LisaForrester has consistently advised clients that they should ensure their own WAN survivability and not rely
on a single provider to ensure a consistently high level of availability.  Two
very recent examples with US providers illustrate

– In one case, the service provider
proactively notified us of a 30 minute availability problem in a section of its
nationwide MPLS network.  The cause was internal — the provider’s process to
upgrade router software was not followed.  This case is not unique — we know of
other, similar instances.

– In the second case, several
clients reported that another provider’s MPLS network wasn’t living up to SLAs
for several hours — this due to a security breach.  The provider in question confirmed this problem with us today.  The
irony: amongst its peers, this provider has some of the toughest network security
technology and processes in place.

Both these instances occurred within
the last 10 days.

Our advice to clients is as

  1. Clients who use the
    same provider for primary and secondary WANs will not ensure WAN survivability,
    because most providers’ MPLS networks are not fully redundant and diverse
    end-to-end, and in the case of public MPLS services, these ride on a provider’s
    Internet backbone.  Thus, either type of problem described above could impact
    both networks. 
  2. If any part of a
    company’s WAN is critical to its business, then the only way to ensure the
    highest level of survivability is through facilities-based provider diversity
    for business critical links.  This is required — it is not optional.
  3. To ensure the highest
    degree of WAN survivability, networks must be completely diverse — beginning
    with access.

By Lisa Pierce

Check out Lisa’s research