The last mile refers to the link between a network service provider and a networked location. For a home user,...
this might be a cable or DSL internet link. For a company and its employees, the last-mile connection might be an internet link or a private WAN connection delivered through MPLS or Ethernet services.
When a location has just one last-mile connection, any problem with that connection affects the location's ability to use the internet. If the link goes down, for example, the location is off the network.
Companies can address this problem by using two different last-mile connections to all of their important branches -- from different providers and along different physical routes, if possible. Redundancy at this level means the branch office can stay online and productive, as one link is likely to be up when the other goes down.
With legacy technology, however, it is difficult to use both links simultaneously. It's also challenging to shift network traffic from a failing link to a good connection without killing all the network sessions. Typically, one link would be the primary connection, and the secondary link would come into service only if the primary failed. In this case, failover takes a noticeable amount of time and kills active sessions.
How last-mile connections play with SD-WAN deployment
Software-defined WAN technology makes use of all available links at the same time. It also has the ability to fail over traffic from bad links to good ones, as automatically and transparently as possible. SD-WAN deployment can typically make use of several WAN-side connections, and it can mix connection types, including MPLS, Ethernet, internet and 4G Long Term Evolution.
With SD-WAN, enterprises can bring redundant connectivity to all locations, because they can easily fold in whatever type of link is available. This minimizes the effect of problems with individual links. SD-WAN services also provide visibility into link uptime and link performance. This gives IT teams detailed support for service-level-agreement discussions with providers.
But SD-WAN deployment also increases the risk of provider sprawl, which is the proliferation of last-mile providers with which an organization has to work. Each provider adds a billing cycle and bill-review relationship, a technical support mechanism, a support team and a contract to manage.
Companies deploying SD-WAN should be wary of provider sprawl and ensure they are getting true route diversity -- not only for the last-mile connections, but also for upstream access to the internet. If providers A and B both send traffic through the same peering point with provider C to reach the wider internet, for example, a problem at C will affect both A and B. This obviously negates the benefit of last-mile redundancy.
Dig Deeper on SD-WAN
Related Q&A from John Burke
Service providers and enterprises use software-defined access to extend SDN beyond the data center and WAN to the edge of the network. Learn about ... Continue Reading
Network engineers have two standard systems utilities that help them check for packet loss. Learn how this is done and, should a problem prove ... Continue Reading
Segment routing steers packets using a list of instructions, and it can be used in MPLS and IPv6 architecture. But does software-defined networking ... Continue Reading
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.