Measuring the performance of IP networks through the use of standard protocols has always been a challenge. The inventors of IP provided some tools, such as ICMP Ping, Traceroute and UDP Echo, as part of the TCP/IP suite of protocols.
Measuring the performance of IP networks through the use of standard protocols has always been a challenge. The inventors of IP provided some tools, such as Internet Control Messaging Protocol ping, Traceroute and User Datagram Protocol (UDP) Echo, as part of the TCP/IP suite of protocols. However, these tools were not intended to conduct overall performance measurements but were designed for simple troubleshooting of IP networks. As a result, there's a need for standards-based, effective performance-monitoring tools within enterprise networks.
The IETF aims to fill that need with a new draft standard developed by the organization's IP Performance Metrics working group. The Two-Way Active Measurement Protocol (TWAMP) defines a flexible method for measuring round-trip IP performance among any two devices in a network that support the standard. With TWAMP, enterprise IT managers can effectively measure the complete IP performance of underlying transport through cooperation between network elements that have already been deployed.
In the past, protocols were proprietary. As such, there was no interoperability among devices from multiple vendors. This approach worked well for equipment providers, as it forced their IT customers to purchase and deploy their products throughout the network.
TWAMP provides operators of large enterprise networks a flexible choice of solutions and full visibility into network performance via interoperability among all devices deployed in their networks. It works by measuring core and edge IP performance through cooperation between the routers and switches in the network.
Any two endpoints can interoperate and therefore obviate the need for managers to deploy systems with closed proprietary protocols for measuring performance.
TWAMP defines two sets of protocols: one for setting up performance-measurement sessions, called the control protocol, and another for transmission and reception of performance-measurement probes.
The control protocol enables endpoints to negotiate and start a performance-monitoring session. The protocol for transmission and reception of probes that measure performance defines the packet format that is needed for measuring round-trip performance. This part of the protocol is designed to accommodate hardware-based implementations in order to offload local CPUs during performance-measurement sessions.
The TWAMP architecture is composed of several entities that are responsible for starting the monitoring session and exchanging packets. TWAMP defines different entities for flexibility, and some of them can be collocated for ease of implementation.
The TWAMP control protocol provides a flexible way to set up monitoring sessions and exchange information between transmitter and receiver of monitoring packets. In such scenarios, it is possible to eliminate the need for some of the entities in TWAMP. This simplified architecture is defined as TWAMP Light within the standard.
TWAMP Light is designed to help implement the standard for entities that act as active responders to TWAMP controllers within the network, thereby enabling the measurement of two-way IP performance from anywhere in the network.
Standards-based performance measurement protocols are needed within large enterprises to create an ecosystem where IP performance metrics are obtained through cooperation among network elements.
Through the use and deployment of TWAMP, enterprises can avoid the costly deployment of performance management systems with proprietary protocols while effectively measuring the IP performance of their network at all locations.
Hedayat is vice president of engineering and CTO of Brix Networks. He can be reached at firstname.lastname@example.org.