NQA Overview

NQA tests are performed between an NQA client and server. An NQA test instance can be configured manually or automatically delivered by an NMS. The NQA client NQA places a type of test instances into a specific test queue and triggers a specific test instance immediately, after a delay timer expires, or at a specified time.

The NQA client generates a protocol-specific packet for a test instance and timestamps the packet based on the local system time before sending it to the NQA server. The test packet must be greater than or equal to the minimum size defined in the specific protocol.

Upon receipt, the NQA server replies to the NQA client with a response packet carrying another timestamp. The NQA client obtains timestamps indicating when test packets were sent and received and calculates the round-trip time (RTT).

In a jitter test instance, both the NQA client and server timestamp each packet based on the local system time when they receive packets and respond to the packets. The NQA client can obtain timestamps to calculate the jitter time.

Users can query test results to learn network operation.

NQA Concepts

• NQA test instance

  An NQA test instance is created before an NQA test is performed. Some test parameters, such as the test type, destination address, and port number, can be configured for the created NQA test instance.

  Each test instance has an instance name and an administrator name. The two names uniquely identify a test instance.

• Test and probe

  An NQA test is initiated and performed at specified intervals. The interval between tests and the number of times the test is performed can be set.

  An NQA test performed each time comprises several consecutive probes. The number of probes in a test can be set.

• NQA client and server

  A device that initiates an NQA test is called the NQA client. NQA test instances are created on the NQA client.

  The NQA server processes the test packets sent by the NQA client. Figure 1 illustrates a network on which the NQA client and server are configured. The NQA server only responds to a test request carrying the specified destination IP address and port number. In most test instances, only the NQA client needs to be configured. In TCP, UDP jitter, and UDP test instances, the NQA server must also be configured.

  Figure 1 Relationships between the NQA client and server
  
  

NQA Characteristics

• Various test types and performance indicators

  Table 1 NQA test types

  Test Type	Description	Example
  Ping test	Checks network connectivity. Ping tests provide only a few performance indicators, and the test duration is short. Ping tests are configured to monitor network connectivity in real time.	Internet Control Message Protocol (ICMP), TCP, and label switched path (LSP) ping tests
  Trace test	Obtains path information. In a trace test, addresses of all reachable nodes along the path to a destination can be obtained. This helps quickly locate network problems.	Trace and LSP trace tests
  Jitter test	Measures network performance. Jitter tests provide a lot of performance indicators, such as the delay time, jitter time, packet loss ratio, and packet sequence number verification result, and the test duration is long. Therefore, jitter tests are usually used to analyze network performance.	UDP jitter, ICMP jitter, and LSP jitter tests

  NQA provides the following performance indicators:

  • Connectivity
  • Path information
  • Path MTU

  • Link performance

    • Delay time (one-way and round-trip delay time)
    • Jitter time
    • Packet loss ratio

• Flexible scheduling

  NQA provides a flexible scheduling mechanism to reduce its interaction with the NMS (this reduces the stress on the NMS) and helps efficiently use network bandwidth. A user can set the start time, end time, and interval for a test instance. In addition, the NQA scheduling mechanism automatically adjusts the time of a test instance based on the device workload to prevent the test instance from affecting other services running on the device.

  An NQA test instance starts in one of the following modes:

  • Start now: An NQA test instance starts immediately.
  • Start at: An NQA test instance starts at a specified time.
  • Start delay: An NQA test instance starts after a specified delay time elapses.
  • Start daily: An NQA test instance starts at a specified time every day.

  An NQA test instance stops in one of the following modes:

  • End at: An NQA test instance stops at a specified time.
  • End delay: An NQA test instance stops after a specified delay time elapses.
  • End lifetime: An NQA test instance stops a specified period of time after its startup.

  If the configuration does not meet the frequency verification formulas listed in Table 2, the test results may be inaccurate. When an NQA test instance is started, the system verifies the frequency configuration and displays messages for user operations. To ensure the accuracy of test results, modify the configuration as needed.

  Table 2 Frequency verification

  Test Instance Type	Frequency Verification Formula	Variable Description
  ICMP LSP ping MAC ping TCP UDP PWE3 ping SNMP VPLS ping VPLS PW ping	frequency > interval x (probeCount – 1) + timeout	frequency: interval between two execution times interval: interval at which packets are sent probeCount: number of test probes timeout: operation timeout period
  DNS	frequency > timeout	frequency: interval between two execution times timeout: operation timeout period
  ICMP jitter Jitter	Packet sending by software: frequency > interval x (jitter-packetnum – 1) + timeout Packet sending by hardware: frequency > ((NegotPktCount –1) x NegotInterval + NegotTimeOut) + ((TestPktCount – 1) x Testinterval + TestTimeOut) + ((InquireCount – 1) x InquireInterval + InquireTimeOut) + SenssionTimeout	jitter-packetnum: number of packets sent in each probe NegotPktCount: number of negotiation packets, which cannot be set and is fixed to 5 NegotInterval: interval at which negotiation packets are sent, which cannot be set and is fixed to 20 milliseconds NegotTimeOut: timeout period for negotiation packets, which cannot be set and is fixed to 3000 milliseconds TestPktCount: number of test packets, which is obtained by multiplying probe-count and jitter-packetnum Testinterval: interval at which test packets are sent, which is set using interval milliseconds or interval seconds TestTimeOut: timeout period for test packets, which is set using timeout InquireCount: number of inquiry packets, which cannot be set and is fixed to 5 InquireInterval: interval at which inquiry packets are sent, which cannot be set and is fixed to 20 milliseconds InquireTimeOut: timeout period for inquiry packets, which cannot be set and is fixed to 3000 milliseconds SenssionTimeout: timeout period for session table delivery, which cannot be set and is fixed to 3000 milliseconds
  LSP jitter	frequency > interval x (probeCount x jitter-packetnum – 1) + timeout	See the preceding variable descriptions.
  Path jitter	frequency > (MaxTTL – firstTTL + 1) x timeout (3000 ms) x probecount (3) + (MaxTTL – firstTTL + 1) x (interval x (probeCount x jitter-packetnum – 1) + timeout)	MaxTTL: maximum TTL firstTTL: initial TTL
  LSP trace PWE3 trace Path MTU Trace VPLS PW trace	frequency ≥ 60	See the preceding variable descriptions.