x
Optical/IP

Report: Users Lukewarm on IP VPNs

Businesses with large investments in frame relay networks aren't looking to change anytime soon, according to a recent report from Vertical Systems Group.

A survey of over 600 U.S. enterprise customers conducted by the firm indicates business are satisfied with frame relay's cost and convenience, compared with the private lines they formerly used.

"Certainly, IP VPN growth is there, but none of our research suggests a large-scale swapout of frame relay for IP VPNs," says Erin Dunne, director of research services at Vertical.



Further, Dunne says businesses willing to launch a complete replacement represent just 7 percent of the total frame relay locations overseen by survey respondents.

"Typically, businesses are making the transition for smaller networks or for a limited number of sites," she says.

Why aren't more big companies moving to IP VPNs? Dunne says many of these customers see frame relay and even Asynchronous Transfer Mode (ATM) as a more reliable means than Internet Protocol (IP) for supporting multiprotocol traffic containing legacy protocols such as IBM's SNA.

It also seems that companies aren't motivated yet by the economics of making a change. If prices are higher without a sizeable payoff, it's no go (see Readers Take Shots at IP VPNs ).

Vertical's still tallying survey results to find out the percentage of customers that may be adopting IP-based "transition services" that allow them to keep their frame relay networks but tack on IP VPNs (see Crossing Over to VPNs).

But Dunne says it's not clear these customers will move to expand their IP VPNs in the future. Indeed, she thinks such transition services are really a compromise for companies with pockets of IP traffic that continue to subscribe to frame relay as the prime mover of multiprotocol data.

The migration to all-IP will take a long time. "Frame relay's been around for ten years. You can expect that sort of ramp-up for IP VPNs," says Dunne.

— Mary Jander, Senior Editor, Light Reading

Page 1 / 2   >   >>
ohub 12/5/2012 | 12:28:54 AM
re: Report: Users Lukewarm on IP VPNs Since Frame Relay can work properly, why do the carriers need to take this bother and risk to replace it with IP VPNs. Just because IP VPNs is a more advanced technology or a new generation technology? Frame Relay is a very mature, stable, and reliable technology. However, IP VPNS is a comparatively new technique. Its stability, reliability and availability need to be proved. IP VPN may have a market to carry some data traffic which overflows the capacity of the current Frame Relay or ATM facilities. But that will be a new deployment instead of replacement. In my pointofview, IP VPN still needs a real provement.

Optical Hub
http://www.ohub.net
teng100 12/5/2012 | 12:28:52 AM
re: Report: Users Lukewarm on IP VPNs "The view I've heard from the carrier side is
that nobody wants a seperate FR network internally
(with the extra cost), but that they want to
keep selling something called FR "service"
forever because people will pay a high margin
price for it regardless of how its actually
being delivered."

How do you define the SLA for such case?
whoever57 12/5/2012 | 12:28:52 AM
re: Report: Users Lukewarm on IP VPNs Is it possible that enterprises are setting up their own IP-VPNs using open source software, such as FreeS/WAN? (www.freeswan.org) and hence have no need for external provision of IP VPNs?
skeptic 12/5/2012 | 12:28:52 AM
re: Report: Users Lukewarm on IP VPNs Since Frame Relay can work properly, why do the carriers need to take this bother and risk to replace it with IP VPNs.
-------------
The theory was that they could offer consolidated
VPN service at a lower price and less equipment.
Sometimes this was a customer driven thing. They
looked at the cost of certain services and saw
an advantage to consolidation on IP. But as
with many customer-driven things, they didn't
really think it through.



But in practice, some people started to finally
figure out that killing margins on a cash cow
like frame relay service wasn't going to help
the bottom line. So (and I have not checked
this recently) they priced it high.

In some places, cisco was almost giving away
equipment to get carriers to deploy VPN services
as well.

What I expect in the long run is that the
carriers will sliently accomplish the goal
of equipment consolidation by reducing the
amount of equipment doing FR inside their
network. Just because you have an FR service
doesn't mean that it has to be end-to-end FR.

The view I've heard from the carrier side is
that nobody wants a seperate FR network internally
(with the extra cost), but that they want to
keep selling something called FR "service"
forever because people will pay a high margin
price for it regardless of how its actually
being delivered.


skeptic 12/5/2012 | 12:28:51 AM
re: Report: Users Lukewarm on IP VPNs Is it possible that enterprises are setting up their own IP-VPNs using open source software, such as FreeS/WAN? (www.freeswan.org) and hence have no need for external provision of IP VPNs?
--------------
It depends on what you mean by VPNs. freeswan
as I remember is just an IPSEC package for
linux.

The article may (or may not) being using a
broader definition of VPNs than just secure
tunnels.
AAL5 12/5/2012 | 12:28:49 AM
re: Report: Users Lukewarm on IP VPNs Teng100 asked: "How do you define the SLA for such case?",

One way is to run FRoMPLS, implement per PVC CIR/EIR on the edge using traffic policing, and reserve enough bandwidth in the core using Traffic Engineering.

The EXP can be used to classify AToM packet treatment in the core to use Low latency queueing. Then shape the traffic as it exits the MPLS core back into the FR network.

This is one way of doing this, not the only way.

AAL5
whoever57 12/5/2012 | 12:28:47 AM
re: Report: Users Lukewarm on IP VPNs << It depends on what you mean by VPNs. freeswan
as I remember is just an IPSEC package for
linux.

The article may (or may not) being using a
broader definition of VPNs than just secure
tunnels. >>

FreeS/WAN can provide a secure tunnel (gateway-to-gateway) and it can provide host-to-host and host-to-gateway ("roadwarrior") configurations, based on IPSEC.

But my point is not that solutions such as FreeS/WAN can provide solutions for all IP VPNs, rather that the number of deployed IP VPNs may be significantly higher than that recognized by the carriers, since many VPNs are deployed without any carrier involvement or knowledge.
teng100 12/5/2012 | 12:28:47 AM
re: Report: Users Lukewarm on IP VPNs "One way is to run FRoMPLS, implement per PVC CIR/EIR on the edge using traffic policing, and reserve enough bandwidth in the core using Traffic Engineering.

The EXP can be used to classify AToM packet treatment in the core to use Low latency queueing. Then shape the traffic as it exits the MPLS core back into the FR network."

What about OAM&P?

AAL5 12/5/2012 | 12:28:46 AM
re: Report: Users Lukewarm on IP VPNs Teng100,

have a look at Draft-ietf-mpls-lsp-ping-01.txt regarding LSP fault detection. Still work to be done in this area but progress is being made...

AAL5
rtfm 12/5/2012 | 12:28:45 AM
re: Report: Users Lukewarm on IP VPNs Would you characterize SSL VPNs as an alternative VPN solution?

It mainly handles the security aspects, not QoS, but for many users (e.g., Healthcare people handling off-line data), that is what they mainly care about.

This is not *free* but the revenues to the software provider, and lie at the CPE level.

The question remains, is it just a matter of a solution out there (e.g., MPLS-VPN) that does more than what people need, rather, will pay for, today? If VPNs mean just security and basic connectivity, why *do* we need so many upgraded technologies? Yes, if we try and do all the things, like Traffic Engineering, QoS, security, then an upgrade might be better, but for many users, I don't know.

Even raw connectivity can have some level of SLA (uptime if not better).

rtfm
Page 1 / 2   >   >>
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE