SMARTPolling™ is a proprietary access algorithm written into the operating system of all point-to-multipoint radios that accomplishes network efficiency, maximum bandwidth delivery and the lowest possible latency for fully-burdened networks at the lowest cost of ownership to service providers and enterprise operators. SMARTPolling™ was developed in conjunction with the development of the Access Series™ product line (5.3 and 5.8 GHz product offering) and then migrated to the M900S and M2400S Series™ of multipoint radios. SMARTPolling™ is an adaptive Time Division Duplex architecture that splits time into variable lengths through the polling process; it is considered a learning algorithm that actively monitors network demand and adjusts itself to accommodate the demand. In essence, the SMARTPolling™ process is a sequential talk/listen architecture that allocates more frequent polls to active subscribers and less polls to less active or dormant subscribers; its uniqueness is such that is pending U.S. and foreign patents.
In the polling process, the SMARTPolling™ algorithm utilizes 6 polling groups that each subscriber falls into. The highest priority polling group, Group 0, experiences the most frequent polls while Group 5, the lowest priority polling group, experiences the least number of polls per second. Subscribers move up or down in priority based on their bandwidth demand. For example, a subscriber that is simply surfing the Internet and doing the occasional hyperlink click will be in a lower priority group than a user that is engaged in a VoIP session via Skype. The elegance of the SMARTPolling™ algorithm is that time (which translates to bandwidth) is expended on those users that have actual data to send vs. the traditional polling method uses by other equipment providers that polls every subscriber regardless of network demand (thereby introducing unmanageable and unpredictable latency).
By focusing the microprocessor's energy on serving active users, network efficiency and latency is improved by over 20% relative to other competing solutions which translates into a 95% utilization of the available aggregate bandwidth. Comparatively, competitive solutions for a similar signaling rate achieve less than 40% efficiency; i.e., 4 Mbps for a 10 Mbps signal rate. Actual performance in the field with independent operators across 60 countries have shown this fact as networks grow over time. Network efficiency is the amount of time (which translates into usable bandwidth) expended on keeping subscribers served while still maintaining the capacity and ability to continue oversubscription. In fact, the SMARTPolling™ algorithm, to our knowledge is still the standard in network efficiency for over 5 years and continues to set the industry benchmark as noted by the SMARTBits™ test suite (SMARTBits™ is the equipment benchmark test bench of choice by Cisco, Foundry Networks and other leading Ethernet equipment providers). Further gains in network efficiency will be introduced in the coming months as SMARTPolling™ is addressed to our next generation of products and we will continue to set the standard by which competing solutions are measured.