Automated Deployment of IP WAN Gets Closer
by William Flanagan, Publisher
After ten years, the problem addressed in VL Issue 1
(when we started with a focus on deployment automation) finally seems
close to a practical solution. To recall: in
the early days, installation of TDM data circuits was very labor
intensive--comparable to switching voice calls through manual patch
The PSTN overcame that problem via direct dialing and automated
switches. Easier provisioning of data services became possible
with the wide availability of packet switching, routing, domain name
service--the Internet--and Virtual Private Lines (VPL's). But
there remained serious problems associated with multiple carriers and
configuration. For example:
Help is almost here. Within one carrier's network, LSP's today
are configured by indicating the end points and some path
attributes. The Network Management System (NMS) takes care of
finding a path and setting up the routing/forwarding tables in the
switches and routers. Automated hand offs between carriers is, as
the ITU puts it, "For further study," but it should be here in your
- Carriers using different code points for prioritization levels,
or deleting these header fields, cause packets passing between carriers
to lose their Quality of Service (QoS) settings. Voice packets
get only best effort service, along with all the data on the network.
- Errors in mapping between a Virtual Private Circuit on the access
side and a Label Switched Path (LSP) in the core can lead to one-way
connections, using the wrong path (adding latency to voice calls), or
inappropriate bandwidth limitations.
- Configuration of Customer Premises Equipment (CPE) that doesn't
match the carrier requirements on SIP trunks can introduce audio and
video impairments that are difficult to troubleshoot.
Internet Telephony and the Future Voice Network
Wiley Interscience published my sixth book earlier this year. For
the list of previous books see the Publications
page on our web site. Together, those five titles have sold over
70,000 copies. Hope you like this one too. Order from Wiley.com
or your favorite technical bookstore.
Provisioning, particularly across multiple carriers, gets done every
day. But then comes the verification of the circuit. If
there's a problem, the location of the root cause may not be
Traditional testing methods relied on each carrier or autonomous
network entity to test within its boundaries and compare notes with
others. You've heard of "finger pointing"? The customer
could verify connections end-to-end based on CPE, but had no way to
locate a problem within a carrier network.
A joint announcement by RAD and EXFO brought back the historical situation
and triggered this issue. Rad makes Network Interface Devices
(NID's) that carriers use to terminate local loops at customer
sites. The NID regenerates the signals, provides monitoring, sets
up loopbacks, and often contains an Ethernet switch. The EXFO
management system (BrixWorx software) collects information from
hardware probes deployed in the core. Together, the two firms
make the twain meet: end-to-end and per-segment tests. Or
they say they will, relatively soon.
Each NID collects frame and error counts as it measures latency and
jitter, per Ethernet Virtual Circuit (EVC). BrixWorx correlates
that information with what the probes in the core provide on their
segments. The result is the ability to confirm operations and
Not yet in hand is full integration with those evolving Operations
Support Systems (OSS's) that provision the circuit in the first
place. The product road map calls for automatic testing of newly
provisioned circuits, VPN's, VPL's, and VPC's. BrixWorx will
issue a "birth certificate" showing the configuration conforms to the
customer order (or where it fails to conform).
Once in operation the goal is to monitor both end-to-end performance
and per-segment performance. Ideally, the OSS sets up the service
to match the Service Level Agreement (SLA) associated with the order
for the circuit. The NMS monitors each segment of each circuit as
well as end-to-end behavior. Any problem triggers an alarm to the
appropriate people before the customer sees any problem.
Flanagan Consulting Can Help You
We understand not only the technology of networks,
the surrounding business processes:
preparation/analysis, statements of work, financial
consensus building around a solution, and more.
Find out more: call +1.703.242.8381 or
Flanagan Consulting Supports Litigation Professionals
Several associates are experienced in analysis of patents,
contracts, and other intellectual property related to IT
We have assisted attorneys preparing claims, depositions,
Queries to +1.703.242.8381.
...to reach over two thousand interesting people in Telecom and
For details, call the Publisher at +1.703.855.0191
Responses to ViewsLetter and Subscriptions
Mail is welcome when addressed to firstname.lastname@example.org.
Because of the number of 'out of office' autoreplies, replies to
this message are discarded.
Special thanks for supporting ViewsLetter to www.Webtorials.com,
your best source for communications tutorials and white papers.
From Desktop to Data Center
3800 Concorde Parkway, Suite
1500, Chantilly, Virginia USA
Consulting is a Service Mark of W. A. Flanagan, Inc.
of false knowledge; it is more dangerous than ignorance."