Critical Considerations for LAN-to-LAN Virtual Private Networks

A pivotal paper by Atkinson and Haller in 1994[1] laid the groundwork for secure transmission over the Internet and Virtual Private Networks (VPNs) were born. Infonetics Research, Inc. projects the worldwide market for Virtual Private Networks to reach $32 billion by 2003.[2] There are two primary reasons for the growing interest in VPNs: the cost savings associated with replacing leased lines with packet-switched IP networks and the flexibility of adding geographically diverse sites to the corporate network.[3] As VPNs have gained in popularity, their definition has evolved. Traditionally, VPNs were simply defined as encrypted data connections over a public network.[4] Today, VPNs include all aspects of a network design strategy, including reliability, flexibility, and latency, as well as security.[5] Applications for VPNs fall into three broad categories: access for remote users; intranets, in which a VPN replaces a leased line; and extranets, allowing business partners access to internal applications and data.[6] Many corporations have already implemented VPNs for remote access, as this is the simplest solution, and are moving on to more sophisticated applications which may involve replacing an existing (Wide Area Network) WAN. Since most corporations already have a WAN in place, they have high expectations for reliability, latency, and throughput. These expectations combined with the need for security on the public Internet make VPN requirements very stringent and the design process very important.