Home > Articles > Cisco Network Technology > General Networking > Interconnecting Data Centers Using VPLS: Data Center Layer 2 Interconnect

Interconnecting Data Centers Using VPLS: Data Center Layer 2 Interconnect


  1. Overview of High-Availability Clusters
  2. Data Center Interconnect: Legacy Deployment Models
  3. Problems Associated with Extended Layer 2 Networks
  4. Summary

Chapter Description

This chapter provides an overview of high availability (HA) clusters, legacy deployment models for interconnecting data centers, and problems related to extending Layer 2 networks.

Data Center Interconnect: Legacy Deployment Models

Several transport technologies are available for interconnecting the data centers, each of which provides various features and allows different distances, depending on factors such as the power budget of the optics, the lambda used for transmission, the type of fiber, and so forth.

Consider the features of the LAN and SAN switches that provide higher availability for the data center interconnect (DCI) before considering some of the available technologies. The convergence time required for the application also is important and should be evaluated.

The list that follows describes common transport options:

  • Dark fiber: Dark fiber is a viable method for extending VLANs over data center or campus distances. The maximum attainable distance is a function of the optical characteristics (transmit power and receive sensitivity) of the LED or laser that resides in a small form-factor pluggable (SFP) or Gigabit Interface Converter (GBIC) transponder, combined with the number of fiber joins, and the attenuation of the fiber.
  • Coarse wavelength-division multiplexing (CWDM): CWDM offers a simple solution to carry up to eight channels (1 Gbps or 2 Gbps) on the same fiber. These channels can carry Ethernet or Fiber Channel. CWDM does not offer protected lambdas, but client protection allows rerouting of the traffic on the functioning links when a failure occurs. CWDM lambdas can be added and dropped, allowing the creation of hub-and-spoke, ring, and meshed topologies. The maximum achievable distance is approximately 60 miles (100 km) with a point-to-point physical topology and approximately 25 miles (40 km) with a physical ring topology.
  • Dense wavelength-division multiplexing (DWDM): DWDM enables up to 32 channels (lambdas), each of which can operate at up to 10 Gbps. DWDM networks can be designed either as multiplexing networks that are similar to CWDM or with a variety of protection schemes to guard against failures in the fiber plant. DWDM also offers more protection mechanisms (splitter protection and Y-cable protection), and the possibility to amplify the channels to reach greater distances.

In nearly all of these deployment models, costs associated with deploying and maintaining a dedicated optical network is one of the biggest concerns. Also, there is no STP isolation. Depending on the nature of the problem, issues in one data center will affect other data centers. Another disadvantage is the lack of load balancing across redundant paths due to blocked links in the core network.

3. Problems Associated with Extended Layer 2 Networks | Next Section Previous Section