Abstract: Speech recognition is performed utilizing a dynamically maintained set of grammar elements. A plurality of grammar elements may be identified, and the grammar elements may be ordered based at least in part upon contextual information. In other words, contextual information may be utilized to bias speech recognition. Once a speech input is received, the ordered plurality of grammar elements may be evaluated, and a correspondence between the received speech input and a grammar element included in the plurality of grammar elements may be determined.

Abstract: In some embodiments a network processor is described that includes cache to store metadata for individual packets. The cached metadata is utilized separately for forwarding and non-forwarding based functional units so that the forwarding and non-forwarding operations can be segregated. The cached metadata may be maintained after an associated individual packet is egressed so that non-forwarding operations can be completed after the associated packet has been egressed and hence impact to forwarding rate minimized. The cached metadata may be utilized separately by parallel processing paths. The parallel processing paths may be forwarding based so that parallel operations can be performed on metadata associated with an individual packet and the packet processing latency can be reduced. Other embodiments are otherwise disclosed herein.

Abstract: Packets may be classified into flows using direct lookup tables. The classification includes receiving a packet including a packet field having a corresponding field value. A direct lookup table (“DLT”) is indexed into at a DLT offset matching the field value to determine whether one or more of classification rules for classifying the packet into one or more flows are indexed at the DLT offset matching the field value. The DLT includes at least a portion of the classification rules indexed at any of multiple DLT offsets within the DLT according to at least one bit matching mask. In some cases, the packet field may be segmented into packet sub-fields having corresponding sub-field values and multiple strided DLTs are indexed into at DLT offsets matching the corresponding sub-field values to determine the matching classification rules for each of the sub-field values.

Abstract: An ordered classification procedure includes receiving a packet including a plurality of packet fields having corresponding field values. At least some of the packet fields are parsed to extract at least some of the field values of the packet. The extracted field values are examined using the ordered classification procedure to classify the packet into a flow. An order of the field examinations scheduled within the ordered classification procedure to examine the field values is based at least in part on the field values themselves.