Corvil » Questions and Answers

A. CorvilNet 5.0 is a latency management platform for networked applications. Network latency is of growing importance due to the emergence of new latency-sensitive applications, such as financial trading, high performance computing, and interactive voice and video. To maximise application performance, network managers need to drive down latency across both LAN and WAN environments. CorvilNet is the first complete latency management solution enabling customers to visualise and control network and application behaviours at millisecond and microsecond time scales.

A. CorvilNet is designed to facilitate pro–active management of low latency network infrastructures. Its primary capabilities are:

• SLA Compliance – Define, monitor, and manage flexible low latency SLAs suitable for the most demanding modern applications.
• Performance Monitoring – Point-to-point microsecond monitoring of applications and service channels across the network.
• Event Trouble-Shooting – Automated Smart Packet Capture and root-cause analysis in response to SLA violation.
• Network Optimization – Determine network provisioning actions required to support low latency performance.

A. The principle benefits of the CorvilNet solution are:

• Pro-active SLA management. CorvilNet enables network managers to escape from the reactive 'break-fix' cycle and instead focus on ensuring their network is engineered for low latency performance.
• Visibility into traffic and network behaviour at ultra-short time scales. CorvilNet's microsecond latency, jitter, and bit-rate measurements are precise enough to monitor performance in high-speed 10 Gbps LANs, and scalable enough to support global WAN deployments.
• Faster and more accurate trouble-shooting. CorvilNet's Smart Packet Capture capability allows you to quickly focus on traffic conditions prevailing during anomalous events. Drill-down graphical review of the comprehensive performance measurements saved alongside the captured data speeds problem diagnosis.
• Reduce the cost and effort required to support latency-sensitive network applications. CorvilNet's quantitative provisioning and modelling features take the guesswork out of dimensioning the network for low latency.

A. CorvilNet keeps track of all the traffic at a micro level using its patent-pending Rapid Traffic Sequencing (RTS) technology. RTS provides a framework for processing high volumes of traffic at very fine time resolution and generates bandwidth and QoS analysis results in real time. It enables:

• Service Channels. Pairs of CorvilNet appliances define Service Channels to measure the latency, jitter, and loss being experienced by traffic flowing between them. Service Channels measure one-way latency and jitter with microsecond precision in both directions for up to 100% of application packets.
• Microburst Detection. CorvilNet reveals network behaviour at ultra-short timescales by measuring traffic bit-rates over adjustable intervals down to 1 millisecond.
• Congestion-Aware Monitoring. Because transient network congestion is a common cause of unexpected latency, CorvilNet provides a built-in congestion-aware monitoring feature which calculates the amount of latency and loss measured traffic will induce at network bottlenecks.
• Bandwidth Sizing. CorvilNet removes the guesswork from network provisioning by directly computing the amount of bandwidth traffic needs to attain user-specified latency targets.
• Smart Packet Capture. CorvilNet's triggered packet capture facility makes sure that all network traffic is saved to disk before, during and after any anomalous performance event. CorvilNet also saves all of its performance measurements and analytics alongside the captured packets and provides a drill-down event analysis GUI to speed up trouble-shooting.
• TCP Micro-Analytics. CorvilNet monitors application response times, network and server delays, goodput, out-of-sequence and retransmitted packets, and makes these measurements available per-packet in drill-down analysis of any anomalous events.
• High-Speed Packet Measurement. A high-speed packet and data processing engine that uses hardware acceleration. This engine can process 10 Gbps rates and provides compact packet descriptors with nanosecond resolution to the subsequent analysis modules.