SAN - IBM TotalStorage Switch ISL Trunking
Published 27 August 2002
Rate and comment
Authors: Jon Tate
Inter Switch Link (ISL) Trunking can dramatically improve the performance, manageability, and reliability for business-critical storage applications. By aggregating up to four ISLs into a single logical 8 Gb/s trunk group, this feature supports efficient high-speed communications throughout Storage Area Networks (SANs). By optimizing available switch resources, ISL Trunking decreases congestion. In addition, ISL Trunking reduces administrative workload since ISLs can be managed as a single entity rather than individually. Moreover, ISL Trunking can significantly increase data availability: even if an individual ISL failure occurs, I/O can continue at a reduced bandwidth as long as at least one ISL in the trunk group remains available. Together, these capabilities give organizations a powerful and flexible tool to extract the greatest performance and value from their SAN investments.
The ISL Trunking feature allows up to four Interswitch Links (ISLs) to merge logically into a single link. An ISL is a connection between two switches through an Expansion Port (E_Port). When using ISL Trunking to aggregate bandwidth of up to four ports, the speed of the ISLs between switches in a fabric is quadrupled. For example, at 2 Gb/s speeds, trunking delivers ISL throughput of up to 8 Gb per second. ISL Trunking supports high-bandwidth, large-scale SANs which include core switches. The primary task of ISL Trunking is to route data and edge switches that aggregate connections to servers and storage. ISL Trunking simplifies network design and reduces the cost of storage management by optimizing bandwidth utilization and enabling load balancing of traffic at the frame-level. The ISL Trunking feature has many advantages, for example, it ensures optimal ISL bandwidth use across trunked links, while preserving in-order delivery. ISL Trunking uses frame-level load balancing, as opposed to Fibre Channel Shortest Path First (FSPF), to achieve faster fabric convergence, as well as higher availability in the fabric.
Routing without the ISL Trunking feature
Prior to the implementation of the ISL Trunking feature, device-level load sharing was done through Fibre Channel networks that created ISLs and operated using the FSPF routing protocol. The FSPF routing protocol established and communicated the shortest paths for data to be carried from source to destination. Although FSPF compliant switches ensure fixed routing paths, and guarantee that all frames are delivered in order, congestion occurs if the aggregation of the stream exceeds the capacity of one of the ISLs in the path. For example, four untrunked ISLs have a maximum capacity of 2 Gb/s which provides for a maximum throughput of 8 Gb per second. Due to traffic that is not trunked, the throughput of the four ISLs is:
- 2 Gb/s + 1.5 Gb/s +.5 Gb/s + 1 Gb/s, which gives a 5 Gb/s total.
This is because two 2 Gb/s data streams are competing for the same path.
Routing with the ISL Trunking feature
With ISL Trunking, four ISLs provide 8 Gb/s of total throughput. With the implementation of ISL Trunking, bandwidth is shared across the trunked ISLs, permitting a total throughput of:
- 2 Gb/s + 1.5 Gb/s + 0.5 Gb/s + 1 Gb/s + 2 Gb/s, for a total 7 Gb/s in this case.
Because the trunk aggregates the four individual paths into one and preserves in-order deliver of frames, the total throughput is increased compared to a non-trunked group of ISLs. ISL Trunking dynamically performs load balancing, at the frame level, across a set of available links between two adjacent switches to establish a trunking group. Ports that belong to a trunking group are called trunking ports. One port is used to assign traffic for the group, and is referred to as the trunking master.
This material has not been submitted to any formal IBM test and is published AS IS. It has not been the subject of rigorous review. IBM assumes no responsibility for its accuracy or completeness. The use of this information or the implementation of any of these techniques is a client responsibility and depends upon the client's ability to evaluate and integrate them into the client's operational environment.
Follow IBM Redbooks
Follow IBM Redbooks