Arris International, Inc.

Arris Enterprises LLC

Arris Group, Inc.

Brian Jay Barker

Steven Paul Nolle

Jeffrey Joseph Howe

William Turner Hanks

Tom Cloonan

Alan Doucette

Keith Olsen

Greg J Gohman

Systems and methods can be operable to provide improved load balancing for network devices. In some implementations, such systems and methods can operate to create success list and/or fail list for use in optimizing improved load balancing for network devices. In other implementations, such systems and methods can also operate to manage and remove entries from success list and/or fail list. The success list and/or fail list can be used to optimize load balancing thereby reducing or preventing interruption in service to customers.

Load balancing for network devices

Different polling rates are applied to high priority BE service flows depending on whether the service flow contains, or is about to contain, traffic packets. A high priority service flow is identified upon boot-up of a user device, and defaults to a slow polling mode. A trigger causes the service flow to be serviced as a fast polling mode. The fast polling mode lasts for a predetermined period of time. If traffic continues at the end of the predetermined period of time, the predetermined period of time is reset so that the fast-polling mode continues until traffic flow on the high priority service flow ceases.

Method and system for providing bandwidth request opportunities for high priority service flows

CMTS software keeps track of the number of call requests accepted during a given monitoring period, this number being referred to as Count. As each call request is received, Count is compared to the current value stored in an overload state MIB (which can change multiple times during a monitoring period based on the actual overload state as reported by fault management). If Count exceeds the MIB value, the call request is denied. After the monitoring period ends, the Count of call requests accepted is reset to zero and the count value begins incrementing at the start of the next monitoring period as calls are accepted.

Method for processor overload control in a voice over internet protocol cable modem termination system

Static balancing of cable modems across upstream channels is made based on the channel's current bandwidth demand compared to a first and/or second CAC threshold level. If both threshold levels are exceeded, the modem is assigned to the channel having the lowest bandwidth demand. After registration, predictive balancing modems according to whether an MTA is part of a given modem avoids concentration of modems having MTAs on certain channels while other channels serve only moderns without MTAs. Modems are also predictively balanced according to whether they have associated a DSA_use_history profile. Modems associated with certain subscribers may be balanced according to the time of day balancing is occurring based on the profile. Thus, light users during working hours may be balanced as heavy users at night if they typically download video content or use VoIP features during the evening.

Growing community of inventors

Naperville, IL, United States of America

Brian Jay Barker