Network Performance
Improve network performance
by Joel Trammell
While infrastructure reliability has
improved to a point where 99.9 percent
availability is not uncommon, networked
application-performance issues are growing
dramatically due to trends such as data
center consolidation, the rise of multimedia
traffic and growing numbers of remote users.
Relying solely on infrastructure
availability and utilization is no longer
enough to address these challenges,
especially as network professionals are
becoming increasingly responsible for
application delivery across the network
infrastructure. Given the improvements in
device availability, organizations are
focusing more on performance management to
address application delivery issues. This
performance-first approach to network
management inverts the traditional
device-monitoring approach and begins with
top-down visibility into application
performance.
A performance-first model is driven by
the fundamental purpose of the network
infrastructure: to transport data from one
end of the system to the other as rapidly as
possible. The more efficiently data flows at
the transport layer, the better the
application performance. Thus, the best
indication of how applications are
performing for the end-user is to measure
response times from an end-user perspective.
A performance-first approach requires
starting with response times to get an
overall view of what is happening on the
network and then drilling down into other
key performance metrics as needed, including
traffic flow analysis, device performance
and deep-packet analysis.
Understanding and base-lining end-to-end
response times is the essential starting
point for making strategic decisions, such
as how to optimize the network, plan new
infrastructure rollouts and upgrades, and
identify the severity and pervasiveness of
problems. For instance, end-to-end
performance monitoring enables network
professionals to decide where WAN
optimization and application-acceleration
technologies are most needed and to measure
before and after impact.
An end-to-end performance-monitoring tool
should tie end-user performance back to the
IT infrastructure. The deconstruction of
total end-user transaction time allows
network professionals to analyze the
behavior of networks, servers and
applications.
With end-to-end performance monitoring in
place, traffic analysis is a key capability
needed to drill into network links and
understand how application traffic is
impacting network performance. Identifying
which applications and users are consuming
bandwidth, and when, enables network
professionals to make informed decisions.
For example, if end-to-end performance
monitoring isolates a problem to the network
infrastructure, traffic analysis is needed
to identify where latency is higher than
normal or expected. This yields the
information needed to redirect or
reprioritize application traffic, or add
capacity. In addition, visibility into new
or anomalous traffic patterns pinpoints
performance problems and identifies security
risks.
Managing network infrastructure, devices
and services is also a critical component of
the performance-first approach, for both
short-term troubleshooting and long-term
planning. If end-to-end performance
monitoring shows the source of latency is
isolated to an infrastructure component
(e.g., a busy router or a server memory
leak), network professionals need
performance-management capabilities to poll
the device in question so corrective action
can be taken.
Retrospective network analysis provides
the ability to view and analyze detailed
packet-level information before, during and
after a problem. This capability enables
easier and faster troubleshooting,
especially when combined with end-to-end
performance monitoring.
While using these key performance metrics
individually is a step in the right
direction, integrating these capabilities
provides a more complete approach. Common
data sources and workflow act as a force
multiplier for benefits, enabling IT staff
to move seamlessly between capabilities for
real-time troubleshooting and long-term
planning purposes.
Without visibility into end-user response
times, traffic flows, infrastructure health
and packet-level traffic, managing
application performance proactively is
challenging. For optimum network and
application performance, IT organizations
should understand how well services are
being delivered and where violations occur.
Whether moving to multiprotocol label
switching, deploying a new application or
consolidating data centers, network
professionals should understand network and
application performance before, during and
after the change to ensure success. Without
access to key performance metrics, they are
operating virtually in the dark.
Joel Trammell is chief executive
officer of NetQoS, Austin, Texas.
For more information
(click here)