|
Reflecting on VPN evolution, we see
that each type had certain limitations. But VPNs based
on MPLS incorporate the best of past versions, along
with new features like smart healing. by Vinod Chandran
Multiprotocol label switching
based IP-VPN (Internet Protocol-Virtual Private Network)
is the latest technology for providing secure VPN connectivity
to closed user groups. It is an Internet Engineering
Task Force (IETF)-specified framework that uses label
switching to forward data through the network. Labels
are short fixed-length identifiers that define the data
path followed through the network. Such paths are known
as label switched paths (LSPs).
MPLS is an integration
of Layer 2 and Layer 3 technologies with reference to
the OSI model for data communications. By making traditional
Layer 2 features available at Layer 3, MPLS enables
the service provider to offer a one-tier network. MPLS
traffic engineering automatically establishes and maintains
a tunnel across the backbone using RSVP (Resource Reservation
Setup Protocol).
The path used by
a given tunnel at any point in time is determined based
on the tunnel resource requirements and network resources
such as bandwidth. Tunnel paths are calculated at the
tunnel ingress based on a fit between required and available
resources (constraint-based routing). The IGP (Interior
Gateway Protocol) automatically routes the traffic into
these tunnels. Typically, a packet crossing the MPLS
Traffic engineering backbone travels on a single tunnel
that connects the ingress point to the egress point.
Why is MPLS
important ?
MPLS enhances the
IP Services offering, with guaranteed Quality of Service
and supports traffic engineering as well as multimedia
services.
It provides a diversified
range of services (Layer-2, Layer-3) to meet the requirements
of the entire spectrum of customers, including security,
quality of service (QoS), and any-to-any connectivity,
with service level agreements and fully managed services.
- Fully integrates IP Routing
& Layer 2 switching
Conventional IP routers
use longest match lookups to perform the routing function.
The router matches the IP address on the incoming
packet with the address on the routing table having
the best match, in terms of number digit matches.
This involves a search algorithm and the routing is
based on the best match. This
process is inherently more time-consuming than the
exact-match lookups performed by MPLS routers. But
now new route lookup engines have emerged that run
just as fast as MPLS lookup engines.
- Utilizes existing IP infrastructure
to deliver multiple services
By just changing the
way packets are assigned to an LSP, new services can
be easily offered without changing the MPLS forwarding
infrastructure. New services can therefore migrate
seamlessly to a common MPLS infrastructure.
- Optimizes IP networks
through traffic engineering
Traffic Engineering
to optimize the flow of traffic through the IP network
is made possible by MPLS. Traffic can be segregated
and processed differently based on network performance
requirements.
- Seamlessly integrates
private and public networks
The public and private
networks can share the same infrastructure without
any possibility of mix-up.
MPLS OBJECTIVE
The IETF wanted
to do a single prefix lookup at the entry into the MPLS
network and make forwarding decisions in the network
based on fixed label values--with the ability to specify
routes.
Historically, ISPs
could not cause traffic to dynamically flow on the path
they chose; the traffic followed the IGP/PNNI shortest
path. Although dynamic protocols are normally very effective
at routing traffic, ISPs could not split traffic between
diverse routes.
Dynamic ATM algorithms
did not permit easy binding of traffic to specific paths.
The carriers defined MPLS to allow them to bind specific
prefixes to specific paths. In other words, they defined
a way for traffic to follow a different path than the
IGP (Interior Gateway Protocol) would dictate—even if
that administrative path were a sub-optima IGP path.
The IETF Working
Group was chartered in the spring of 1997 with the following
objectives:
- Enhance the performance
and scalability of IP routing.
- Facilitate explicit routing
and traffic engineering
- Separate control (routing)
from the forwarding mechanism so each can be modified
independently
- Develop a single forwarding
algorithm to support a wide range of routing and switching
functionality.
The writer is DGM,
BSNL.
|
VPNs were initially
being set up using physical leased lines hired
from service providers by the user organizations.
All the hardware and management of the network
was from the user companies.
However such VPNs
could not meet the highly dynamic connectivity
requirements and also in cases of failure of the
physical leased line, no protection of traffic
was possible.
Internet VPNs then
emerged as an alternate solution. In this type
of interconnection the nodes are interconnected
over the Internet using encryption techniques
to provide the necessary security.
However, despite the
low cost and wide reach, such VPNs do not meet
the complete requirements, since the Quality of
Service is not guaranteed when the connections
are across the Internet. At present tunneling
VPNs are popular among many business units. Secure
tunnels are set up end-to-end across the Internet
for exchange of information between the source
and destination.
Such tunneling VPNs
are highly complex and costly especially due to
the problem of setting up squared tunnels for
connecting locations. The performance of these
tunneling VPNs is limited by the delay and bandwidth
congestion seen on the Internet.
VPNs based on MPLS
provides the best of both worlds as it has inherent
self-healing ability by automatically re-routing
traffic and also ensures Quality of Service including
high throughput with low latency and jitter. This
is possible because unlike the public Internet,
the routers are interconnected over fully dedicated
long distance backbone transmission network of
the service provider. The traffic forwarding is
based on label swapping across label switched
paths.
|
|
|| Feedback
|| Subscribe-
