Windows 2000 RRAS Routing

To add a persistant route, use
	ROUTE ADD -p 10.1.1.0 mask 255.255.255.0 10.1.1.1

Windows 2000 IP routing only ever has one routing table regardless of how
many routing interfaces or protocols it has. This is single-route routing.
Cisco routers maintain a routing table for each interface (multiple-route
routing).

You do not need to install routing protocols or define static routes to be
able to route packets from one interface to another when using multihomed
Windows 2000 routers.

Silent RIP listens for RIP but doesn't talk back, workstations can listen to
RIPs "silently" and update their routing table. If the silent listener is a
router, other routers will not know about the silent listening router.

ICMP Router discovery can update or set the default gateway on Windows 98
and 2000 computers. The benefits are fault-tolerance and minimum
administrative overhead.

	Workstation: 224.0.0.1  -> router advertisement -> router 224.0.0.1
	Router:      224.0.0.2  -> router solicitation  -> workst 224.0.0.2

IGMP multicast with a Windows 2000 router w/ 2 NICs: On external NIC,
use IGMP Proxy Mode, on Internal NIC, use IGMP Router Mode.

Convergence is achieved when all routers have the correct routing
information. Convergence time is how long it takes a network to
reconfigure for a loss in connectivity with another router.

When 2 different routes from the same routing protocol are found that provide
paths to the same destination, the metric is used to determine which is the
preferred route.

When 2 different routes from different routing protocols are found, that
provide paths to the same destination, the metric is ignored, the route
comes from the preferred source (protocol) is used.

Only install IGMP when you want to provide multicast from another network.
You do not need IGMP for routing protocols using multicast.

Routing Protocols

     Distance Vector based protocols periodically send out their known routes from their routing table, and update their own routing table after receiving similar information from other routers. The exchange is unsynchronized and unacknowledged. Link State based routing protocols exchange their link state using multicast traffic. The data includes a logical network ID. The exchange is synchronized and acknowledged.

     A group of routers and networks under the same administration using a common routing protocol is referred to as an Autonomous System (AS). The protocol si known as the Interior Gateway Protocol (IGP), the routers are known as Interior Gateways. They handle "intra-AS routing" (The IGPs can be RIP or OSPF).

     Exterior Gateway Protocols (EGPs) are used when 2 ASs are used to exchange routing information. Examples of EGPs are Exterior Gateway Protocol (EGP) and Border Gateway Protocol (BGP). The routers involved are Exterior Gateways and they handle "inter-AS routing".

An AS can only have one routing protocol.
Two ASs using different IGPs can communicate via EGPs.
Rule-of-Thumb: 1 area = 40 routers.

A router that sits in between 'areas' (on the border between 2 or more areas) is called an Area Border Router (ABR) and is responsible for exchanging routing information from one area to another. If an ABR joins an area with a backbone, the router communicates "directly" with the backbone. If not, it will communicate "indirectly" with multiple border routers. The area between the backbone and a non-connected (to a backbone) area is an transit area.

RIP v1

Uses broadcast rather than multicast announcements.
Subnet mask not announced with the route.
No protection from rouge routers.

RIP v2

Has multicast option.
Subnet mask is announced (CIDR / VLSM).
Authentication.
Route Tag.

OSPF (Open Shortest Path First)

OSPF uses a more efficient routing algorithm, has faster convergence time, and uses less bandwidth than RIP.

OSPF sends multicast and not broadcast over a broadcast network. OSPF never sends broadcasts.
The use of OSPF over non-persistant links is not recommended (eg, ISDN, POTS).
Uses areas to divide networks into manageable sections. The Links State Advertisements (LSAs) is the network boundary of these areas.
Each area has a unique ID (similar to IP format) and router.
If(area>1), you must have a "backbone" area with id 0.0.0.0. This will be the hub for AS to AS transfers.
The Hello interval, the Area ID, and the stub setting must all be the same on neighboring OSPF routers.
The type of network determines the OSPF message formats:
- Broadcast Multiple Access Networks
  - Ethernet
  - Token ring
  - FDDI
- Non-Broadcast Multiple Access Networks
  - X.25
  - Frame Relay
  - ATM
- Point-to-Point Networks
  - (eg, 2 router connection only)
  - Leased lines (WAN)
  - T1, T3, Fractional T1, etc...

 Rip Terminology       -> OSPF Terminology
  Routing announcement -> Link State Advertisement (LSAs)
  Routing table        -> Link State Database (LSD)
  Hop                  -> Cost

There should be no ABRs connected without going through the backbone. These
are called backdoor routers.

OSPF Router Types:

Area Border Router (ABR) - Sits on border between 2 areas.
Internal Router (IR) - Inside its area, only in its area, and handles "intra-AS routing".
AS Border Router (ASBR) - Connects different ASs. Data from outside the AS is referred to as "external routes".
Designated Routers (DRs) - DRs minimize data exchange when routers are powered on. The DR holds the data which powered-on routers will synchronize their LSDs. The DR is designated by interface priority numbers 0 - 255. The default is 1.
Backup Designated Routers (BDRs) - Designated in case a DR fails. At least 1 router on an OSPF network shoud have a router priority of 1 or greater.

if(networks<=10 && servers <=50)
	RIP
else
	OSPF

Demand-dial Routing

Demand-dial connections have the benefit of offering cost-efficient dial-up WAN links when configured with idle time-out.
A two-way initiated connection offers the greatest flexibility, but requires more configuration. A one-way initiated connection offers tighter control for the calling router (more secure), with less configuration.
Demand-dial filters are different from IP packet filters in that they define what traffic initiates a demand-dial connection. IP packet filters define what traffic is allowed in and out of the demand-dial interface ONCE it is connected.
On-demand connections require static routes, a persistant connection should use a routing protocol.
If a routing protocol is used with demand-dial connection, use:
- RIP set operation mode to "periodic update mode" and enable triggered updates.
- OSPF - Advanced properties of an OSPF interface, you may want to increase the values of the transit delay, the retransmit interval, the hello interval, and the dead interval.

Configuration

	RRAS->Server->Properties->General Tab->LAN & Demand Dial->OK.

	RRAS->Server->Ports->Properties->Modem->Configure both
	 inbound and outbound + phone #.

	Rclick routing interfaces->New->Demand dial interface,
	 check route ip packet on this interface.

	Rclick Server->Properties->IP Tab->IP Pool->Create a static pool

	When using RIPv2 on a demand dial interface, use RIP autostatic
	update mode with no multicast.