CN104796276B - A kind of link switch-over method and system - Google Patents

Abstract

The invention discloses a link switching method and system, which are used to solve the problem in the prior art that a blade host will not actively switch when a currently used link fails, resulting in interruption of data transmission. In this solution, there are active IB Switch modules and standby IB Switch modules in the blade host, and there is a link between any IB Switch module and each IB Switch in the IB subnet where the blade host is located. When the blade host determines that the currently used first link fails according to the predetermined network topology structure of the IB subnet where it is located, it determines the second link; wherein, the first link and the The second links are all connected to the main IB Switch module in the blade host; switching from the first link to the second link enables the blade host to use the second link to transmit data.

Description

A link switching method and system

technical field

The present invention relates to the technical field of information and service support, in particular to a link switching method and system.

Background technique

Interconnect structure (InfiniBand, referred to as IB) network is an open standard high-bandwidth, high-speed network interconnection technology, this technology is not used for general network connection, its main design purpose is to solve the problem of server-side connection. Therefore, InfiniBand technology is mainly applied to servers and servers, servers and storage devices (such as storage area network (Storage Area Network, SAN) and direct storage attachments) and between servers and networks (such as local area network (Local Area Network, LAN), Wide Area Network (Wide Area Network, WAN) and Internet (internetwork, referred to as the Internet)) communication.

As shown in Figure 1, it is a topological structure diagram of an existing IB network. Wherein, the two network cards in the host are respectively connected to two different switches (that is, there is only one link between the host and each switch), and the two switches are cascaded. Network card binding is performed on the host side. The binding mode is active-standby mode (Active-Standby mode). When the active link fails, the standby link takes over the failed active link and changes from standby mode to active mode at the same time. , to ensure that data transmission will not be interrupted due to a link failure during data transmission. Referring to Figure 1, when the link or network card corresponding to 9000 fails, the data transmission is switched to the link and network card corresponding to 9002.

For the above-mentioned network topology, for traditional hosts that use the Peripheral Component Interconnect (PCI) standard, once the currently used link fails, the host will be notified of the network failure through the uplink detection mechanism of the IP protocol. , so when the main switch fails, the data transmission can be switched to the standby switch, and when the main network card or the link corresponding to the main network card fails, the data transmission can be switched to the standby network card and the corresponding link.

However, for a host (referred to as a blade host) using a blade architecture (multiple card-type server units inserted in a standard-height rack cabinet to achieve high availability and high density), if the blade host is applied to In the above-mentioned traditional network topology, once the currently used link fails, the uplink detection mechanism of the IP protocol will also notify the blade host that the network has failed. However, because the internal hardware architecture of the blade host is different from the traditional host architecture, its blade The host channel adapter (Host Channel Adapter, HCA) card is integrated on the blade host. The HCA card is connected to the IBSwitch (interconnect structure switch) module inside the blade host, and the IB Switch module does not have the function of the uplink detection mechanism, so no matter the link currently used Whether a fault occurs, the connection port between the HCA card and the IB Switch module on the blade is always connected (up), so the blade host cannot be informed of network faults. In this case, the blade host will always send data to the link corresponding to the active IBSwitch module, and the network service cannot be switched from the active IBSwitch module to the standby IBSwitch module, resulting in interruption of data transmission.

Contents of the invention

Embodiments of the present invention provide a link switching method and system to solve the problem in the prior art that a blade host does not actively switch when a currently used link fails, resulting in interruption of data transmission.

Embodiments of the present invention adopt the following technical solutions:

A link switching method, a main interconnection structure switch IB Switch module and a standby IBSwitch module exist in a blade host, and there is a link between any IB Switch module and each IB Switch in the IB subnet where the blade host is located way, said method includes:

When the blade host judges that the first link currently used fails according to the network topology structure of the IB subnet determined in advance, it determines the second link; wherein, the first link and the The second links are all connected to the main IB Switch module in the blade host;

Switching from the first link to the second link enables the blade host to use the second link to transmit data.

Wherein, the method also includes:

When the blade host determines that the main IB Switch module connected to the first link fails, determine the link connected to the standby IB Switch module in the blade host;

switching from the first link to the link connected to the standby IB Switch module, so that the blade host can use the link connected to the standby IB Switch module to transmit data.

Wherein, there are at least two links connected to the standby IB Switch module; then

Switching from the first link to the link connected to the standby IB Switch module, so that the blade host can use the link connected to the standby IB Switch module to transmit data, specifically includes:

Determine the weight of each link connected to the standby IB Switch module;

Switching from the first link to the link with the highest weight connected to the standby IB Switch module, so that the blade host can use the link with the highest weight connected to the standby IB Switch module to transmit data.

Wherein, the network topology is determined as follows:

The subnet manager SM in the IB subnet detects the nodes in the IB subnet by sending subnet detection packets to the IB subnet, and generates the network topology of the IB subnet according to the detection results .

Wherein, each IB Switch in the IB subnet is cascaded, and for each IBSwitch module in the blade host:

Links existing between the current IB Switch module and each IB Switch in the subnet where the blade host is located are bound into a virtual link.

A link switching system, applied in an interconnection structure IB subnet, said system comprising:

At least two interconnection structure switches IB Switch and at least one blade host, where there are a master IB Switch module and a standby IB Switch module in the blade host, and there is a link between any IB Switch module and each IB Switch, wherein:

The blade host is configured to determine a second link when it is determined that the currently used first link fails according to the predetermined network topology structure of the IB subnet; switch from the first link to the second link, so that the blade host can use the second link to transmit data; wherein, both the first link and the second link are connected to the main IB Switch in the blade host module connection.

Wherein, the blade host is also used for:

When it is determined that the main IB Switch module connected to the first link fails, determine the link connected to the standby IB Switch module in the blade host; switch from the first link to the link connected to the standby IB The link connected to the Switch module enables the blade host to use the link connected to the standby IBSwitch module to transmit data.

Wherein, there are at least two links connected to the standby IB Switch module; then

The blade host is specifically used for:

Determine the weight of each link connected to the standby IB Switch module; switch from the first link to the link with the largest weight connected to the standby IB Switch module, so that the blade host can use the The link with the largest weight connected to the backup IBSwitch module transmits data.

Wherein, the system also includes:

The subnet manager SM is configured to detect the nodes in the IB subnet by sending subnet detection packets to the IB subnet, and generate the network topology of the IB subnet according to the detection results.

Wherein, each IB Switch in the IB subnet is cascaded, and for each IBSwitch module in the blade host:

Links existing between the current IB Switch module and each IB Switch in the subnet where the blade host is located are bound into a virtual link.

The beneficial effects of the embodiments of the present invention are as follows:

In the embodiment of the present invention, when the first link connected to the main IB Switch module currently used by the blade host fails, it can be switched to the second link connected to the main IB Switch module without switching the main and standby IB Switch modules. This method of adding redundant links in the IB subnet solves the problem in the prior art that the blade host will not actively switch when the currently used link fails, resulting in interruption of data transmission .

Description of drawings

Figure 1 is a diagram of the existing IB network topology;

FIG. 2 is an implementation flowchart of a link switching method provided by an embodiment of the present invention;

FIG. 3 is a schematic diagram of implementing a link switching method provided in an embodiment of the present invention in practical applications;

FIG. 4 is a schematic structural diagram of a link switching system provided by an embodiment of the present invention.

Detailed ways

In order to solve the problem in the prior art that the blade host does not actively switch when the currently used link fails, resulting in interruption of data transmission, embodiments of the present invention provide a link switching method and system. In this technical solution, when the first link connected to the main IB Switch module currently used by the blade host fails, it can switch to the second link connected to the main IB Switch module without switching the main and standby IB Switch modules. This method of adding redundant links in the IB subnet solves the problem in the prior art that the blade host will not actively switch when the currently used link fails, resulting in interruption of data transmission .

The embodiments of the present invention will be described below in conjunction with the accompanying drawings. It should be understood that the embodiments described here are only used to illustrate and explain the present invention, and are not intended to limit the present invention. And in the case of no conflict, the embodiments and the features of the embodiments in this description can be combined with each other.

The embodiment of the present invention provides a link switching method, as shown in Figure 2, which is a flow chart for the implementation of the method, wherein, there are a main IB Switch module and a standby IB Switch module in the blade host, and any one IB Switch module and There is a link between each IB Switch in the IB subnet where the blade host is located, and the method specifically includes the following steps:

Step 21, when the blade host determines that the first link currently in use fails according to the network topology of the IB subnet where it is located, determine the second link; the first link and the second link Both are connected to the main IB Switch module in the blade host;

In the embodiment of the present invention, the network topology of the IB subnet where the blade host is located can be determined in the following manner:

The main subnet manager SM in the IB subnet detects the nodes in the IB subnet by sending subnet detection packets to the IB subnet, and generates the network topology structure of the IB subnet according to the detection results.

Among them, SM is one of the common functions in the IB network, which realizes the centralized management of the IB subnet. The most commonly used type of subnet detection packets sent by SM is directional routing packets. There is an initial path field (init_path) in the directional routing packet, which indicates the path of the packet, and a hop count field (hop_count), which indicates the number of hops the packet passes through. , there is a return path field (return_path), which records the path of the response packet. Relying on the above information, the subnet detection packet can complete all the topology search and initialization work. The init_path and return_path records are actually port numbers, which indicate the port from which the current node should send the packet.

The detection process of the subnet manager is mainly as follows: SM first sends a Get(Nodeinfo) subnet management detection packet SMP to the local node where it resides. node, it will be added to the topology database. If the node has been detected, SM will obtain the port status information of the node through Get (Portinfo) SMPs. If it is a CA port, it will be added to the port list. If it is the management port of the switch, Then configure the port, set parameters, and then add the discovered port to the entry of the port object. Then, through the obtained port status information, the SM judges whether the other end of the port is connected to other devices. If it is connected to other devices, it will continue to send the SMP of Get (nodeinfo) for detection. If not, it will judge whether the detection is over. If not, continue to detect , if it ends, the detection process ends.

Generally, SM can exist on any port of any CA, switch or router, and in order to prevent a single point of failure, there can be multiple SMs in the IB subnet, one primary SM and multiple backup SMs, and the backup SM detects When the primary SM dies, a backup SM is selected as the primary SM to take over the management right of the subnet.

Step 22, switch from the first link to the second link, so that the blade host can use the second link to transmit data.

Since both the first link and the second link are connected to the main IB Switch module in the blade host, the IB Switch module is not actually switched. The first link and the second link can also be pre-bound into a virtual link through the dual-link binding technology, that is, no matter whether the first link is used to transmit data or the second link is used to transmit data. Data is transmitted through the bound virtual link.

In the embodiment of the present invention, when the first link connected to the main IB Switch module currently used by the blade host fails, it can be switched to the second link connected to the main IB Switch module without switching the main and standby IB Switch modules. This method of adding redundant links in the IB subnet solves the problem in the prior art that the blade host will not actively switch when the currently used link fails, resulting in interruption of data transmission .

As shown in FIG. 3 , it is a schematic diagram of realizing the above-mentioned link switching method in practical application. Among them, there are two IB Switch modules in the blade host 1, one is the main IB Switch module, and the other is the standby IB Switch module. For link 9001, a link 9003 is added between IBSwitch2 and the main IB Switch module of blade host 1, and IB Switch1 and IBSwitch2 are cascaded.

Referring to Figure 3, when the IB subnet is initialized, the SM performs node detection on the entire IB subnet to generate a network topology, and within the blade host 1, links 9000 and 9003 are bound into A virtual link, links 9001 and 9002 are bound to form another virtual link.

When the link 9000 between IB Switch1 and the active IB Switch module of blade host 1 fails, because the IB Switch module does not have an uplink detection mechanism, the active IB Switch module and the standby IB Switch module will not switch over, but the SM will Node detection is performed according to the network detection process again, and a new network topology is regenerated according to the preset routing algorithm. Thus, the link 9003 connected to the main IB Switch module is determined, and the link 9003 is switched to enable data to be transmitted from the standby IB Switch module of the blade host 1 to the IB Switch2.

Therefore, in this IB subnet, when the currently used link is interrupted, the network service of the blade host will not be interrupted, thereby achieving the purpose of ensuring safe and stable data transmission.

Furthermore, under this IB network architecture, when the blade host judges that the main IBSwitch module linking links 9000 and 9003 fails, it will switch to the link connected to the standby IB Switch module, such as transferring data from the link to Link 9000 is switched to link 9002 to ensure uninterrupted data transmission.

Since there are two links connected to the standby IB Switch module, you can first determine the weight of each link connected to the standby IB Switch module, and then switch to the link with the highest weight connected to the standby IB Switch module. transfer data.

corresponding. The embodiment of the present invention also provides a link switching system, as shown in FIG. 4 , which is a schematic structural diagram of the system. The system is applied to the IB subnet of the interconnection structure, including:

At least two interconnected structure switches IB Switch41 and at least one blade host 42, there are main IB Switch module and standby IB Switch module in the blade host 42, and there is a link between any IB Switch module and each IBSwitch41, wherein :

The blade host 42 is configured to determine a second link when it is determined that the currently used first link fails according to the predetermined network topology structure of the IB subnet; from the first link switch to the second link, so that the blade host can use the second link to transmit data; wherein, both the first link and the second link are connected to the master IB in the blade host Switch module connection.

Wherein, the blade host 42 is also used for:

When it is determined that the main IB Switch module connected to the first link fails, determine the link connected to the standby IB Switch module in the blade host; switch from the first link to the link connected to the standby IB The link connected to the Switch module enables the blade host to use the link connected to the standby IB Switch module to transmit data.

Wherein, when there are at least two links connected with the standby IB Switch module, the blade host 42 is specifically used for:

Determine the weight of each link connected to the standby IB Switch module; switch from the first link to the link with the largest weight connected to the standby IB Switch module, so that the blade host can use the The link with the largest weight connected to the backup IBSwitch module transmits data.

Further, the system may also include:

The subnet manager SM is configured to detect the nodes in the IB subnet by sending subnet detection packets to the IB subnet, and generate the network topology of the IB subnet according to the detection results.

Wherein, each IB Switch41 in the IB subnet is cascaded, and for each IB Switch module in the blade host 42:

The links existing between the current IB Switch module and each IB Switch 41 in the subnet where the blade host 42 is located are bound into a virtual link.

Those skilled in the art should understand that the embodiments of the present invention may be provided as methods, systems, or computer program products. Accordingly, the present invention can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each procedure and/or block in the flowchart and/or block diagram, and combinations of procedures and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions may be provided to a general purpose computer, special purpose computer, embedded processor, or processor of other programmable data processing equipment to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing equipment produce a Means for realizing the functions specified in one or more steps of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the instructions The device realizes the function specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions can also be loaded onto a computer or other programmable data processing device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby The instructions provide steps for implementing the functions specified in the flow chart flow or flows and/or block diagram block or blocks.

While preferred embodiments of the invention have been described, additional changes and modifications to these embodiments can be made by those skilled in the art once the basic inventive concept is appreciated. Therefore, it is intended that the appended claims be construed to cover the preferred embodiment as well as all changes and modifications which fall within the scope of the invention.

Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and equivalent technologies thereof, the present invention also intends to include these modifications and variations.

Claims (6)

1. a link switching method, it is characterized in that, there is main interconnection structure exchange IB Switch module and standby IB Switch module in the blade host, and any one IB Switch module and each in the IB subnet where described blade host is located Links exist between IB Switches, and the methods include: When the blade host judges that the first link currently in use fails according to the network topology of the IB subnet where it is located, it determines the second link and switches from the first link to the The second link enables the blade host to use the second link to transmit data; wherein both the first link and the second link are connected to the main IB Switch module in the blade host; When the blade host determines that the main IB Switch module connected to the first link fails, determine the link connected to the standby IB Switch module in the blade host, and switch from the first link to the link connected to the first link. The link connected to the standby IBSwitch module enables the blade host to use the link connected to the standby IB Switch module to transmit data; Wherein, the network topology is determined as follows: the subnet manager SM in the IB subnet detects the nodes in the IB subnet by sending a subnet detection packet to the IB subnet, and according to The detection result generates the network topology of the IB subnet. 2. The method according to claim 1, wherein there are at least two links connected with the standby IB Switch module; then Switching from the first link to the link connected to the standby IB Switch module, so that the blade host can use the link connected to the standby IB Switch module to transmit data, specifically includes: Determine the weight of each link connected to the standby IB Switch module; Switching from the first link to the link with the highest weight connected to the standby IB Switch module, so that the blade host can use the link with the highest weight connected to the standby IB Switch module to transmit data. 3. The method according to claim 1 or 2, wherein each IB Switch in the IB subnet is cascaded, and for each IB Switch module in the blade host: Links existing between the current IB Switch module and each IB Switch in the subnet where the blade host is located are bound into a virtual link. 4. A link switching system, characterized in that it is applied in the interconnection structure IB subnet, the system includes at least two interconnection structure switches IB Switch and at least one blade host, and there is a main IB Switch module in the blade host and standby IB Switch modules, and there is a link between any IB Switch module and each IB Switch, where: The blade host is configured to determine a second link and switch from the first link when it is determined that the currently used first link fails according to the predetermined network topology structure of the IB subnet to the second link, so that the blade host can use the second link to transmit data; wherein, both the first link and the second link are connected to the main IB Switch in the blade host module connection; The blade host is further configured to, when it is determined that the main IB Switch module connected to the first link fails, determine the link connected to the standby IB Switch module in the blade host, from the first link Road switching to the link connected to the standby IB Switch module, so that the blade host can use the link connected to the standby IB Switch module to transmit data; The system also includes: a subnet manager SM, configured to detect nodes in the IB subnet by sending a subnet detection packet to the IB subnet, and generate a network of the IB subnet according to the detection results Topology. 5. system as claimed in claim 4, is characterized in that, the link that is connected with described spare IB Switch module has at least two; Then The blade host is specifically used for: Determine the weight of each link connected to the standby IB Switch module; switch from the first link to the link with the largest weight connected to the standby IB Switch module, so that the blade host can use the The link with the largest weight connected to the backup IB Switch module transmits data. 6. The system according to claim 4 or 5, wherein each IB Switch in the IB subnet is cascaded, and for each IB Switch module in the blade host: Links existing between the current IB Switch module and each IB Switch in the subnet where the blade host is located are bound into a virtual link.