locked
High latency in NLB unicast mode RRS feed

  • Question

  • Hi Folks,

    I have setup Exchange organization with two HUC-CAS physical server which are managed by NLB in Unicast mode.

    facing issue with High latency on mailbox servers (database are equally distributed on both MBX server, mail queue's are normal, all ports are enable in NLB...)

    Please let me know to reduce traffic without any service failure.

    Thanks in Advance = Kapil

    Monday, February 9, 2015 10:17 AM

Answers

  • Hi Shweta,

    Thanks for your valuable comment but I have resolved issue after change NLB mode from Unicast to Multicast

    Unicast use single mac - address for traffic transition while multicast distribute the traffic between nodes fairly and Unicast use to broadcast the traffic as well on network.

    • Marked as answer by Kapil_Vats Tuesday, February 10, 2015 8:08 AM
    Tuesday, February 10, 2015 8:08 AM

All replies

  • Hi Kapil,

    According to your description, I understand you experience high latency in your Exchange server.
    If I misunderstand your concern, please do not hesitate to let me know.

    Regardless of their geographic location relative to other DAG members, each member of the DAG must have round trip network latency no greater than 500 milliseconds between each other member. Configurations with higher latency values may require special tuning of DAG, replication, and network parameters, such as increasing the number of databases or decreasing the number of mailboxes per database, to achieve the desired goals.

    Please try to run Outlook Connection Status and pay attention to “Avg Resp” parameter. Run Network Load Balancing control program to configure the setting of NLB, or run tools to monitor.
    For high latency for NLB, this may be related to several factors, for example network bandwidth, AD replication, PF frequency. More details about How Network Load Balancing Technology Works, for your reference: https://technet.microsoft.com/en-us/library/cc756878(v=ws.10).aspx

    Best Regards,
    Allen Wang

    Tuesday, February 10, 2015 3:40 AM
    Moderator
  • Hi,

    As per the information and details provided by you, to solve the high latency in NLB Unicode mode, please follow these steps: -

    During packet reception, Netork Load Balancing’s fully pipelined implementation overlaps the delivery of incoming packets to TCP/IP and the reception of other packets by the network adapter drive.

    This speeds up overall processing and reduces latency because latency because TCP/IP can process a packet while the network driver interface specification (NDIS) drive receives a subsequent packet. 

    It also reduces the overhead required for TCP/IP and the NDIS drive to coordinate their actions, and in many cases, it eliminates an extra copy of packet data in memory.

    During packet sending, Network Load Balancing also enhances throughput and reduces latency and overhead by increasing the number of packets that TCP/IP can send with one NDIS call.

    To achieve these performance enhancements, Network Load Balancing allocates and manages a pool of packet buffers and descriptors that it uses to overlap the actions of TCP/IP and NDIS driver.

    Network Load Balancing Performance.

    The performance impact of Network Load Balancing can be measured in four key areas: -

    1. CPU overhead
    2. Response time
    3. Throughput
    4. Switch occupancy

    I hope this information will be helpful for you.
    Thanks and regards
    Shweta@G 

    Tuesday, February 10, 2015 7:14 AM
  • Hi Shweta,

    Thanks for your valuable comment but I have resolved issue after change NLB mode from Unicast to Multicast

    Unicast use single mac - address for traffic transition while multicast distribute the traffic between nodes fairly and Unicast use to broadcast the traffic as well on network.

    • Marked as answer by Kapil_Vats Tuesday, February 10, 2015 8:08 AM
    Tuesday, February 10, 2015 8:08 AM