Methods, systems and computer readable media for active queue management

Abstract

The subject matter described herein includes methods, systems, and computer readable media for active queue management. A method for active queue management occurs at a network test system (NTS). The method includes transmitting, by a test packet generator and during a test session, a test packet to a system under test (SUT); receiving queue performance information associated with a SUT queue related to the test packet; generating, utilizing the queue performance information, control signaling information usable for controlling the test packet generator; and controlling the test packet generator using the control signaling information.

Description

TECHNICAL FIELD

The subject matter described herein relates to network testing. More particularly, the subject matter described herein relates to methods, systems, and computer readable media for active queue management.

BACKGROUND

Network operators typically test network nodes for reliability and other characteristics before deploying the network nodes to production environments (e.g., non-test environments). Generally, it is important to test networks nodes with various amounts of traffic and different types of traffic. For example, a test platform, such as an IxNetwork™ platform manufactured by Keysight, may be usable for network topology testing and traffic analysis and may generate test traffic for testing various network nodes using one or more protocols.

Data centers may be a term for distributed systems (e.g., multiple servers, switches, and/or other devices in same building) used for performing various functions. Within a data center, traffic can be significantly impacted by network switches and their related components. For example, queue related logic during heavy traffic periods can affect when and how traffic congestion is managed and/or can significantly impact system performance.

Issues can arise when testing a network switch center or a related queue. In particular, it may require significant time and testing resources to effectively test network switch queues and related logic during congestion events associated with various traffic scenarios, especially scenarios involving test and non-test traffic.

SUMMARY

The subject matter described herein includes methods, systems, and computer readable media for active queue management. A method for active queue management occurs at a network test system (NTS). The method includes transmitting, by a test packet generator and during a test session, a test packet to a system under test (SUT); receiving queue performance information associated with a queue related to the test packet; generating, utilizing the queue performance information, control signaling information usable for controlling the test packet generator; and controlling the test packet generator using the control signaling information.

A system for active queue management includes an NTS implemented using at least one processor. The NTS is configured for: transmitting, by a test packet generator and during a test session, a test packet to a SUT; receiving queue performance information associated with a queue related to the test packet; generating, utilizing the queue performance information, control signaling information usable for controlling the test packet generator; and controlling the test packet generator using the control signaling information. The subject matter described herein may be implemented in software in combination with hardware and/or firmware. For example, the subject matter described herein may be implemented in software executed by a processor. In one example implementation, the subject matter described herein may be implemented using a computer readable medium having stored thereon computer executable instructions that when executed by the processor of a computer control the computer to perform steps. Example computer readable media suitable for implementing the subject matter described herein include non-transitory devices, such as disk memory devices, chip memory devices, programmable logic devices, and application-specific integrated circuits. In addition, a computer readable medium that implements the subject matter described herein may be located on a single device or computing platform or may be distributed across multiple devices or computing platforms.

As used herein, the term “node” refers to at least one physical computing platform including one or more processors and memory.

As used herein, each of the terms “function”, “engine”, and “module” refers to hardware, firmware, or software in combination with hardware and/or firmware for implementing features described herein.

BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments of the subject matter described herein will now be explained with reference to the accompanying drawings, wherein like reference numerals represent like parts, of which:

FIG. 1 is a block diagram illustrating an example environment for active queue management;

FIG. 2 is a message flow diagram illustrating an example in-band process associated with active queue management;

FIG. 3 is a message flow diagram illustrating an example out-of-band process associated with active queue management; and

FIG. 4 is a flow chart illustrating an example process for active queue management.

DETAILED DESCRIPTION

In accordance with some aspects of the subject matter described herein, techniques, systems, methods, or mechanisms for active queue management are provided. For example, a network test system or a related entity in accordance with aspects described herein may be configured to test a network switch or other node with one or more network traffic related queues by using a servo or feedback controller to adjust a rate of test traffic being generated and sent to a system under test (SUT). In this example, the servo or feedback controller may utilize queue performance information (e.g., queue depth or utilization information, latency information, processing throughput information, and/or dropped packet counts) associated with a SUT queue, e.g., via in-band telemetry, an application programming interface (API), a test agent at the SUT, and/or other another entity.

Reference will now be made in detail to various embodiments of the subject matter described herein, examples of which are illustrated in the accompanying drawings. Wherever possible, the same reference numbers will be used throughout the drawings to refer to the same or like parts.

FIG. 1 is a diagram illustrating an example environment 100 for active queue management. Referring to FIG. 1, environment 100 may include a network testing system (NTS) 102 for performing various monitoring and/or testing operations. For example, NTS 102 may represent any suitable entity or entities (e.g., one or more testing platforms, nodes, or devices) associated with monitoring and/or testing one or more system(s) under testing (SUT) 118. In some embodiments, NTS 102 may include network monitoring and/or related analysis functionality.

In some embodiments, NTS 102 may include feedback controlled testing functionality. For example, NTS 102 or a related entity may generate and send test traffic (e.g., background traffic) to SUT 118 or related SUT queues during a test session. In this example, during the test session NTS 102 may obtain queue performance information and may use the queue performance information to generate control signaling information for dynamically or periodically adjusting or modifying aspects of testing, e.g., by reducing or increasing test traffic or by changing the makeup of the test traffic. Continuing with this example, NTS 102 or a related entity may monitor performance of SUT 118 and/or related SUT queues during the test session to evaluate various SUT performance aspects, e.g., programmability for queue related functionality, queue allocation logic, packet discard logic, and/or congestion related mechanisms.

In some embodiments, NTS 102 may perform testing of SUT 118 in various types of environments or scenarios. For example, NTS 102 may perform testing of SUT 118 in open-loop scenarios where SUT 118 receives test traffic and non-test traffic (e.g., production traffic from a live network). In another example, NTS 102 may perform testing of SUT 118 in a closed-loop scenarios where SUT 118 receives test traffic but no non-test traffic.

In some embodiments, NTS 102 may be a stand-alone tool, an in-line device, a testing device, a testing platform, a network tap, or software executing on at least one processor. In some embodiments, NTS 102 may be a single node or may be distributed across multiple computing platforms or nodes.

NTS 102 may include a user interface 104, a test system controller (TSC) 106, a traffic generator (TG) 108, a feedback controller and/or SUT analyzer (FCSA) 110, one or more communications interface(s) 112, a data stor age 114, and/or one or more network monitoring agents (NMAs), e.g., NMA 116.

In some embodiments, NTS 102 may provide user interface 104 for communicating with a test operator and/or another entity. For example, an analysis operator may represent any entity (e.g., an automated system or a device or system controlled or controllable by a human user) for selecting and/or configuring various aspects associated with configuring and/or executing one or more tests or for analyzing observed network information. For example, user interface 104, e.g., an application user programming interface (API) and a graphical user interface (GUI), may allow a user (e.g., a human or computer system) to provide and/or modify configuration information, such as test settings, SUT related information, network traffic latency emulation metrics, feedback control algorithms, traffic patterns, service emulation settings, etc. In some embodiments, user interface 104 may support automation (e.g., via one or more scripting languages), a representational state transfer (REST) or other API, a command line, and/or a web-based GUI.

TSC 106 may represent any suitable entity or entities (e.g., software executing on a processor, a field-programmable gateway array (FPGA), and/or an ASIC, or a combination of software, an FPGA, and/or an ASIC) for performing one or more aspects associated with testing SUT 118, controlling or configuring a test session or test system entities, and/or various aspects thereof. In some embodiments, TSC 106 may be implemented using one or more processors and/or memory. For example, TSC 106 may utilize one or more processors (e.g., executing software stored in memory) to generate traffic patterns or scenarios for various message streams (e.g., flows or sessions). In another example, TSC 106 may also utilize one or more processors to perform or initiate various tests and/or analyses involving test packets and/or related responses from SUT 118. In this example, TSC 106 may send instructions to various modules or entities, e.g., testing applications, TG 108, FCSA 110, and/or NMA 116, in NTS 102 for controlling (e.g., to pause, (re)start, or stop) a test session.

TG 108 may represent any suitable entity or entities (e.g., software executing on a processor, an ASIC, an FPGA, or a combination of software, an ASIC, or an FPGA) for performing one or more aspects associated with generating or synthesizing test sessions, test cases, or related test packets. For example, TG 108 may receive user input (e.g., a test intent or objective, like a test scenario declaration) via user interface 104 to obtain user intent or other information about a test case. In this example, TG 108 may use the user input and predefined test case templates or related data to generate one or more test cases and/or test sessions.

In some embodiments, TG 108 or a related entity may generate and/or use one or more test case templates or related test flows based on observed or predefined traffic patterns. For example, live intra-data center traffic flows may be monitored and captured (e.g., entire packets and/or flows may be captured, packet and/or flow meta data may be captured, etc.) and used, at least in part, to construct a traffic model that can be used to test SUT 118 or SUT queues therein. In this example, TG 108 or a related entity may also use topology information obtained from analysis of the observed traffic data and/or may be provisioned by a test operator or other entity.

FCSA 110 may represent any suitable entity (e.g., a servo implemented using a CPU or in a programmable datapath) or entities (e.g., software executing on a processor, an ASIC, an FPGA, or a combination of software, an ASIC, or an FPGA) for performing one or more aspects associated with analyzing queue performance information related to SUT 118 and/or for determining control signaling information for adjusting one or more aspects of TG 108.

In some embodiments, FCSA 110 may receive queue performance information from SUT 118 (e.g., in-band telemetry information and/or by reading ASIC registers out-of-band) and/or NMA 116. In this example, FCSA 110 may execute a feedback control algorithm (e.g., discrete-time control loop technique) that utilizes the queue performance information to generate control signaling information. Continuing with this example, the control signaling information may be utilized in controlling TG 108 for achieving a desired SUT state indicative of various SUT related metrics or values.

In some embodiments, a desired SUT state can represent one or more steady-state condition(s) or one or more time-varying condition(s). For example, during a test session for simulating well-known diurnal patterns of traffic congestion, a desired SUT state may be represented by conditions (e.g., queue depth values or packet drop counts) that vary or change over a particular time period. In this example, during the test session, FCSA 110 may periodically or dynamically generate control signaling information for controlling TG 108 or a related entity to generate and send test traffic to achieve or attempt to achieve the current desired SUT state.

In some embodiments, a feedback control algorithm may involve computing an error term, e.g., desired SUT state—actual (e.g., monitored or derived) SUT state), based at least in part on queue performance information associated with SUT 118. Continuing with this example, the feedback control algorithm may use the error term to identify various changes or adjustments (e.g., adjusting a traffic generation rate) needed to TG 108 for achieving the desired SUT state and may generate control signaling information for enacting these changes or adjustments at TG 108, e.g., by adjusting a traffic generation rate of TG 108.

In some embodiments involving open-loop scenarios, adjustments or changes to test traffic generation may consider a derived traffic rate associated with non-test traffic. For example, FCSA 110 may compute a derived non-test traffic rate by subtracting a known test traffic rate from a monitored total traffic rate at SUT 118 or a SUT queue therein. In this example, FCSA 110 may use the derived non-test traffic rate when determining control signaling information for adjusting or changing aspects of test traffic generation for achieving or attempting to achieve a set of desired SUT state characteristics.

In some embodiments involving closed-loop scenarios, adjustments or changes to test traffic generation may consider a known and/or a monitored test traffic rate associated with generated test traffic. For example, FCSA 110 may compute a derived SUT traffic rate by averaging a known test traffic rate and a monitored traffic rate at SUT 118 or a SUT queue therein. In this example, FCSA 110 may use the derived SUT traffic rate when determining control signaling information for adjusting or changing aspects of test traffic generation for achieving or attempting to achieve a set of desired SUT state characteristics. In another example, FCSA 110 may use the known test traffic rate or the monitored traffic rate when determining control signaling information for adjusting or changing aspects of test traffic generation for achieving or attempting to achieve a set of desired SUT state characteristics.

Communications interface(s) 112 may include or utilize any suitable entity or entities (e.g., one or more network interface cards (NICs), physical processors, pluggable jacks, physical processors, transceiver modules, direct-attach cables (DACs) and/or other hardware) for sending or receiving communications. For example, NTS 102 or FCSA 110 may use one or more communications interface(s) 112 for receiving and sending test packets, configuration instructions, observed traffic, queue performance information, or other data.

In some embodiments, communications interface(s) 112 may send or receive IP messages, Ethernet frames, Ethernet messages, packet data units (PDUs), datagrams, user datagram protocol (UDP) messages, TCP messages, IP version 4 (v4) messages, IP version 6 (v6) messages, stream control transmission protocol (SCTP) messages, real-time transport protocol (RTP) messages, or reliable data protocol (RDP) messages, messages using a tunneling protocol, and/or other data units.

In some embodiments, communications interface(s) 112 may include multiple port modules for interacting with SUT 118 and/or NMA 116. In some examples, a port module may include one or more transmit ports for sending test packets to SUT 118 or a queue thereof and one or more receive ports for receiving test packets back from SUT 118 or a queue thereof. In some examples, a port module may include one or more transmit ports for sending information to one or more SUT queues and one or more receive ports for receiving information from one or more SUT queues. In some examples, each port module or port(s) thereof may be assigned to a particular application, service, test flow, and/or IP address and port or for handling communications associated with a particular function or role.

NTS 102 and/or related entities may include functionality for accessing data stor age 114. Data stor age 114 may represent any suitable entity or entities (e.g., a stor age device, a non-transitory computer readable medium, or a stor age system) for maintaining or storing information related to testing and/or related metrics. For example, data stor age 114 may contain simulated test traffic, test cases, test session data, configuration settings, historical test session data, feedback control logic, performance metrics (e.g., statistics), and/or other information usable for generating control signaling information for controlling TG 108 or other aspects of testing, e.g., adjusting a test traffic rate or changing makeup of test traffic. In some embodiments, data stor age 114 may be located at NTS 102, FCSA 110, another node, or distributed across multiple platforms or devices.

SUT 118 may represent any suitable entity or entities (e.g., devices, systems, or platforms) for communicating with NTS 102 or related entities and/or receiving, processing, forwarding, and/or sending test traffic, non-test traffic, or other data. For example, SUT 118 may include a network router, a network switch, a traffic forwarding device, a NIC, a network node or related hardware containing one or more traffic related data queues (e.g., data structures for handling ingress or egress packets or frames). In another example, SUT 118 may include one or more systems and/or computing platforms, e.g., a data center or a group of servers and/or switches connected via a network.

In some embodiments, SUT 118 may represent one or more transit nodes (e.g., network switches) that are INT capable and/or can provide queue performance information, e.g., queue depth or utilization information, latency information, processing throughput information, and/or dropped packet counts. For example, SUT 118 may be configured for detecting INT instructions in packet headers or payloads and may insert into the packets INT metadata corresponding to the INT instructions. Example INT metadata may include a switch or node identifier (ID), ingress information (e.g., an ingress port ID, an ingress timestamp, etc.), egress information (e.g., an egress port ID, an egress timestamp, a hop latency, an egress port transmit link utilization value, etc.), buffer information (e.g., queue occupancy, queue congestion status, etc.), and/or other device related information. In some embodiments, INT metadata or other telemetry information may be stored in any portion of a packet, including, for example, a header, a payload, or a combination thereof. Additional details regarding INT is found in “In-band Network Telemetry (INT)” by Kim et al. (June 2016); the disclosure of which is incorpor ated herein by reference in its entirety. In another example, SUT 118 may provide an API or another mechanism for providing queue performance information, e.g., via in-band or out-of-band.

NMA 116 may represent any suitable entity or entities (e.g., software executing on a processor, an ASIC, an FPGA, or a combination of software, an ASIC, or an FPGA) for monitoring and/or providing performance related information about SUT 118 or related SUT queues. For example, NMA 116 may be configured to sniff traffic, filter, and/or classify traffic, and record metadata and precise timestamps for relevant packets. In this example, NMA 116 may capture entire packets or portions therein, e.g., headers, parameters, particular attribute value pairs (AVPs), payload portions, etc. In another example, NMA 116 may be capable of reading queue performance information from via an internal API or another mechanism. In another example, NMA 116 may be configured to include INT metadata containing queue performance information.

In some embodiments, NTS 102 and/or related entities may test SUT 118 or queues therein for determining queue related performance aspects before, during, and/or after a congestion event. For example, NTS 102 may test an internal processing queue associated with SUT 118 in an open-loop scenario, e.g., where SUT 118 or the processing queue is concurrently receiving some test traffic and some production (e.g., non-test) traffic. In this example, when uncongested, no packets from either the production or test packet flows may be dropped from the processing queue associated with SUT 118 due to queue saturation or overflow. However, when congested, the processing queue may fill completely and overflow, thereby causing packets to be dropped or discarded. In both scenarios, performance of the queue may be measured via various metrics, e.g., cross-switch latency times and queue depth levels.

In some embodiments, NTS 102 and/or related entities may perform traffic engineering when testing and evaluating performance aspects of SUT 118 or related queues. For example, in an open-loop scenario and prior to sending test traffic, NTS 102 may monitor SUT 118 to establish a baseline for performance aspects of an internal processing queue associated with SUT 118. In this example, after determining a performance baseline, NTS 102 and/or related entities may configure a queue, a port, a port group, and/or an ECMP group for rated traffic throughput and queueing allocations, may calculate background traffic mixes to produce the desired emulation of datacenter workloads which compete concurrently with a production workload of interest; and may systematically or dynamically vary the background traffic while monitoring the corresponding performance of SUT 118 along with the impact that the background traffic has on the production traffic transiting SUT 118, e.g., relative to the determined performance baseline.

In some embodiments, NTS 102 and/or related entities may test SUT 118 or queues therein using injection rates and drain rates. For example, prior to testing SUT 118 or a related queue, NTS 102 may be aware (e.g., from baseline testing or test operator input) of a drain rate (e.g., how quickly packets are removed from a queue) associated with an internal processing queue associated with SUT 118. In this example, NTS 102 may be aware (e.g., from baseline testing or test operator input) of an injection rate (e.g., how quickly packets are added to a queue and which may be strongly correlated to test traffic generation rate) associated with the internal processing queue associated with SUT 118. After determining these queue related rates, NTS 102 and/or related entities may configure a queue, a port, a port group, and/or an ECMP group for rated traffic throughput and queueing allocations and then send test traffic at a configured steady state test traffic rate for yielding an injection rate equivalent to the known drain rate such that a SUT queue maintains consistent queue occupancy, e.g., if the SUT queue was initially empty, then the SUT queue will maintain at most a fractional packet in it.

In some embodiments where test sessions are configured to utilize test traffic rates that yield an injection rate equivalent to a known drain rate, NTS 102 and/or related entities (e.g., TG 108) may send a batch or burst of extra packets as an “adjustment” event (e.g., higher than the steady state injection rate) which should raise the average depth level of the queue assuming the queue drains at a precise drain rate.

In some embodiments where test sessions utilize test traffic rates that yield an injection rate equivalent to a known drain rate, NTS 102 and/or related entities (e.g., TG 108) may send fewer packets as an “adjustment” event (e.g., lower than the steady state injection rate) which should reduce the average depth level of the queue assuming the queue drains at a precise drain rate.

It will be appreciated that FIG. 1 is for illustrative purposes and that various nodes and/or modules, locations, and/or functionality described above in relation to FIG. 1 may be changed, altered, added, or removed.

FIG. 2 is a message flow diagram illustrating an example in-band process 200 associated with active queue management. In some embodiments, SUT 118 may be configured to allow NTS 102 or a related entity to access or obtain queue performance information via one or more in-band processes or techniques (e.g., by adding or appending queue performance information to at least some packets that traverses SUT 118).

In some embodiments, accessing or obtaining queue performance information using an in-band process may not require communicating with one or more NMAs (e.g., NMA 116) deployed within SUT 118 and/or the surrounding testing environment. For example, SUT 118 may support one or more telemetry protocols that can be utilized by NTS 102 or another entity, e.g., FCSA 110.

In some embodiments, accessing or obtaining queue performance information using an in-band process may involve communicating with one or more NMAs (e.g., NMA 116) deployed within SUT 118 and/or the surrounding testing environment. For example, NMA 116 may be used in environments where SUT 118 does not support INT or does not support other in-band techniques for obtaining queue performance information.

In some embodiments, an NMA may be deployed in-line between NTS 102 and SUT 118 and act as an telemetry source (e.g., an entity that inserts INT headers, trailers, payloads, or other information to trigger INT metadata collection), while another NMA (e.g., NMA 116) may be deployed at or near SUT 118 and act as an telemetry sink (e.g., an entity that removes INT headers or trailers and INT metadata so as to make INT transparent to other layers or entities).

In some embodiments, during a test session or set of test sessions, FCSA 110 (e.g., a servo or feedback sub-system) may dynamically or periodically modify test traffic generation characteristics at TG 108 for achieving or attempting to achieve a set of desired SUT state characteristics. For example, desired SUT state characteristics may include an average queue depth for one or more of SUT queues; a workload latency value associated with one or more of SUT queues, a drop rate associated with one or more of SUT queues, a traffic or packet throughput value associated with one or more of SUT queues, etc.

Referring to FIG. 2, in step 201, TG 108 or a related entity may generate one or more test packets with telemetry instructions (e.g., in an INT header, a trailer, a payload, or one or more portions of a packet or frame) for triggering SUT 118 (e.g., a network switch) to collect and store telemetry information (e.g., INT metadata) associated with a SUT queue in at least some of the test packets as they transit SUT 118.

In step 202, NTS 102 or a related entity may send or forward the test packets to SUT 118.

In step 203, SUT 118 or a related entity (e.g., NMA 116) may receive the test packets, detect the telemetry instructions, and, in response to the telemetry instructions, insert telemetry information into the test packets. In some embodiments, inserted telemetry information may indicate various queue performance metrics or related data associated with one or more SUT queues.

In step 204, SUT 118 or a related entity may send or forward the test packets containing the telemetry information toward a destination (e.g., NTS 102 or FCSA 110).

In step 205, FCSA 110 or a related entity may receive the modified packets and extract the telemetry information. In some embodiments, NTS 102 or FCSA 110 may use the telemetry information for generating control signaling information to control TG 108 (e.g., by reducing or increasing a test traffic generation rate in an attempt to reach a desire level or depth at one or more queues of SUT 118 or by adjusting a test traffic pattern) and/or for other purposes. In some embodiments, FCSA 110 may strip or remove telemetry related information from the packets and then forward the “stripped” packets to another destination.

In some embodiments, test system traffic generation rates or other aspects of TG 108 may be controlled by FCSA 110 or a related feedback loop (servo) sub-system. For example, NTS 102 or a related entity (e.g., TSC 106) using operator input may set or define ideal or desired SUT queue state characteristics during a test session. In this example, FCSA 110 may dynamically or periodically during a test session or set of test session may compute an error term (e.g., desired SUT state—actual (e.g., monitored or derived) SUT state), based at least in part on the received telemetry information. Continuing with this example, FCSA 110 may use the error term, a predetermined table of possible adjustments and probable effects, classical discrete-time control-loop computations (e.g., a proportional-integral-derivative (PID) controller algorithm or other arbitrarily complex algorithms) and/or prediction logic to determine various changes or adjustments (e.g., adjusting a traffic generation rate) needed to TG 108 for achieving the desired SUT state, which may be outputted as or converted to control signaling information for enacting these changes or adjustments at TG 108.

In step 206, FCSA 110 or a related entity may send control signaling information to TG 108 for controlling TG 108. For example, control signaling information may trigger TG 108 to adjust a current test traffic generation rate during a test session. In another example, control signaling information may trigger TG 108 to adjust types of test packets generated, test traffic patterns, and/or traffic related characteristics (e.g., destination port numbers and/or IP addresses) during a test session.

It will be appreciated that FIG. 2 is for illustrative purposes and that different and/or additional steps other than those depicted in FIG. 2 may occur. Further, it will be appreciated that some steps may occur in a different order than depicted in FIG. 2 and that functionality described above in relation to FIG. 2 may be changed, altered, added, or removed.

FIG. 3 is a message flow diagram illustrating an example out-of-band process 300 associated with active queue management. In some embodiments, SUT 118 may be configured to allow NTS 102 or a related entity to access or obtain queue performance information via one or more out-of-band processes or techniques (e.g., by reading ASIC registers via a CPU or in the dataplane itself or another mechanism). For example, NMA 116 may periodically read, generate, and/or provide SUT queue performance information to FCSA 110 via an out-of-band connection, e.g., outside the datapath.

In some embodiments, accessing or obtaining queue performance information using an out-of-band process may not require communicating with one or more NMAs (e.g., NMA 116) deployed within SUT 118 and/or the surrounding testing environment. For example, SUT 118 may provide a network operating system (NOS) and/or a related API that can be accessed by NTS 102 or another entity, e.g., FCSA 110.

In some embodiments, accessing or obtaining queue performance information may involve communicating with one or more NMAs (e.g., NMA 116) deployed within SUT 118 and/or the surrounding testing environment. For example, NMA 116 may be used in environments where SUT 118 or a queue therein does not have a test system accessible API or is unable to support an out-of-band technique for obtaining queue performance information.

In some embodiments, during a test session or set of test sessions, FCSA 110 (e.g., a servo or feedback sub-system) may dynamically or periodically modify test traffic generation characteristics at TG 108 for achieving or attempting to achieve a set of desired SUT state characteristics. For example, desired SUT state characteristics may include an average queue depth for one or more of SUT queues; a workload latency value associated with one or more of SUT queues, a drop rate associated with one or more of SUT queues, a traffic or packet throughput value associated with one or more of SUT queues, etc.

Referring to FIG. 3, in step 301, TG 108 or a related entity may generate one or more test packets that will be added to a queue and potentially processed at SUT 118.

In step 302, NTS 102 or a related entity may send or forward the test packets to SUT 118.

In step 303, SUT 118 or a related entity (e.g., NMA 116) may receive and process the test packets.

In step 304, SUT 118 or a related entity may send or forward the test packets toward NTS 102 or a destination 299.

In step 305, FCSA 110 or a related entity may request, e.g., via an out-of-band connection, queue performance information from SUT 118 or another entity (e.g., NMA 116).

In step 306, SUT 118 or a related entity may provide the queue performance information to FCSA 110 or a related entity.

In step 307, NTS 102 or FCSA 110 may use the queue performance information for generating control signaling information to control TG 108 (e.g., by reducing or increasing a test traffic generation rate in an attempt to reach a desire level or depth at one or more queues of SUT 118 or by adjusting a test traffic pattern) and/or for other purposes.

In some embodiments, test system traffic generation rates or other aspects of TG 108 may be controlled by FCSA 110 or a related feedback loop (servo) sub-system. For example, NTS 102 or a related entity (e.g., TSC 106) using operator input may set or define ideal or desired SUT queue state characteristics during a test session. In this example, FCSA 110 may dynamically or periodically during a test session or set of test session may compute an error term (e.g., desired SUT state—actual (e.g., monitored or derived) SUT state), based at least in part on queue performance information received via an out-of-band process. Continuing with this example, FCSA 110 may use the error term, a predetermined table of possible adjustments and probable effects, classical discrete-time control-loop computations (such as PID controller algorithm or other arbitrarily complex algorithms) and/or prediction logic to determine various changes or adjustments (e.g., adjusting a traffic generation rate) needed to TG 108 for achieving the desired SUT state, which may be outputted as or converted to control signaling information for enacting these changes or adjustments at TG 108.

In step 308, FCSA 110 or a related entity may send control signaling information to TG 108 for controlling TG 108. For example, control signaling information may trigger TG 108 to adjust a current test traffic generation rate during a test session. In another example, control signaling information may trigger TG 108 to adjust types of test packets generated, test traffic patterns, and/or traffic related characteristics (e.g., destination port numbers and/or IP addresses) during a test session.

It will be appreciated that FIG. 3 is for illustrative purposes and that different and/or additional steps other than those depicted in FIG. 3 may occur. Further, it will be appreciated that some steps may occur in a different order than depicted in FIG. 3 and that functionality described above in relation to FIG. 3 may be changed, altered, added, or removed.

FIG. 4 is a diagram illustrating an example process 400 for active queue management. In some embodiments, process 400, or portions thereof, may be performed by or at NTS 102 (e.g., a network test system) and/or another node or module. In some embodiments, process 400 may include steps 402, 404, 406, and/or 408.

Referring to process 400, in step 402, a test packet may be transmitted, by a test packet generator and during a test session, to a SUT. For example, NTS 102 may execute a test session involving TG 108 sending test traffic to SUT 118 for testing queue performance aspects of SUT 118.

In step 404, queue performance information associated with a SUT queue related to the test packet may be received. For example, queue performance information may include latency information, processing throughput information, discard or drop information, or queue utilization information.

In some embodiments, receiving queue performance information associated with a SUT queue may occur using a out-of-band process. In such embodiments, the out-of-band process may request (e.g., by FCSA 110) queue performance information associated with the SUT queue and may receive (e.g., from SUT 118, a switch API, or NMA 116) the queue performance information.

In some embodiments, receiving queue performance information associated with a SUT queue may occur using a out-of-band process. In such embodiments, the in-band process may include sending a test packet containing INT instructions for instructing SUT 118 or a related entity (e.g., NMA 116) to provide INT metadata related to the SUT queue and wherein the queue performance information may include the INT metadata provided by SUT 118 or the related entity.

In some embodiments, a SUT queue may handle traffic received on a receive port of SUT 118, traffic sent in response to the traffic received on a transmit port of SUT 118, or traffic that may be similar to the test packet. For example, a SUT queue may handle traffic associated with a port number, a port group, or an equal-cost multi-path routing (ECMP) group.

In step 406, control signaling information usable for controlling the test packet generator may be generated utilizing the queue performance information. In some embodiments, generating the control signaling information may include computing a change for the test traffic rate in an attempt to achieve a desired traffic rate at SUT 118 based on a known test traffic rate and the queue performance information.

In some embodiments, e.g., where SUT 118 also receives and processes non-test traffic during the test session (e.g., an open-loop scenario), generating the control signaling information may include computing a change for the test traffic rate in an attempt to achieve a desired traffic rate at SUT 118 based on a known test traffic rate, a derived non-test traffic rate, and the queue performance information.

In step 408, the test packet generator may be controlled using the control signaling information. For example, controlling TG 108 may include adjusting a test traffic rate during a test session and/or adjusting types or characteristics of generated test traffic during the test session.

In some embodiments, SUT 118 may include a network switch, a traffic forwarding device, a NIC, an FPGA, an ASIC, or a processor.

It will be appreciated that process 400 is for illustrative purposes and that different and/or additional actions may be used. It will also be appreciated that various actions described herein may occur in a different order or sequence.

It should be noted that NTS 102, TSC 106, TG 108, FCSA 110, and/or functionality described herein may constitute one or more special purpose computing devices. Further, NTS 102, TSC 106, TG 108, FCSA 110, and/or functionality described herein can improve the technological field of testing traffic related queues in various network nodes (e.g., network switches) by providing various techniques for active queue management. For example, NTS 102, TSC 106, TG 108, FCSA 110, and/or functionality described herein may test a network switch or other node with one or more network traffic related queues by using a servo or feedback controller to adjust a rate of test traffic being generated and sent to SUT 118. In this example, FCSA 110 may utilize queue performance information (e.g., queue depth or utilization information, latency, and/or dropped packet counts) associated with a SUT queue, e.g., via in-band telemetry, an API, NMA 116, and/or other another entity.

It will be understood that various details of the subject matter described herein may be changed without departing from the scope of the subject matter described herein. Furthermore, the foregoing description is for the purpose of illustration only, and not for the purpose of limitation, as the subject matter described herein is defined by the claims as set forth hereinafter.

Claims

1. A method for active queue management, the method comprising: at a network test system (NTS): prior to a test session, determining a baseline for one or more performance metrics associated with performance of a system under test (SUT), wherein the one or more performance metrics include a drain rate associated with a SUT queue of the SUT;transmitting, by a test packet generator and during a test session, test packets to the SUT;receiving queue performance information associated with the SUT queue related to the test packets;generating, utilizing the queue performance information, a known test traffic rate, and the drain rate, control signaling information usable for controlling the test packet generator, wherein generating the control signaling information includes deriving a non-test traffic rate by subtracting the known test traffic rate from a monitored total traffic rate at the SUT and using the derived non-test traffic rate in generating the control signaling information; andcontrolling the test packet generator using the control signaling information to achieve or maintain an average queue depth value at the SUT queue.

2. The method of claim 1 wherein receiving the queue performance information associated with the SUT queue occurs using an out-of-band process comprising: requesting, by a feedback controller or a SUT analyzer associated with the NTS, queue performance information associated with the SUT queue; andreceiving, from the SUT, an application programming interface (API), or a network monitoring agent associated with the SUT, the queue performance information associated with the SUT queue.

3. The method of claim 1 wherein the test packet includes telemetry instructions for instructing the SUT or a related entity to provide telemetry information related to the SUT queue and wherein the queue performance information includes the telemetry information provided by the SUT or the related entity.

4. The method of claim 1 wherein controlling the test packet generator includes adjusting a test traffic rate or adjusting types or characteristics of generated test traffic.

5. The method of claim 4 wherein generating the control signaling information includes computing a change for the test traffic rate in an attempt to achieve a desired traffic rate at the SUT based on a known test traffic rate and the queue performance information.

6. The method of claim 1 wherein the queue performance information includes latency information, processing throughput information, discard or drop information, or queue utilization information.

7. The method of claim 1 wherein the SUT queue handles traffic received on a receive port of the SUT, traffic sent in response to the traffic received on a transmit port of the SUT, or traffic that is similar to the test packet.

8. The method of claim 1 wherein the SUT includes a network switch, a traffic forwarding device, a network interface card (NIC), a programmable gate array (FPGA), an application-specific integrated circuit (ASIC), or a processor.

9. A system for active queue management, the system comprising: at least one processor; anda network test system (NTS) implemented using the at least one processor, the NTS configured for: prior to a test session, determining a baseline for one or more performance metrics associated with performance of a system under test (SUT), wherein the one or more performance metrics include a drain rate associated with a SUT queue of the SUT;transmitting, by a test packet generator and during a test session, test packets to the SUT;receiving queue performance information associated with the SUT queue related to the test packets;generating, utilizing the queue performance information, a known test traffic rate, and the drain rate, control signaling information usable for controlling the test packet generator, wherein generating the control signaling information includes deriving a non-test traffic rate by subtracting the known test traffic rate from a monitored total traffic rate at the SUT and using the derived non-test traffic rate in generating the control signaling information; andcontrolling the test packet generator using the control signaling information to achieve or maintain an average queue depth value at the SUT queue.

10. The system of claim 9 wherein the NTS is configured for receiving the queue performance information associated with the queue using an out-of-band process comprising: requesting, by a feedback controller or a SUT analyzer associated with the NTS, queue performance information associated with the SUT queue; andreceiving, from the SUT, an application programming interface (API), or a network monitoring agent associated with the SUT, the queue performance information associated with the SUT queue.

11. The system of claim 9 wherein the test packet includes telemetry instructions for instructing the SUT or a related entity to provide telemetry information related to the SUT queue and wherein the queue performance information includes the telemetry information provided by the SUT or the related entity.

12. The system of claim 9 wherein controlling the test packet generator includes adjusting a test traffic rate or adjusting types or characteristics of generated test traffic.

13. The system of claim 12 wherein the NTS is configured for computing a change in the test traffic rate in an attempt to achieve a desired traffic rate at the SUT based on a known test traffic rate and the queue performance information.

14. The system of claim 9 wherein the queue performance information includes latency information, processing throughput information, discard or drop information, or queue utilization information.

15. The system of claim 9 wherein the SUT queue handles traffic received on a receive port of the SUT, traffic sent in response to the traffic received on a transmit port of the SUT, or traffic that is similar to the test packet.

16. The system of claim 9 wherein the SUT includes a network switch, a traffic forwarding device, a network interface card (NIC), a programmable gate array (FPGA), an application-specific integrated circuit (ASIC), or a processor.

17. A non-transitory computer readable medium having stored thereon executable instructions that when executed by at least one processor of at least one computer cause the at least one computer to perform steps comprising: at a network test system (NTS): prior to a test session, determining a baseline for one or more performance metrics associated with performance of a system under test (SUT), wherein the one or more performance metrics include a drain rate associated with a SUT queue of the SUT;transmitting, by a test packet generator and during a test session, test packets to the SUT;receiving queue performance information associated with the SUT queue related to the test packets;generating, utilizing the queue performance information, a known test traffic rate, and the drain rate, control signaling information usable for controlling the test packet generator, wherein generating the control signaling information includes deriving a non-test traffic rate by subtracting the known test traffic rate from a monitored total traffic rate at the SUT and using the derived non-test traffic rate in generating the control signaling information; andcontrolling the test packet generator using the control signaling information to achieve or maintain an average queue depth value at the SUT queue.

18. The non-transitory computer readable medium of claim 17 wherein receiving the queue performance information associated with the SUT queue occurs using an out-of-band process comprising: requesting, by a feedback controller or a SUT analyzer associated with the NTS, queue performance information associated with the SUT queue; andreceiving, from the SUT, an application programming interface (API), or a network monitoring agent associated with the SUT, the queue performance information associated with the SUT queue.