Over the last 30 years computer networks have established themselves as a cornerstone
of modern business. With the commercialization of the Internet in the 1990s, networks
have become fundamental in how companies share knowledge, operate, and interact
with customers, partners, suppliers and other third parties.
Local area networks (LAN), which linked computers in the same or adjacent buildings,
were the first to make an impact in businesses. Nowadays, the wide area network
(WAN) is just as important as it allows companies to share information and business
applications countrywide or even across the globe. Smaller versions of WANs that
offer citywide or regional connectivity are sometimes labeled as metropolitan
area networks (MAN).
Underlying the flow of all this information is the physical network. Copper cable,
fiber optics, satellites and radio waves support the network technologies that
send and receive data and give access to information.
Which enterprise network technologies are most common?
Over the last five years, two technologies have come to dominate fixed enterprise
networks:
multiprotocol label switching (MPLS) and
Ethernet.
Demand for both MPLS IPVPN and Ethernet services are booming.
Analyst Infonetics says that worldwide Ethernet service revenue grew 33% to $12.5 billion in 2007,
and IP MPLS VPN service revenue grew 20% to $13 billion ; it predicts that both
will grow strongly through to 2011.
What's so special about MPLS?
MPLS was first conceived in the late 1990s to improve the performance of IP traffic
over the network. For this it uses classes of service that allow enterprises to
put time-sensitive applications such as
VoIP
in a priority class, and batch traffic such as email in a 'best-effort' class.
MPLS IP VPNs are popular with enterprises because they offer quality of service
and security guarantees, and allow them to utilize a single network for all voice,
video and data traffic. Unlike older hub-and-spoke networks, MPLS IP VPNs allow
individual sites to communicate directly, which cuts down on the total network
traffic and improves performance.
MPLS is a packet-switched network; packets entering the network are assigned
a label and passed along a label switched path (LSP). This allows a service provider
to create private networks for their customers. Another advantage of MPLS is that
it can create end-to-end circuits across multiple networks, including ATM, Frame
Relay and Ethernet, which allows enterprises to integrate different networks into
a single solution. For more details on how MPLS works and its advantages
read this tutorial.
And how about Ethernet?
Ethernet is a much more venerable network technology, having been first conceived
in the 1970s. For many years it was primarily a LAN technology, where it eventually
beat off all comers to completely dominate in-building networking. More recently,
advances in
Ethernet technology have seen it take over
the MAN market, establish itself as the next-generation
backhaul network of choice for mobile operators and make major inroads into the
enterprise WAN. The main drivers for its popularity are: ubiquity, which makes
equipment economical; flexibility, as enterprises can scale bandwidth up and down
easily; and its support for IP.
Carriers have marketed Ethernet services under a variety of names over the years,
so to clear up the confusion the
Metro
Ethernet Forum has defined three distinct services: point-to-point services called E-Line,
multipoint-to-multipoint services called E-LAN, and point-to-multipoint services
called E-Tree. E-LANs are ideal for multi-site WANs and support differentiated
classes of services, much like MPLS IP-VPNs.
Why is IP fundamental to enterprise networking?
Internet Protocol (IP) is a protocol for communicating data over a packet-switched
network and is the fundamental language of Internet communications. Packet switched
means that data is divided up into discrete segments and sent across the network
and reassembled at the other end. It is distinct from
circuit
switching, where a fixed channel is set up between two points for the time of the transmission.
The advantage of packet switching is that it allows multiple transmissions to
share the same network, so that one ‘converged’ network can carry all information
and services – voice, data, and video. It is simpler and more economical for a
large enterprise to run operations for the entire organization on a single network.
The convergence of voice and data services also opens the way for new multimedia
applications like presence-based technology and video-conferencing.
What is the difference between a managed and an unmanaged network?
Quite simply a managed network is where a service provider will manage the network
for its customers. They will offer a service level agreement (SLA) for the network
and handle all the traffic engineering, deployment, management and maintenance.
The customer will be provided with a router for their site, through which all
the traffic will pass. Unmanaged, or self-managed networks, were traditionally
popular in the US, but even here managed networks are becoming more popular. Self-managed
networks require the enterprise to build, manage and design the network themselves.
While the enterprise will have total control, they will also need to have the
necessary internal skills and resources. As networks have become more complex,
the majority of enterprises have chosen the route of managed network services.
What are the legacy network technologies that MPLS and Ethernet are replacing?
The two main network legacy technologies are asynchronous
transfer mode (ATM) and
Frame Relay. ATM was a backbone network technology used by carriers and large enterprises
to connect local-area networks. It provided guaranteed capacity for voice, data
and multimedia applications, and dealt well with suddenly demanding levels of
traffic. Frame Relay circuits were often deployed in conjunction with an ATM backbone.
Sites were permanently connected, using capacity on a service provider’s network
at the bandwidth required; any site could communicate with many destinations via
one access link. Voice, data and video were relayed in packets called frames which
carried the information needed to send them to their destination.
What is a virtual private network (VPN)?
The virtual in a VPN allows you to create a private network within a much larger
network, such as the Internet or a service provider's backbone. There are two
types of VPNs: trusted VPNs and secure VPNs. MPLS IP VPNs are an example of the
former. They allow service providers to create virtual circuits within their IP
network and sell these as VPN services. MPLS isolates the traffic streams from
one another so that customers can share the same network, much like legacy ATM
or Frame Relay services.
And what about secure VPNs?
Secure VPNs are designed to carry traffic over unsecured public networks such
as the Internet, and
include IPSec and
SSL VPNs. IPSec technology requires software on the remote machine and uses it to create
a secure tunnel through the public network. This encrypted tunnels means that
no third party is able to intercept data traveling across the network. Although
secure, the IPSec client software is costly to maintain and manage. SSL VPNs are
much newer invention and use the secure SSL protocol that is already built into
Internet browsers. Instead of providing a tunnel to the corporate network, SSL
VPNs provide remote links into specific applications using a browser. Users get
secure access to corporate applications through any Internet connected terminal
and IT departments get better control over which resources users can access.
What is the difference between an access and core network?
A multi-site enterprise WAN will typically comprise an access and a core network.
Also called backbones, core networks such as MPLS IP VPNs and Ethernet consist
of the service providers' high-speed worldwide infrastructure, to which it provides
access through a series of points-of-presence (PoP). Access networks are essentially
the link between those core network PoP and the enterprise site. Network technologies
used here include
xDSL,
T1/E1,
fiber, Ethernet,
ISDN, dial-up, wireless networks such as
WiMax and microwave, and satellite.
How do I know which access network is best for me?
The choice of access network depends primarily on the nature of the site being
connected to the network. Increasingly enterprises are choosing one DSL broadband
to connect smaller sites. The most common version of this is symmetric DSL (SDSL)
that offers the same upload and downloads speeds and doesn't suffer from the same
contention issues as consumer ADSL. The older staple of T1/E1 leased lines is
less common now.
Large sites have many more access network options and the choice will depend
on the network services they plan on running and the bandwidth required. Companies
running MPLS networks are well served by connectivity options as the service can
be delivered over just about any network from copper pairs to fiber. Ethernet
services typically run over fiber links, with fully scalable bandwidth widely
available up to 1Gbps. Sites, especially important locations, will also have a
backup network, and these are typically ISDN, or for small locations dial-up.
Both of these offer low bandwidth and as such are nowadays only suitable for emergency
use.
What do you do where there is no fixed infrastructure?
In countries where there is little or no telecoms infrastructure, enterprises
are increasingly reliant on satellites. Fixed satellite broadband systems use
very small aperture terminals
(VSAT) technology to
provide interactive multimedia communications and broadband access
to the Internet. VSAT technology requires a dish attached outside a building that
can download data at up to 4 Mbps, linking to the corporate intranet or Internet.
The PC or network is connected to a broadband modem linked to a satellite dish
that transmits data to a satellite ground station hub. Here high-speed connections
provide Internet access, email, VPNs, VoIP, broadcast and satellite navigation
services.
How do mobile workers connect to their VPNs?
Doing business away from the workplace is convenient and productive. IDC expects
the global mobile worker population to increase to more than 1 billion in 2011
- just over 30 percent of the
worldwide
workforce. Mobile workers can connect to their corporate networks through a variety of
methods, including traditional dial-up, Wi-Fi hotspots and mobile broadband.
Wi-Fi
hotspots are available at all sorts of locations such as hotels, airports and coffee
shops. Users pay for access to the local Wi-Fi network and can access the Internet,
and through a VPN, their own corporate networks. Wi-Fi cannot match the ubiquity
of
mobile
broadband. Using
3G HSPA
technology mobile operators are offering multimegabit, low-cost broadband on
the move. The key disadvantage currently is the high cost of data when roaming
in another country.
Are there any other network technologies I should know about?
An emerging contender for the ultimate solution for broadband wireless networks
is
WiMAX (Worldwide Interoperability for Microwave Access), Wi-Fi
has a limited range
compared to WiMax, which can connect devices at multimegabit speeds up to 50 kilometers
from an access point. WiMax has two varieties: fixed and mobile. In its fixed
form, it is useful as an access network for enterprise WANs. In its mobile form,
it can provide hotzone-type access to laptops and smartphones.
3G mobile provides wider geographical coverage and higher bandwidth
. Successful bidding for more radio spectrum
could decide whether WiMAX, or an
upgrade of 3G,
Long Term Evolution (LTE)
will prove the winning technology. LTE builds on existing systems and its greater
efficiency could lead to affordable wireless broadband services with faster downloads,
video sharing, and mobile TV.