Kansas reins in its networks
State gives users a common view by consolidating network management
ABOUT NINE MONTHS AGO, Kansas agencies were ordered to
consolidate information technology operations and increase
'Share resources, do more with less; it's a
clich', but we're starting to implement it,'
said Scott Steves, network management software consultant for the
Information Systems and Communications Division of the
state's Department of Administration.
Steves' division provides the data, voice and video
network for 40 state agencies. It supports a Multiprotocol Label
Switching network of about 700 routers and 250 switches, serving
20,000 users across all agencies in 105 counties and across 82,000
square miles. Until recently, each agency went its own way in
managing its network segments.
'We do not mandate what tools each agency uses to manage
its network,' Steves said. But that clearly is not the most
efficient or effective way to do the job. 'When you have
multiple people polling at different intervals, you get information
that might not be accurate.'
As part of its consolidation efforts, the state has begun
standardizing its network management tools.
'The state will save money by consolidating on a single or
just a few network management solutions,' he said.
The primary tool selected for that job is the Network Node
Manager i (NNMi) series, a Hewlett-Packard Open View product that
enables large scale, multivendor network discovery, monitoring and
management. It is being deployed in the eight network operations
centers that support the state's agencies.
'Everybody is looking at the same event at the same
time,' but on their own portion of the network, Steves said.
And every agency is talking the same language and looking at the
same data. 'It saves the state of Kansas tons of
NNMi uses the Simple Network Management Protocol and
automatically discovers the network environment, mapping Layer 2
connectivity and Layer 3 relationships, and performs active and
passive monitoring of network performance and availability. The
tool can discover and begin monitoring a 1,000- node network in
three hours and does continuous discovery so that new devices can
be found and assigned to node groups in near real time.
NMMi's secret sauce is its automated root-cause analysis
engine that uses HP-created algorithms to pinpoint network
problems, their causes and their effects based on its understanding
of the network topology and relationships.
The software runs on Windows, Unix, Solaris and Linux, and a
single server can scale to manage as many as 12,000 network devices
from multiple vendors. Gathering and analyzing data on network
faults and availability in a single engine and displaying the
results in a single interface reduces overhead and simplifies
administration, Steves said.
Each agency still manages its own resources, but management data
comes from a common source so all users see the same events at the
same time on their portion of the network.
The state tested a number of products and conducted a
competition among vendors, but NMMi's cross-platform
performance in Kansas' network environment had no real
competitors. 'Its out-ofthe- box functionality is
unmatched,' Steves said. 'It frees up administrators
for moves, adds and changes and decreases the net time to
Although child labor is not part of the Kansas
costsaving's plan, Steves said the NNMi interface is simple
enough that his 10-year-old daughter is able to find what devices
are down in the network, what is working and what neighborhoods
they are in, and what performance is not within normal
A standard network management tool is not the end of the
state's consolidation plans. Steves said the next step is to
integrate data from intrusion-detection system and
intrusion-prevention systems data into NNMi to take advantage of
the root-cause analysis engine and the single interface.
On a broader scale, the state is now changing its class B
addresses to private addresses and is standardizing the
infrastructure with Cisco switches.
'It's a process,' Steves said.
'We're busy right now.'