Cisco Cisco Prime Network Analysis Module 6.2 for Nexus 1100 Series White Paper
Page 2
©2014 Enterprise Management Associates, Inc. All Rights Reserved. | www.enterprisemanagement.com
2
Network Visibility in the Data Center:
Best Practices for Staying Ahead
gathered from the network perspective. In fact, despite all the layers of virtualization and change, there
are still a few essential truths upon which networking pros can rely. In particular, application traffic and
activity will, at some point, travel across a physical network link and through physical or virtual network
components. While an executable may live within one hypervisor now and another in the next hour, and
an external cloud the day thereafter, the sessions and transactions associated with that application must
still travel across a network, either virtual or physical. This means that strategically placed, application-
aware monitoring technologies can continue to provide essential visibility and valuable insights for
characterizing and troubleshooting performance. That said, network-based monitoring technologies that
were developed for more static, lower-speed infrastructures cannot necessarily be directly adapted and
deployed successfully in the dynamic data centers of today and tomorrow.
Best Practices: Network Monitoring in the Data Center
In order to keep pace and fulfill the value required, network monitoring tools, technologies, and practices
must advance and evolve alongside the changes that are occurring within the managed environment. In
particular, network and application activity monitoring must contend with new demands for scalability,
new visibility barriers represented by virtualization and cloud services, the rising focus upon application
and service awareness, and the need for facilitated collaboration and communications. Following are
more detailed assessments of each of these areas of new and evolving requirements.
1. Planning for growth: Scalability represents a key challenge to network monitoring tools, as
physical network speeds continue to rise at a torrid pace. Recent EMA research indicates rapid
adoption of 40G Ethernet and fast-rising interest in 100G. While network monitoring tools
that rely purely on SNMP, log files, or NetFlow as data sources are only indirectly affected
by such escalation, those that provide packet analysis insights must directly accommodate this
shift. As of early 2014, most packet-based network monitoring solutions have been upgraded to
accommodate 10G Ethernet speeds, but only a very few have begun to address 40G and higher
speeds. Creative approaches exist to extend the effective use of lower-speed packet inspection
technologies, but a clear path towards reliable, lossless monitoring is ideal.
adoption of 40G Ethernet and fast-rising interest in 100G. While network monitoring tools
that rely purely on SNMP, log files, or NetFlow as data sources are only indirectly affected
by such escalation, those that provide packet analysis insights must directly accommodate this
shift. As of early 2014, most packet-based network monitoring solutions have been upgraded to
accommodate 10G Ethernet speeds, but only a very few have begun to address 40G and higher
speeds. Creative approaches exist to extend the effective use of lower-speed packet inspection
technologies, but a clear path towards reliable, lossless monitoring is ideal.
2. Accommodating constant change: The rate of change occurring in today’s dynamic data center
impacts network monitoring in a couple of important ways. First, new services and applications
are being deployed on a regular basis, and application-aware network management and monitor-
ing systems must be updated to properly recognize and track them. Second, with the advent of
programmability and a greatly expanded virtual networking tier, network monitoring systems
must be able to accommodate rapid rates of change as well as increasingly short life spans of virtual
connectivity, such as VLANs. Finally, the advent of cloud services requires network monitoring
tools and practices to adapt to operate across a hybrid mix of internal and external resources.
are being deployed on a regular basis, and application-aware network management and monitor-
ing systems must be updated to properly recognize and track them. Second, with the advent of
programmability and a greatly expanded virtual networking tier, network monitoring systems
must be able to accommodate rapid rates of change as well as increasingly short life spans of virtual
connectivity, such as VLANs. Finally, the advent of cloud services requires network monitoring
tools and practices to adapt to operate across a hybrid mix of internal and external resources.
3. Optimizing visibility/response: As operations teams converge and coalesce around solidifying
a better and more complete understanding of the performance of hosted/delivered applications
and services, coupled with the end user/customer experience, network monitoring must be
aligned to support this shift. Key to success is having enabling features for organizing and pri-
oritizing monitoring based on the most important and most sensitive applications and services,
including the identification and subsequent tracking of key performance indicator metrics, such
as connections, volume, and response times. Also of tremendous value here is an ability of the
network monitoring system to provide related context, revealing potential negative influences
and services, coupled with the end user/customer experience, network monitoring must be
aligned to support this shift. Key to success is having enabling features for organizing and pri-
oritizing monitoring based on the most important and most sensitive applications and services,
including the identification and subsequent tracking of key performance indicator metrics, such
as connections, volume, and response times. Also of tremendous value here is an ability of the
network monitoring system to provide related context, revealing potential negative influences