– Tara Van Unen, senior manager of market development at Ixia (www.ixia.com), says:
There is no such thing as a perfect network. Physics imposes limitations such as noise on copper or fiber optic cables and the speed of electrical pulses on wires or light on fiber optics. Packets may be lost or delayed by transient congestion in switching elements of the net. Packets may also be lost, replicated, or reordered by changes in routing or transitions between slow-path and fast-path routing mechanisms. Packet reordering may also be caused by “load balancing” of traffic between a pair of routers using a set of parallel links.
These conditions create delays from a few milliseconds to a few minutes. There are times when there is no usable route for packets to flow from some point A to some other point B on the net. Reordering caused by parallel telecommunications links can last for as long as those links are in place.
Data center switches need to deliver ultra-low latency for mission critical applications. They need to deliver the best latency performance to analyze and compute time-sensitive data. They also must also work effectively to minimize delay in the WAN to ensure prompt delivery of data.
In order to correctly assess WANs before depending on their performance, WAN emulation must occur. Thorough WAN emulation means validating several items:
- Pre-deployment testing of applications that work over WAN
- Identifying application behavior that can vary widely from LAN to WAN
- Delay and delay variations
Enterprises need to validate the impact of real network characteristics on enterprise applications before deployment. They also must characterize how WAN acceleration works over a real link with delay and other impairments. Service providers need to understand the impact of impairments on services being deployed such as video, voice, and web applications. They also must characterize the robustness of Ethernet-based backhauled services under real world network environments. NEMs need to validate resiliency and robustness of their equipment using diverse, real-world mixes of traffic, services, and users, as well as verify the functionality of Ethernet and MPLS OAM operation.
WAN emulation should necessarily include elements such as:
- Voice, video, and data transmission over TCP/IP
- Variety of TCP and HTTP acceleration techniques
- Buffering techniques for bandwidth optimization
- Data compression
- Transparent caching
- Bandwidth on demand
- Asymmetric links
- Dynamic bandwidth allocation
- WAN impairment
Injecting WAN impairment into existing networks allows IT managers, carrier operators, and equipment manufacturers to measure network recovery times and overall end user QoE. WAN impairment includes:
- Delay and delay variations
- Packet Drop
- Packet reordering
- Packet duplication
- Bit error insertion to emulate network errors
- Link bandwidth limiting
- WAN forwarding error interruption
- Ethernet CRC auto-correction
Equipment manufacturers, service providers, and enterprises have long needed to validate the functionality of their LAN and LAN applications. With the continued growth of world-wide connectivity, along with new models such as cloud-based and virtualized infrastructures, services over WANs must be managed as well. These factors, in addition to the growing nature of bandwidth intensive applications, multiplying wireless devices, and exponential increase in users, mean that networks can no longer afford to ignore how delays in WAN networks affect their customer QoE.
Today, testing solutions exist that can combine layer 2 through 3 load generation and network impairment in to the same solution to deliver “real world” accuracy when testing networks prior to or during deployment. These solutions provide unprecedented scaling to test network impairments for extensive wide area network emulations. Network administrators or testing engineers can save time and effort with an integrated user interface, which enables rapid configuration of both layer 2 and layer 3 traffic, as well as network impairments. An easy-to-use interface gives users access to a unique integration of traffic generation and associated impairment configuration. This also ensures simplified impairment test creation by generating many protocol fields of interest in the traffic classifier configuration, which users can customize to meet a variety of testing needs.