May 2014
Let us now look at different Type Length Values
(TLVs) which are blocks of information embedded in
GRDP advertisements
Table 1: Different Type Length Values (TLVs)
TLV
Device-ID
TLV
Address
TLV
Port-ID
TLV
Capabilities
TLV
Platform
TLV
IP Network
Prefix TLV
Full/half
Duplex TLV
Definition
Identifies the device name in the form
of a character string
Contains a list of network addresses of
both receiving & transmitting devices
Identifies the port number on which the
GRDP packet is sent
Describes the devices functional
capability in the form of a device type
(A switch)
Describes the hardware platform name
of the device (Cisco 7500)
Contains a list of network prefixes to
which the sending device can forward
IP packets. This information is in the
form of the interface protocol and port
number (Eth 1/0)
Indicates status (duplex configuration)
of CDP broadcast interface. It is used
by network operators to diagnose
connectivity problems between adjacent
network elements
7. Preventing Address Conflicts
After the CR receives an assignment from the MZR, it
sends an Address Resolution Protocol (ARP) request to
the address that it has been assigned. If a reply to the
ARP request is received, the CR has detected a conflict
and sends a DRCP-Decline message to the MZR. The
CR is then assigned the next available address in the
scope. A scope must be properly defined and activated
before CRs can use the DRCP for automatic TCP/IP
configuration. A DRCP scope is an administrative
collection of IP addresses and TCP/IP configuration
parameters that are available for assignment to the CRs.
A scope has the following properties:
- A scope name, assigned when it is created.
- A range of possible IPs from which to include or
exclude addresses used in DRCP offers.
- A unique subnet mask, which determines the
network ID for an IP address in the scope.
8. Auto Subnetting
The concept of auto-subnetting is essential here, because
we are letting both Manual & Automatic allocation of
network configuration to be a part of DRCP. And when
we say Automatic Allocation it refers to allocation of IP
as well as Subnet Mask based on the requirements of the
CRs.
What it does is when a request for assignment
comes; at first it checks the class of the requested block
of IP Address as well as the number of available
unassigned IPs of that class. And then a subnet mask is
set to release that IP from the general pool of IPs and
assign to that specific CR.
But things to note here is the overlapping of IP
among different regions after the specification of subnet
mask. An improper masking may drag one IP from one
region to another resulting in IP Allocation mishaps.
Sequence of Events:
CHECK whether the requested IP is valid or not
CHECK whether the IP is from expected class or not
FIND which of the 5 zones holds that IP
CHECK availability of the IP in that particular zone
FIND number of IPs needed
FIND subnet mask to isolate those IPs
APPLY IP & Mask to the requesting CR
9. Blending the concept of DRCP with OSPF
As we know from OSPF routing protocol design that
among the areas of OSPF one area needs to be
designated as Backbone Area. I am actually proposing to
extend this concept to one step further.
I would like to suggest that there should be Five
Backbone Areas associated with the three MZRs. For
example, the MZR associated with Backbone Area 1
should have full control over the routers inside Area 1
and so on. Each backbone area can have one or more
Backbone Area Border Router (BABR) associated with
it as shown in Figure. 2. Actually I am proposing to add
an extra level to Area Hierarchy where there are three
BABRs spanned across five backbone areas creating a
network of BABRs. Each Master Area Border Router
(MABR) of a particular BABR will eventually
communicate with the Sub Area Border Routers (SABR)
www.ijsret.org
273
International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 3, Issue 2, May 2014
Fig. 2 Proposed network of BABR, MABR & SABR
under its supervision. Here the function of the MABR is
to maintain the Routing Table based on which zone it is
in as well as which portion of which class of IP it is
supposed to distribute to SABRs controlled by it. And
the function of the SABR would be to distribute IPs
properly across the routers around the world as per need.
I will leave the details for future work for time being.
10. Conclusion
This is mostly an imagination based on some real life
system of course. The whole idea was to automate the
entire process of IP assignment amongst the routers in
the world. I know it needs a lot of synchronization with
lot of things. But if we can look at things one at a time it
might make some sense in the near future where I
believe this whole imagination would not sound as weird
as it is now. We cannot just rule out human interaction
with the DRCP Protocol at all but what I am suggesting
here is to do so at the minimum level.
References
I am not including any reference from any research
paper or anything similar to that, because I have not used
any. I got the above mentioned ideas from the existing
technologies as listed below:
[1] DHCP (Dynamic Host Configuration Protocol)
[2] CDP (Cisco Discovery Protocol)
[3] IP Subnetting Technique
[4] OSPF Routing Protocol
www.ijsret.org
274
