iSCSI SAN Design Considerations

  1. A basic difference between iSCSI and Fibre Channel is that when an iSCSI path is overloaded, the TCP/IP protocol drops packets and requires them to be resent. Fibre Channel communications over a dedicated path are not at risk of being overloaded. When a network path carrying iSCSI storage traffic is oversubscribed, a bad situation quickly grows worse and performance further degrades as dropped packets must be resent.
  2. Another potential disadvantage with implementing software-initiator iSCSI (but not hardware-initiator iSCSI) is that standard 10/100 Ethernet interfaces do not have enough throughput for practical iSCSI work. Gigabit Ethernet interfaces are required, and those interfaces tend to consume large amounts of CPU time. One way of overcoming this demand for CPU resources is to use TOEs (TCP/IP offload engines).
  3. TCP/IP offload engines shift TCP packet processing tasks from the server CPU to specialized TCP processors on the network adapter or storage device. The QLA4050 hardware initiator, which is supported on ESX Server 3.0, uses a TOE.
  4. iSCSI does not work well over most shared wide area networks. one should consider iSCSI a local-area technology, not a wide-area technology, because of latency issues and security concerns.
  5. One should also segregate iSCSI traffic from general traffic. Layer 2 VLANs are a particularly good way to implement this segregation.
  6. Best practice is to have a dedicated LAN for iSCSI traffic and not share the network with other network traffic.
    It is also best practice not to oversubscribe the dedicated LAN.
  7. f

Leave a Reply

Your email address will not be published. Required fields are marked *