The MPLS answer

 

The new business continuity Downtime no more Remembering every instant in time The MPLS answer IT monitoring with a twist Your opinion: Talk it up The new data center by the numbers Awesome in Austin Filling the app gap

When faced with recovery after the Sept. 11 terrorist attacks, Syed Ghaus had to reroute his users to a New Jersey branch office so they could access the corporate network using dial-up connections from laptops. With a private Multi-protocol Label Switching-based network, he says, failover would have happened seamlessly.

For Ghaus, who has since joined Resun Leasing, MPLS has become a facet of business continuity. The flexible, decentralized routing structure of MPLS allows connectivity among 34 branch offices in the U.S. without requiring routing through a hub, says Ghaus, who is director of business technology at the Dulles, Va., supplier of modular buildings. Thus, MPLS eliminates the single point of failure found in traditional spoke-and-wheel networks.

This any-to-any connectivity presents a great opportunity for redundancy and seamless transition to back-up systems and applications at other locations. But as with any new technology, early adopters warn, MPLS comes with trade-offs. And without proper planning, MPLS can create more problems for business continuity than it solves.

Scott Peterson, director of network services for Accenture, points to standards as one weak point. MPLS is still new, so standards for vendor peering and multi-carrier QoS only now are making service-level agreements (SLA) achievable. So while you might have a strong MPLS network out in the WAN cloud, you could have trouble with last-mile local connections, he says.

Planning for the apps

Defining how mission-critical applications are transported over the MPLS infrastructure becomes highly important because of IP's nebulous nature, Peterson says. So you must take the time to map the IP infrastructure with performance and end-to-end management. "Then you can achieve a continuous quality of service across the infrastructure, into your data centers and remote facilities, and pass it off into the cloud" for the carrier to uphold, he says.

IT departments already should know the applications most critical to the business and how they're being used. So defining QoS should be easy. But to make your MPLS network even more fail-safe, you should take the opportunity to standardize platforms, applications and versions, suggests Jermaine Mason, IT manager at Wilson's Sporting Goods, a division of Ameri Sports in Chicago.

Then you need to select a carrier that can guarantee end-to-end service quality. The carrier also should be able to accommodate security requirements (primarily IP VPN) and support new, demanding applications (such as VoIP and videostreaming) with minimal latency, early adopters say.

Large carriers offer two ways to allocate bandwidth - static and dynamic. MCI, for example, offers committed access rate for static bandwidth allocation and supports differentiated services for dynamically allocating bandwidth to applications that otherwise would be affected by link failures or congestion. Some carriers offer the ability to directly run over the Internet using a VPN, which most companies chose for better privacy and reliability.