The present disclosure relates generally to providing selective build fabric (BF) and reconfigure fabric (RCF) frame flooding in a data network.
A fiber channel fabric typically uses a set of interconnected host bus adapters ((HBAs) which connects a host in/out bus to a computer terminal's memory), bridges, storage devices and switches. A storage area network (SAN) installation may include multiple fabrics for redundancy. Typically, a fabric may be considered as a set of switches sharing inter-switch links (ISLs) together with the devices to which they are connected. A SAN with multiple switches not connected by ISLs provide multiple fabrics.
Existing fiber channel fabrics have scalability limitations with the continued increase in the support of ports and virtual SANs (VSANs) on each switch. Indeed, when the principal link on a switch fails, there is a flood to fabric (Build-fabric) reconfiguration phase on all ports, all VSANs and all switches in the fabric. In turn, this leads to increase in the system load and may potentially result in control frame drops, instability, or large re-convergence times for the fabric on each VSAN.
More particularly, in a fibre-channel fabric, when a switch starts a build fabric (BF) phase, it will flood the network with BF frames by sending a BF_REQ frame through each interface and, for each interface, the switch will wait for a corresponding BF_RESP response frame. In a similar manner, when a switch starts a reconfigure fabric (RCF) phase, it will flood the network with RCF frames by sending a reconfigure request frame (RCF_REQ frame) through each connected interface, and also, for each interface, it will wait for a corresponding reconfigure response frame (RCF_RESP frame).
A method in particular embodiments may include detecting an initialization of a synchronization phase of a network device in a fiber channel fabric, performing synchronization phase optimization of the network device to select a predetermined number of interfaces, and transmitting a synchronization request on the predetermined number of interfaces.
A method in particular embodiments may include detecting an initialization of a synchronization phase of a network device in a fiber channel fabric, the synchronization phase including one of a build fabric phase or a reconfigure fabric phase, performing synchronization phase optimization of the network device to select a predetermined number of interfaces, transmitting a synchronization request on the predetermined number of interfaces, and receiving a synchronization acknowledgement on the predetermined number of interfaces.
An apparatus in particular embodiments may include a network interface, one or more processors coupled to the network interface, and a memory for storing instructions which, when executed by the one or more processors, causes the one or more processors to detect an initialization of a synchronization phase of a network device in a fiber channel fabric, perform synchronization phase optimization of the network device to select a predetermined number of interfaces of the network interface, and transmit a synchronization request on the predetermined number of interfaces.
An apparatus in particular embodiments includes means for detecting an initialization of a synchronization phase of a network device in a fiber channel fabric, means for performing synchronization phase optimization of the network device to select a predetermined number of interfaces, and means for transmitting a synchronization request on the predetermined number of interfaces.
These and other features and advantages of the present disclosure will be understood upon consideration of the following description of the particular embodiments and the accompanying drawings.
Referring again to
For example, in a fiber channel, device addressing may be handled in one of several ways. The first approach uses the worldwide name (WWN) of the device which is a 64-bit address that uniquely identifies each device globally to ensure that no duplicate WWNs are in the fiber channel network. Referring back to
In particular embodiments, during the BF or RCF synchronization phase of the switches in the fiber channel fabric, bandwidth and/or processing load optimization may be provided by, for example, performing a principal link only flooding, or alternatively performing a peer-aware flooding, as discussed in further detail below. More specifically, when the BF or RCF phase is initiated either automatically (for example, when a principal link goes down), or manually (for example, performed by a network administrator to apply one or more configuration parameters), the BF_REQ frames may be configured to send (and also to receive a corresponding BF_RESP frames) on the principal links only, or alternatively, a single BF_REQ frame may be sent to each peer switch (and a corresponding BF_RESP frame received) even if the peer switches are reachable via additional links or interfaces.
After successful BF or RCF synchronization phase, the principal switch selection may be performed as well as domain ID distribution in the fibre channel fabric.
Referring again to
In particular embodiments, as discussed in further detail below, the memory or storage unit 120A of the switch 120 may be configured to store instructions which may be executed by the processing unit 120C to detect an initialization of a synchronization phase of a network device in a fiber channel fabric, perform synchronization phase optimization of the network device to select a predetermined number of interfaces, and transmit a synchronization request on the predetermined number of interfaces.
Referring again to
Referring back to
In this manner, in particular embodiments, the synchronization phase optimization including, for example, build fabric (BF) or reconfigure fabric (RCF) phase in a fiber channel fabric may be performed, for example, but not limited to a subset of network device or switch interfaces that are associated with a predetermined world wide name (WWN) for example, or that are principal links. While there are other approaches for the synchronization phase optimization in determining a subset of interfaces to transmit BF or RCF request frames (and to receive a corresponding BF or RCF acknowledgement frames), in particular embodiments, two approaches are described herein. More specifically, in particular embodiments, one approach may be configured to send the synchronization frames only via the principal links, while another approach may be configured to send the synchronization frames based on a predetermined or randomly selected link from a number of links that have the same peer-WWN.
Referring back to
In this manner, in one aspect of the present disclosure, since a switch in a fibre channel fabric need only receive one single BF and RCF frame in order to enter the corresponding phase, the amount of bandwidth used by a fabric to propagate the BF and/or RCF frames to all switches in the network may be optimized by, for example, pruning redundant frames. More particularly, in accordance with one aspect of the present disclosure, optimization techniques for reducing redundant BF request (BF_REQ) or RCF request (RCF_REQ) frames sent by a switch in a fibre-channel fabric when the corresponding BF phase or the RCF phase are started, including, for example, selective flooding in the fibre-channel fabric during the BF or the RCF phase. In particular, in one aspect of the present disclosure, the selective flooding during the BF or the RCF phase may include configuring each switch in the fabric to send the BF or RCF request frames through a subset of its respective available interfaces, while insuring that all switches in the fabric receive at least one request frame.
Accordingly, in particular embodiments, the bandwidth and the processing load may be reduced in particular, when trunked ports are used to carry multiple VSANs. In this manner, a simple, efficient approach to optimizing bandwidth and processing load associated with BF or RCF phases in a fibre-channel fabric are provided, and which does not require new or additional frame or messaging.
A method in particular embodiments include detecting an initialization of a synchronization phase of a network device in a fiber channel fabric, performing synchronization phase optimization of the network device to select a predetermined number of interface, and transmitting a synchronization request on the predetermined number of interfaces.
The synchronization phase may include one of a build fabric phase or a reconfigure fabric phase.
Each of the predetermined number of interfaces of the network device may be a principal link.
In a further embodiment, each of the predetermined number of interfaces of the network device may be associated with a predetermined parameter, where the predetermined parameter may include a world wide name (WWN).
The method in particular embodiments may include receiving a synchronization acknowledgement on the predetermined number of interfaces.
Also, the predetermined number of interfaces may be a subset of the total number of interfaces of the network device in the fibre channel fabric.
A method in particular embodiments may include detecting an initialization of a synchronization phase of a network device in a fiber channel fabric, the synchronization phase including one of a build fabric phase or a reconfigure fabric phase, performing synchronization phase optimization of the network device to select a predetermined number of interfaces, transmitting a synchronization request on the predetermined number of interfaces, and receiving a synchronization acknowledgement on the predetermined number of interfaces.
An apparatus in particular embodiments may include a network interface, one or more processors coupled to the network interface, a memory for storing instructions which, when executed by the one or more processors, causes the one or more processors to detect an initialization of a synchronization phase of a network device in a fiber channel fabric, perform synchronization phase optimization of the network device to select a predetermined number of interfaces of the network interface, and transmit a synchronization request on the predetermined number of interfaces.
Further, each of the predetermined number of interfaces of the network device may be a principal link. Alternatively, each of the predetermined number of interfaces of the network interface may be associated with a predetermined parameter, where the predetermined parameter may include a world wide name (WWN).
The memory for storing instructions which, when executed by the one or more processors, may cause the one or more processors to receive a synchronization acknowledgement on the predetermined number of interfaces.
The predetermined number of interfaces may be a subset of the total number of interfaces of the network device in the fibre channel fabric.
An apparatus in particular embodiments may include means for detecting an initialization of a synchronization phase of a network device in a fiber channel fabric, means for performing synchronization phase optimization of the network device to select a predetermined number of interfaces, and means for transmitting a synchronization request on the predetermined number of interfaces.
The various processes described above including the processes performed by the switch 110 or switch 120 in the software application execution environment in the example system 100 including the processes and routines described in conjunction with
Various other modifications and alterations in the structure and method of operation of the particular embodiments will be apparent to those skilled in the art without departing from the scope and spirit of the disclosure. Although the disclosure has been described in connection with specific particular embodiments, it should be understood that the disclosure as claimed should not be unduly limited to such particular embodiments. It is intended that the following claims define the scope of the present disclosure and that structures and methods within the scope of these claims and their equivalents be covered thereby.
Number | Name | Date | Kind |
---|---|---|---|
4965788 | Newman | Oct 1990 | A |
5150360 | Perlman et al. | Sep 1992 | A |
6032194 | Gai et al. | Feb 2000 | A |
6388995 | Gai et al. | May 2002 | B1 |
6400681 | Bertin et al. | Jun 2002 | B1 |
6535491 | Dutt et al. | Mar 2003 | B2 |
6697339 | Jain | Feb 2004 | B1 |
6882630 | Seaman | Apr 2005 | B1 |
6976088 | Gai et al. | Dec 2005 | B1 |
7428658 | Nagin et al. | Sep 2008 | B2 |
7457322 | Flood et al. | Nov 2008 | B1 |
20010056503 | Hibbard | Dec 2001 | A1 |
20020181503 | Montgomery, Jr. | Dec 2002 | A1 |
20030123389 | Russell et al. | Jul 2003 | A1 |
20030218986 | DeSanti et al. | Nov 2003 | A1 |
20050265260 | Zinin et al. | Dec 2005 | A1 |
20070207591 | Rahman et al. | Sep 2007 | A1 |
20080072000 | Osaki et al. | Mar 2008 | A1 |
20090141700 | Baldemair et al. | Jun 2009 | A1 |
Number | Date | Country | |
---|---|---|---|
20090052474 A1 | Feb 2009 | US |