Transformation of the enterprise WAN with dynamic-path networking

Table of Contents

  1. Summary
  2. How dynamic-path networking addresses current problems
  3. Service capabilities
  4. Key takeaways
  5. About Greg Ferro


An emerging market of products addresses the limitations of packet routing in traditional wide area networks (WAN) with a “dynamic-path management” technology that is based on programmable network devices. These devices have APIs that enable new applications to configure a WAN reliably and predictably.

WAN services consume a large part of the IT telecommunications budget, which accounts for around 40 percent of total IT spend. WAN will continue growing and consuming budget with the adoption of cloud-based applications, so managing cost while extracting value from these networking resources is vital. At the same time, applications are consuming network resources and must be able to request and provision network services on demand; sometimes they require additional bandwidth to fulfill service requirements. Other types of application traffic, such as IP telephony, unified communication, and market-trading data, are sensitive to network latency and need specific network management capabilities to ensure that they deliver services that meet business objectives.

This report looks at the business value of dynamic-path networking, flow networking, the “atomic element” of path control, and the opportunity to transform the nature of the WAN so that it can better deliver applications instead of packets. It will help IT executives understand the limitations of packet routing in traditional WAN networks as well as the key factors they should consider when adopting dynamic-path networking.

Key findings include:

  • Routing protocols can calculate only a single path. Load sharing or multi-path selection would reduce costs and improve services, but current technology cannot deliver. Flow networking can.
  • Today’s network services must match services and application requirements, each potentially requiring different reliability, application control, and capacity requirements. The corporate network is typically built with a mix of networks from multiple carriers that use differing transmission technologies and performance characteristics. Service-provider networks are increasingly diverse in access technology and use a variety of multiprotocol label switching (MPLS) technologies to present Ethernet or IP connectivity or, in some cases, older time-division multiplexing (TDM) technology.
  • WAN complexity is increasing, while reliability and resiliency are perceived to be declining. The current approaches to carrier diversity and path redundancy increase complexity and deliver unreliable networks.
  • The most common industry approach is an MPLS overlay network for path management that further increases complexity and operational expense (OpEx). For many, the MPLS reality falls far short of business expectations or need.

Full content available to GigaOm Subscribers.

Sign Up For Free