CN108259336B

CN108259336B - Data center interconnection method and device

Info

Publication number: CN108259336B
Application number: CN201711172805.4A
Authority: CN
Inventors: 杨正广; 陈军
Original assignee: Hangzhou H3C Technologies Co Ltd
Current assignee: Hangzhou H3C Technologies Co Ltd
Priority date: 2017-11-22
Filing date: 2017-11-22
Publication date: 2020-12-29
Anticipated expiration: 2037-11-22
Also published as: CN108259336A

Abstract

The application provides a data center interconnection method. In this application, the protection group addresses configured by different EDs in the same data center are the same, which is equivalent to forming a virtual protection group, and the source end of a tunnel established between two different data centers is the protection group address configured by the ED in the local data center (which is equivalent to the local protection group), and the destination end is the protection group address configured by the ED in the opposite data center (which is equivalent to the opposite protection group), which means that the tunnel between two different data centers is established as a whole with the protection group, so for the ED of any data center, when subsequently forwarding traffic to the opposite data center through the tunnel, traffic is shared between the EDs in the protection group of the opposite data center (which corresponds to the protection group address configured by the ED in the opposite data center), and finally, load sharing between different EDs in the same data center is achieved.

Description

Data center interconnection method and device

Technical Field

The present application relates to network communication technologies, and in particular, to a data center interconnection method and apparatus.

Background

Data Center Interconnection (DCI) is realized by Data centers deployed at different positions through an Ethernet Virtual Private Network (EVPN), and the Data Center Interconnection can improve service robustness, reduce Network delay, reduce Network jitter and the like.

EVPN is a two-layer Virtual Private Network (VPN) technology, in EVPN, a control plane adopts a multi-Protocol Border Gateway Protocol (MP) to announce EVPN routing information, and a data plane adopts an eXtensible Virtual local area Network (VXLAN) encapsulation mode to forward messages. The EVPN provides interconnection for data centers in different locations when the different data centers are dispersed in different locations.

Disclosure of Invention

The application provides a method and a device for realizing load sharing among a plurality of different edge devices of the same data center in data center interconnection application.

The technical scheme provided by the application comprises the following steps:

a data center interconnection method is applied to an edge device ED of a data center and comprises the following steps:

configuring a protection group address; the protection group address is the same as the protection group addresses of other EDs in the same data center;

advertising an EVPN route with the next hop set as the protection group address;

receiving an EVPN route announced by an opposite end ED in an opposite end data center; wherein, the next hop of the EVPN route advertised by the opposite end ED is a protection group address configured for the opposite end ED;

and adding the received EVPN route to a local EVPN route table, checking whether a tunnel with a source address configured for an ED (enhanced data access) of the local end and a destination address configured for the ED of the opposite end is established between the local end data center and the opposite end data center, and if so, setting an outlet interface of the EVPN route added to the local EVPN route table as the tunnel.

An interconnection apparatus of a data center, which is applied to an edge device ED of the data center, includes:

a configuration unit, configured to configure a protection group address at a home end ED; the protection group address is the same as the protection group addresses of other EDs in the same data center;

a route notification unit for notifying an EVPN route with the next hop set as the protection group address;

the route receiving unit is used for receiving an EVPN route announced by an opposite end ED in an opposite end data center; wherein, the next hop of the EVPN route advertised by the opposite end ED is a protection group address configured for the opposite end ED;

and the route processing unit is used for adding the EVPN route received by the route receiving unit into a local EVPN route table, checking whether a tunnel with a source address being a protection group address of the local end ED and a destination address being a protection group address configured by the opposite end ED is established between the local end data center and the opposite end data center, and if so, setting an output interface of the EVPN route added into the local EVPN route table as the tunnel.

It can be seen from the above technical solutions that, in the present application, the addresses of the protection groups configured by different EDs in the same data center are the same, which is equivalent to forming a virtual protection group, the source end of the tunnel established between two different data centers is the protection group address configured by the ED in the local data center (equivalent to the local protection group), the destination end is the protection group address configured by the ED in the opposite data center (equivalent to the opposite protection group), this means that the tunnel between two different data centers is established with the protection group as a whole, and as such, for an ED of any data center, when forwarding traffic to an opposite-end data center through a tunnel subsequently, traffic is shared among all EDs in a protection group (corresponding to a protection group address configured for the ED in the opposite-end data center) of the opposite-end data center, and finally, load sharing among different EDs in the same data center is realized.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and together with the description, serve to explain the principles of the disclosure.

FIG. 1 is a schematic diagram of a data center networking;

FIG. 2 is a schematic diagram of a data center networking configured with protected group addresses provided herein;

FIG. 3 is a flow chart of a method provided herein;

FIG. 4 is a schematic diagram of application networking of an embodiment provided by the present application;

FIG. 5 is a block diagram of the apparatus provided in the present application;

fig. 6 is a schematic hardware structure diagram of the apparatus shown in fig. 5 provided in the present application.

Detailed Description

FIG. 1 illustrates a typical networking of data center interconnects. The networking shown in fig. 1 is exemplified by interconnecting DC1 and DC 2. In FIG. 1, Edge devices (ED: Edge Device)110, ED111 are the outlets of DC 1. ED112, ED113 are the outlets of DC 2.

In DC1 or DC2 shown in fig. 1, whether an ED or Leaf device (Leaf), it configures a unique device identifier, which may be a loopback (loopback) address. For example, in DC1 shown in fig. 1, the only device identifier configured by ED110 is hookback 1a, the only device identifier configured by ED111 is hookback 2a, the only device identifier configured by Leaf120 is hookback 3a, the only device identifier configured by Leaf121 is hookback 4a, in DC2, the only device identifier configured by ED112 is hookback 5a, the only device identifier configured by ED113 is hookback 6a, the only device identifier configured by Leaf122 is hookback 7a, and the only device identifier configured by Leaf123 is hookback 8 a.

When a plurality of EDs exist in the same data center, the network reliability can be effectively improved by carrying out load sharing among the plurality of EDs. However, EVPN is a two-layer virtual private network that currently has no perfect load sharing scheme.

In order to realize load sharing among multiple EDs in the data center, the routing distribution of the EVPN is improved, and through the improvement, the load sharing among the multiple EDs in the data center can be realized. Before describing the method provided by the present application, a description will be given of a data center to which the present application relates:

the data center is divided into two parts of a base layer (underwritay) network and an Overlay network. The Overlay network and the Underlay network are independent, and the Overlay network is formed by superposing (Overlay) a layer of logic network on the Underlay network.

The ED or Leaf in the data center generates an Underlay route with reachable unique device identifier (for example, a hookback address) for indicating its configuration according to the Underlay routing protocol, and advertises the Underlay route to the opposite end, and the next hop of the Underlay route is an interface address of a physical port advertising the Underlay route. The Underlay routing protocol herein includes static routing protocols (such as preconfigured routing protocols) and also includes dynamic routing protocols (such as IS-IS protocols, OSPF protocols, etc. commonly used protocols).

And when the opposite terminal receives the Underlay route, adding the received Underlay route to the local Underlay route table. And the opposite terminal establishes an EVPN neighbor with the received Lookback address indicated by the Underlay route so as to issue the EVPN route in the Overlay network through the EVPN neighbor.

Taking the ED110 in the DC1 shown in fig. 1 as an example, the ED110 generates an Underlay route indicating that the unique device identifier configured by the ED110, i.e., the hookback 1a, is reachable according to the Underlay routing protocol, and advertises the Leaf120, the Leaf121, and the ED112 and the ED113 in the DC2 connected in the same DC 1. Taking the example that the ED110 advertises to the ED112 the Underlay route indicating that the loopback address Lookback1 configured by the ED110 is reachable, where the next hop of the Underlay route advertised by the ED110 is the interface address of the physical port interconnected with the ED112 on the ED110, when the ED112 receives the Underlay route advertised by the ED110, the ED112 adds the received Underlay route to the local Underlay route table, and the next hop of the Underlay route added to the local Underlay route table is the interface address of the physical port interconnected with the ED112 on the ED 110. The ED112 also establishes EVPN neighbors with the hookback 1a to publish EVPN routes in the Overlay network through the EVPN neighbors.

Currently, the EVPN technology defines 5 EVPN routes, and there are three routes affecting forwarding, which are:

MAC/IP Advertisement Route: MAC/IP release route, 2 type route for short, used for announcing MAC address and host routing information;

inclusive Multicast Ethernet Tag Route: the inclusive multicast Ethernet label routing is also called IMET routing, 3 types of routing for short, and is used for announcing the VTEP and the VXLAN information of the VTEP so as to realize automatic discovery of the VTEP, automatic establishment of a VXLAN tunnel and automatic association of the VXLAN and the VXLAN tunnel;

IP Prefix advertisement route: the IP prefix routing, for short, the routing of 5 types, is used for announcing BGP IPv4 unicast routing or BGP IPv6 unicast routing in the form of an IP prefix;

in existing solutions, whether ED or Leaf, the next hop for the advertised EVPN route is to use a configured hookback address.

In the present application, to achieve load sharing among multiple EDs in the same data center, additional configuration is made for each ED in each data center, specifically see step 301 below.

The methods provided herein are described below:

referring to fig. 3, fig. 3 is a flow chart of a method provided by the present application. The flow applies to the ED of the data center.

Step 301, configuring a protection group address; the protection group address is the same as the protection group address of other EDs of the same data center.

In the present application, to achieve load sharing among multiple EDs in the same data center, the following configuration is made for each ED in each data center: a protected group address is configured.

Different EDs in the same data center are configured with the same protection group address, and the ED in different data centers are configured with different protection group addresses. Still in the networking shown in fig. 1, ED110 and ED111 in DC1 configure the same protection group address (denoted as protection group address 1), and ED112 and ED113 in DC2 configure the same protection group address (denoted as protection group address 2). Where protected group address 1 is different from protected group address 2.

In the present application, configuring the same protection group address on each ED in the same data center is equivalent to forming a virtual protection group between different EDs in the same data center, where an ED is a member of the protection group, and the address of the protection group is the protection group address configured on the ED. Virtual protection groups formed by different EDs in the same data center work as a whole. FIG. 2 is a schematic diagram of a data center networking configured with protected group addresses based on the networking shown in FIG. 1

In this application, the address of the protection group configured by the ED in the data center is different from the address of the lookup back configured by the ED itself, and is a virtual address relative to the address of the lookup back configured by the ED itself, and is not used for identifying the ED.

Step 302, advertising an EVPN route with the next hop set as the protected group address.

As described above, in the current EVPN, when the ED advertises an EVPN route, the next hop of the advertised EVPN route is a loopback address Lookback configured by the ED itself. In the application, the scheme that the ED advertises the EVPN route is improved, and for the EVPN route advertised by the ED, the next hop of the EVPN route is not the loopback address Lookback configured by the ED but the protection group address.

And when the ED or Leaf in the same data center with the local end ED receives the EVPN route, establishing a tunnel with the next hop of the EVPN route.

When receiving the EVPN route, the opposite end ED, which is located in a different data center from the home end ED, performs the operations performed by the home end ED in step 303.

Similarly, the opposite end ED will also issue the EVPN route with the next hop being the configured protection group address for the opposite end ED according to the description of step 302, and after receiving the EVPN route advertised by the opposite end ED, the local end ED executes step 303 below.

Step 303, the home end ED receives an EVPN route advertised by the opposite end ED in the opposite end data center, adds the received EVPN route to the local EVPN route table, and checks whether a tunnel having a source address configured for the home end ED and a destination address configured for the opposite end ED has been established between the home end data center and the opposite end data center, if yes, sets an outgoing interface of the EVPN route added to the local EVPN route table as the tunnel; wherein, the next hop of the EVPN route advertised by the opposite end ED is the protection group address configured for the opposite end ED.

As an embodiment, if it is checked that the tunnel is not established between the local data center and the peer data center, the tunnel is established between the local data center and the peer data center, and an egress interface of the EVPN route added to the local EVPN routing table is set as the tunnel.

As described above, the address of the protection group in each different ED configuration of the same data center is the same, which is equivalent to forming a virtual protection group, in combination with the description in step 303, it can be known that the source end of the tunnel between two different data centers is the protection group address (corresponding to the home protection group) configured for the ED in the home data center, the destination end is the protection group address (corresponding to the peer protection group) configured for the ED in the peer data center, this means that the tunnel between two different data centers is established with the protection group as a whole, and as such, for an ED of any data center, when forwarding traffic to an opposite-end data center through a tunnel subsequently, traffic is shared among all EDs in a protection group (corresponding to a protection group address configured for the ED in the opposite-end data center) of the opposite-end data center, and finally, load sharing among different EDs in the same data center is realized.

The flow shown in fig. 3 is completed.

In this application, an EVPN route advertised by the local ED may be an EVPN route locally generated by the local ED, or may be an EVPN route (for short, a non-local EVPN route) advertised by an ED in another opposite-end data center received by the local ED, and how to advertise an EVPN route whose next hop is set as a protection group address in step 302 is described below by subdividing the local EVPN route and the non-local EVPN route:

as an embodiment, the EVPN route advertising the next hop set to the protected group address in step 302 includes the following steps b1 through b 3:

step b1, if the EVPN route to be advertised is the inclusive multicast ethernet label route (local type 3 route) defined by the EVPN generated by the home end ED, setting the next hop of the inclusive multicast ethernet label route (local type 3 route) as the protection group address configured by the home end ED and advertising the protection group address.

In order to ensure that the next hop of the EVPN-defined inclusive multicast ethernet label route (local type 3 route) generated by the home ED is set as the protection group address configured by the home ED, the following configuration needs to be made in advance: the local end ED configures VXLAN ID and routing attribute (RT) related to a two-layer Virtual Switch Instance (VSI), and configures VXLAN ID and MAC address of a three-layer VXLAN; the VXLAN ID and RT related to the two-layer VSI are respectively the same as the VXLAN ID and RT related to the VSI configured by other EDs in the same data center; the VXLAN ID and MAC address of the three-tier VXLAN are the same as the VXLAN ID and MAC address of the three-tier VXLAN disposed by other EDs in the same data center.

Through the configuration, it is finally ensured that the inclusive multicast ethernet label routes (local type 3 routes) generated by the EDs in the same data center are all the same, and based on this, when the inclusive multicast ethernet label routes (local type 3 routes) defined by the EVPN generated by the home end ED, the next hop of the generated inclusive multicast ethernet label routes (local type 3 routes) can be set as the configured protection group address of the home end ED. Based on the fact that the protection group addresses configured by the different EDs in the same data center are the same, which is equivalent to forming a virtual protection group, the home end ED sets the next hop of the generated inclusive multicast ethernet label route (local type 3 route) as the protection group address configured by the home end ED, which means that each ED in the home end data center externally operates as a whole with the protection group, and a basis is provided for realizing load sharing between different EDs in the same data center.

Similarly, VXLAN ID and RT associated with the two-layer VSI configured by each ED in the peer data center are the same, and VXLAN ID and MAC address of the three-layer VXLAN are the same.

Step b2, if the EVPN route to be advertised is the MAC/IP distribution route (local type 2 route) or the IP prefix route (local type 5 route) defined by the EVPN generated by the home end ED, when it is checked that the MAC/IP distribution route or the IP prefix route satisfies the preset routing policy, setting the next hop of the MAC/IP distribution route (local type 2 route) or the IP prefix route (local type 5 route) satisfying the preset routing policy as the configured protection group address of the home end ED and advertising.

Step b3, if the EVPN route to be advertised is the MAC/IP distribution route (non-local type 2 route) or the IP prefix route (non-local type 5 route) defined by the EVPN received by the home end ED, setting the next hop of the received MAC/IP distribution route (non-local type 2 route) or IP prefix route (non-local type 5 route) as the configured protection group address of the home end ED and advertising the next hop.

As an embodiment, in step b2 described above, the home end ED generally generates the MAC/IP publish route (native type 2 route) and the IP prefix route (native type 5 route) when acting as a gateway.

When the local end ED generates the MAC/IP distribution route (local type 2 route), instead of collectively setting the next hop of the generated MAC/IP distribution route (local type 2 route) as the configured protection group address of the local end ED, it is determined whether the generated MAC/IP distribution route (local type 2 route) satisfies the preset routing policy, and only when the preset routing policy is satisfied, the next hop of the generated MAC/IP distribution route (local type 2 route) is set as the configured protection group address of the local end ED.

Similarly, when the home-end ED generates an IP prefix route (local 5-class route), instead of collectively setting the next hop of the generated IP prefix route (local 5-class route) as the protection group address configured by the home-end ED, it is determined whether the generated IP prefix route (local 5-class route) satisfies the preset routing policy, and only when the preset routing policy is satisfied, the next hop of the generated IP prefix route (local 5-class route) is set as the protection group address configured by the home-end ED.

In one example, the preset routing policy includes a specified network segment for instructing a route matching the specified network segment to perform the following operations: the next hop is modified to the protection group address already configured by the home ED. The designated network segment included in the preset routing strategy may be one network segment or more than two different network segments.

Based on the above description of the preset routing policy, in the present application, the determining, by the local end ED, whether the generated MAC/IP distribution route (local type 2 route) or IP prefix route (local type 5 route) meets the preset routing policy may include: and the local end ED checks whether the generated MAC/IP release route or the network segment to which the IP prefix route belongs is matched with the specified network segment, and if so, determines that the MAC/IP release route or the IP prefix route meets a preset routing strategy. For example, the description is as follows:

if the preset routing policy includes more than two specified network segments, the home end ED checks whether the network segment to which the generated MAC/IP distribution route (local 2-type route) belongs is included in one specified network segment included in the routing policy for the generated MAC/IP distribution route (local 2-type route), and if so, determines that the generated MAC/IP distribution route (local 2-type route) satisfies the preset routing policy; similarly, for the generated IP prefix route (local 5-type route), checking whether the network segment to which the generated IP prefix route (local 5-type route) belongs is included in a specified network segment included in the routing policy, and if so, determining that the generated IP prefix route (local 5-type route) satisfies the preset routing policy.

Through the routing strategy, only the MAC/IP release route (local type 2 route) and the IP prefix route (local type 5 route) which meet the routing strategy and are locally generated by the local end ED can be set as the next hop to be the protection group address configured by the local end ED, and the flexibility is good.

And when the local end ED judges that the MAC/IP distribution route (local type 2 route) and the IP prefix route (local type 5 route) which do not satisfy the routing policy, executing according to the existing EVPN routing scheme, that is, setting the next hop of the MAC/IP distribution route (local type 2 route) which does not satisfy the routing policy as the unique identifier configured by the local end ED, such as loopback address hookback, and setting the next hop of the IP prefix route (local type 5 route) which does not satisfy the routing policy as the unique identifier configured by the local end ED, such as loopback address hookback.

Based on the above description, how to implement load sharing between EDs in the same data center is described below based on one embodiment:

fig. 4 shows an application networking diagram of an embodiment provided in the present application. In the networking shown in fig. 4, DC401 and DC402 are interconnected as an example. In fig. 4, ED411, ED412 are the outlets of DC 401. The ED413, ED414 are the outlets of the DC 402.

In the DC401 or DC402 shown in fig. 4, whether ED or Leaf, it configures a unique device id, which may be a hookback address. For example, as shown in fig. 4, the only device identifier configured by the ED411 is hookback 1c, the only device identifier configured by the ED412 is hookback 2c, the only device identifier configured by the ED413 is hookback 3c, and the only device identifier configured by the ED414 is hookback 4 c. The relationship between Leaf and the application is not so great, and the description is not repeated.

In this application, the EDs 411 and 412 in the DC401 may also be configured with the same protection Group address (denoted as Group1d) to form one protection Group (denoted as Group1), and the EDs 413 and 414 in the DC402 may also be configured with the same protection Group address (denoted as Group2d) to form one protection Group (denoted as Group2), where Group1d is different from Group2 d.

In this application, ED411, ED412 in DC401 have the same VXLAN ID (denoted as VXLAN100), RT (denoted as RT1) associated with the two-tier VSI and the same VXLAN ID (denoted as VXLAN200), MAC address (denoted as MAC1) of the three-tier VXLAN.

In this application, ED413, ED414 in DC402 have the same VXLAN ID (denoted as VXLAN300), RT (denoted as RT2, different from RT1) associated with the two-tier VSI, and the same VXLAN ID (denoted as VXLAN400), MAC address (denoted as MAC2) of the three-tier VXLAN.

The route advertisement is described by taking ED411 as an example:

according to the Underlay routing protocol, in the present embodiment, the ED411 generates an Underlay route indicating that the unique device identifier configured by the ED110 is reachable by the hookback 1c, and advertises each Leaf connected in the same DC1 and the ED413 and the ED414 in the DC 402. Take the example that the ED411 advertises to the ED413 an Underlay route that indicates that the unique device identifier (for example, loopback address Lookback1 c) configured by the ED411 is reachable, where the next hop of the Underlay route advertised by the ED411 is an interface address of a physical port interconnected with the ED413 on the ED411, and when the ED413 receives the Underlay route advertised by the ED411, the ED413 adds the received Underlay route to the local Underlay route table, and the next hop of the Underlay route added to the local Underlay route table is an interface address of a physical port interconnected with the ED413 on the ED 411. The ED413 may also establish an EVPN neighbor with the loopback address hookback 1c indicated by the Underlay route to publish EVPN routes in the Overlay network through the EVPN neighbor.

According to the Underlay routing protocol, in this embodiment, the ED411 generates an Underlay route indicating that the protection Group address Group1d configured by the ED110 is reachable, and advertises each Leaf connected in the same DC1 and the EDs 413 and 414 in the DC 402. Taking the example that the ED411 advertises to the ED413 an Underlay route indicating that the protection Group address Group1d configured by the ED411 is reachable, wherein a next hop of the Underlay route advertised by the ED411 is an interface address of a physical port interconnected with the ED413 on the ED411, when the ED413 receives the Underlay route advertised by the ED411, checking whether the local Underlay routing table has an Underlay route indicating that the protection Group address Group1d is reachable, if not, the ED413 adding the received Underlay route to the local Underlay routing table, and a next hop of the Underlay route added to the local Underlay routing table is an interface address of a physical port interconnected with the ED413 on the ED 411; if yes, the next hop of the received Underlay route (the interface address of the physical port interconnected with the ED413 on the ED411 mentioned above) is newly added to the Underlay routes already in the local Underlay routing table to indicate that the protected Group address Group1d is reachable. The ED412 will also advertise the Underlay route in such a way that the ED411 advertises the Underlay route by indicating the reachable protection Group address Group1d, and finally, implemented in the Underlay route table local to the ED413, the Group1d is finally associated with two next hops, one is the interface address of the physical port interconnected with the ED413 on the ED411, and the other is the interface address of the physical port interconnected with the ED413 on the ED 412. The two next hops with which Group1d is ultimately associated are shown as follows:

next hop 1 e: interface addresses for physical ports on the ED411 that interconnect with the ED 413;

next hop 2 e: is the interface address of the physical port on the ED412 that is interconnected with the ED 413.

These two next hops are equivalent next hops.

According to the EVPN routing protocol, when the ED411 advertises the EVPN route, the next hop of the EVPN route is set as the protection Group address Group1d configured by the local ED411, and advertises the modified EVPN route. Take the example that the ED411 advertises the EVPN route to the ED413, where the EVPN route advertised by the ED411 is a type 2 route locally generated by the ED411, and the next hop is Group1 d. For ease of description, the EVPN route advertised by ED411 to ED413 is referred to herein as EVPN route 500.

When ED413 receives EVPN route 500 advertised by ED411, EVPN route 500 is added to the local EVPN routing table, and the next hop of EVPN route 500 added to the local EVPN routing table is Group1 d.

The ED413 checks whether a tunnel with a protection Group address, i.e. Group2d, configured by the ED413 as a source address and a destination address, i.e. Group1d, has been established between the local data center, i.e. DC402, and the opposite data center (the data center where the ED411 is located), i.e. DC 401.

The ED413 checks that a tunnel with a source address of Group2d and a destination address of Group1d is not established between the DC402 and the DC401, establishes a tunnel with a source address of Group2d and a destination address of Group1d between the DC402 and the DC401, and sets an outgoing interface of the EVPN route 500 added to the local EVPN route table as the established tunnel.

Similarly, when the ED412 advertises the EVPN route, the next hop of the EVPN route is set as the protection Group address Group1d configured by the local ED412, and advertises the modified EVPN route. Take the example that the ED412 advertises the EVPN route to the ED413, where the EVPN route advertised by the ED412 is a type 2 route locally generated by the ED412, and the next hop is Group1 d. For ease of description, the EVPN route advertised by ED412 to ED413 is referred to herein as EVPN route 600.

When the ED413 receives the EVPN route 600 advertised by the ED412, the EVPN route 600 is added to the local EVPN route table, the next hop of the EVPN route 600 added to the local EVPN route table is Group1d, and whether a tunnel with a source address of Group2d and a destination address of Group1d already established between the local data center (DC 402) and the opposite data center (DC 401), which is the data center where the ED412 is located, is detected.

The ED413 checks that a tunnel with a source address of Group2d and a destination address of Group1d is established between the DC402 and the DC401, and directly sets the outgoing interface of the EVPN route 600 added to the local EVPN routing table as the established tunnel without repeatedly establishing the same tunnel.

The manner in which the ED413, ED414 advertise routes is similar to the ED411, ED412 and will not be described in detail.

Thereafter, DC402 communicates with DC401, taking as an example the ED413 in DC402 forwards the data stream to DC 401:

when the ED413 forwards the data stream, the interface is determined to be a tunnel and the next hop is Group1d according to the local EVPN routing table.

The ED413 finds that Group1d associates two interface addresses according to the local Underlay routing table: an interface address for the physical port on the ED411 that interconnects with the ED 413; the other is the interface address of the physical port on the ED412 that is interconnected with the ED413, then the ED413 selects one of the two interface addresses and sends the data stream to the selected interface address through the tunnel as the outgoing interface. Here, there are many ways for the ED413 to select one of the two interface addresses, for example, according to a load sharing algorithm. The final ED413 selects one of the two interface addresses to enable load sharing between the two EDs in DC 401.

Thus, the description of the embodiments is completed.

The methods provided herein are described above. The following describes the apparatus provided in the present application:

referring to fig. 5, fig. 5 is a diagram illustrating a structure of the apparatus according to the present invention. The device is applied to the ED of a data center, wherein different EDs in the same data center are configured with the same protection group address.

As shown in fig. 5, the apparatus may include:

As an embodiment, the route processing unit further establishes the tunnel between the local data center and the peer data center when it is checked that the tunnel is not established between the local data center and the peer data center, and sets an egress interface of the EVPN route added to the local EVPN routing table as the tunnel.

As an embodiment, the route advertisement unit is specifically configured to:

if the EVPN route to be notified is an inclusive multicast Ethernet label route defined by the EVPN generated by the home end ED, setting the next hop of the inclusive multicast Ethernet label route as a protection group address configured by the home end ED and notifying;

if the EVPN route to be advertised is the MAC/IP release route or the IP prefix route defined by the EVPN generated by the local end ED, when the generated MAC/IP release route or the IP prefix route is detected to meet the preset routing strategy, setting the next hop of the MAC/IP release route or the IP prefix route meeting the preset routing strategy as the configured protection group address of the local end ED and advertising the next hop;

if the EVPN route to be advertised is the MAC/IP distribution route or the IP prefix route defined by the EVPN received by the local ED, setting the next hop of the received MAC/IP distribution route or IP prefix route as the protection group address configured by the local ED and advertising the protection group address.

As an embodiment, the preset routing policy includes a specified network segment, and is used to instruct EVPN routing matching the specified network segment to perform the following operations: setting the next hop as the configured protection group address of the home end ED;

the step of the route notification unit checking that the generated MAC/IP release route or IP prefix route meets the preset routing strategy comprises the following steps:

and checking whether the generated MAC/IP release route or the network segment to which the IP prefix route belongs is matched with the specified network segment, and if so, determining that the generated MAC/IP release route or the IP prefix route meets a preset routing strategy.

As an embodiment, the configuration unit further configures VXLAN ID and routing attribute RT related to the layer two virtual switch instance VSI, and configures VXLAN ID and MAC address of the layer three VXLAN;

the VXLAN ID and RT related to the VSI are respectively the same as the VXLAN ID and RT related to the VSI configured by other EDs in the same data center;

the VXLAN ID and MAC address of the three-layer VXLAN are the same as the VXLAN ID and MAC address of the three-layer VXLAN configured by other EDs in the same data center.

Thus, the description of the structure of the apparatus shown in fig. 5 is completed.

Correspondingly, the application also provides a hardware structure diagram of the device shown in FIG. 5. As shown in fig. 6, the hardware structure includes:

may include a processor 601, a machine-readable storage medium 602 having machine-executable instructions stored thereon. The processor 601 and the machine-readable storage medium 602 may communicate via a system bus 603. Also, the processor 601 may perform the data center interconnect method described above by reading and executing machine-executable instructions in the machine-readable storage medium 602 corresponding to the data center interconnect logic.

The machine-readable storage medium 602 referred to herein may be any electronic, magnetic, optical, or other physical storage device that can contain or store information such as executable instructions, data, and the like. For example, the machine-readable storage medium may be: random Access Memory (RAM), volatile Memory, non-volatile Memory, flash Memory, a storage drive (e.g., a hard drive), a solid state drive, any type of storage disk (e.g., an optical disk, dvd, etc.), or similar storage media, or a combination thereof.

Up to this point, the description of the hardware configuration shown in fig. 6 is completed.

Also provided in this application is a machine-readable storage medium, such as machine-readable storage medium 602 in fig. 6, comprising machine-executable instructions that are executable by processor 601 in a data center interconnection apparatus to implement the data center interconnection method described above.

In particular, the processor 601 may perform the operations in the above data center interconnection method by invoking and executing machine executable instructions in a machine readable storage medium corresponding to data center interconnection logic.

The above description is only exemplary of the present application and should not be taken as limiting the present application, as any modification, equivalent replacement, or improvement made within the spirit and principle of the present application should be included in the scope of protection of the present application.

Claims

1. A data center interconnection method is applied to an edge device ED of a data center and comprises the following steps:

2. The method of claim 1, further comprising:

and if the tunnel is not established between the local data center and the opposite data center, establishing the tunnel between the local data center and the opposite data center, and setting an outgoing interface of the EVPN route added to the local EVPN route table as the tunnel.

3. The method of claim 1, wherein advertising the EVPN route with the next hop set to the protected group address comprises:

4. The method of claim 3, wherein the preset routing policy comprises a specified network segment for instructing an EVPN route matching the specified network segment to perform the following operations: setting the next hop as the configured protection group address of the home end ED;

the step of checking that the generated MAC/IP release route or the IP prefix route meets the preset routing strategy comprises the following steps:

5. The method according to any one of claims 1 to 4,

configuring VXLAN ID and routing attribute RT related to two-layer virtual switch instance VSI, wherein the VXLAN ID and RT related to the two-layer VSI are respectively the same as VXLAN ID and RT related to VSI configured by other EDs in the same data center;

and configuring VXLAN ID and MAC address of the three-layer VXLAN, wherein the VXLAN ID and MAC address of the three-layer VXLAN are the same as the VXLAN ID and MAC address of the three-layer VXLAN configured by other EDs in the same data center.

6. A data center interconnection device is applied to an edge device ED of a data center, and comprises:

7. The apparatus of claim 6, wherein the route processing unit further establishes the tunnel between the local data center and the peer data center after checking that the tunnel is not established between the local data center and the peer data center, and sets an egress interface of the EVPN route added to the local EVPN routing table as the tunnel.

8. The apparatus of claim 6, wherein the route advertisement unit is specifically configured to:

9. The apparatus of claim 8, wherein the predetermined routing policy comprises a specified network segment, and wherein the EVPN route matching the specified network segment is instructed to: setting the next hop as the configured protection group address of the home end ED;

10. The apparatus according to any one of claims 6 to 9,

the configuration unit further configures VXLAN ID and routing attribute RT related to the VSI of the two-layer virtual switching instance, and configures VXLAN ID and MAC address of the three-layer VXLAN;