Abstract

Dyband commissioned The Tolly Group to perform a functional evaluation of its Dyband bandwidth management solution.

CrossKeys Dyband 2.0 is presented in this Tolly evaluation as bandwidth management software designed for service providers that need fine-grained control over subscriber traffic, congestion, and service-tier enforcement. The report highlights Dyband’s ability to control bandwidth both per individual user and across hierarchical aggregation points, enabling operators to apply guaranteed, preferred, and equitable allocation models across shared network infrastructure. A rate-ramping mechanism is also included to prevent aggressive users from monopolizing available capacity by reducing sustained usage when demand rises.  

The product also provides real-time monitoring and historical reporting intended to support both immediate troubleshooting and longer-term capacity planning. According to the report, performance statistics include average rate, peak rate, congestion percentage, and congestion delay, with visibility across 60 one-second intervals, 60 one-minute intervals, and 24 one-hour intervals. Subscriber reports can be viewed in real time, while broader historical analysis can be performed for the overall network, aggregation points such as groups or subnets, and selected subscribers.  

Tolly’s testing also emphasized automation and resiliency. Dyband automatically discovered network elements and new subscribers, displayed them in a hierarchical topology, and supported policy inheritance through service profiles, allowing operators to define and propagate bandwidth rules efficiently. The report states that this functionality worked in both DHCP and statically addressed environments. In fault-tolerance testing, traffic shaping was transferred from an active unit to a standby unit within 10 seconds after a simulated hardware failure, allowing bandwidth control and performance statistics collection to continue.  

In the bandwidth-control test, a 33-Mbyte FTP download was limited first to 400 Kb/s, completing in 718 seconds, then increased to 800 Kb/s, cutting the transfer time to 359 seconds. A midstream rate change from 400 Kb/s to 800 Kb/s was reflected immediately in the statistics display and validated by the shortened transfer duration. The test bed diagram on page 4 shows a redundant Dyband deployment connected through 100Base-TX Ethernet hubs, a Cisco 2524 router, an FTP proxy server, and grouped client systems.