Wired

The figure shows basic BGP peer configuration (iBGP and eBGP)
Things to remember in a BGP configuration
Enable BGP with an AS number
Start by configuring the BGP process with its AS number. Remember, the company must obtain this number from a regional authority. You can create only one process and AS on the switch.

Configure a router ID
Although optional, you are strongly encouraged to assign a BGP router-ID. Like OSPF, the loopback 0 address is typical.
If you do not set the router ID manually, the AOS-CX switch uses the highest IP address on a loopback interface, or, if no loopback interfaces exist, the
highest IP address on any interface.

Configure peers
Unlike IGPs like OSPF and RIP, BGP does not discover peers.
You must configure each peer's IP address and AS number. This is how devices know whether they are iBGP or eBGP peers.
In the diagram above, Agg-1 knows that its AS 64500, and you configured peer 10.5.0.3 (Agg-2) with the same AS. These are iBGP peers.
You configured Agg-2 to peer with Router A at 10.255.102.12, AS 64512. Different ASes, so these are eBGP peers.
Configured peers do not communicate until you activate them. You see this with the neighbor activate command.

Use loopbacks for iBGP peers and set iBGP update source
You should configure iBGP sessions to peer with each other's loopback address, to maximize resiliency.
Suppose you configure Agg-1 and Agg-2 to peer via their physical addresses. If that link fails, peering is broken.
If you peer as shown in the figure, via loopback addresses, the routers can maintain the peer relationship via some alternate path.
The reachability between the loopback address is maintained via IGP like OSPF.
BGP peers will always use the IP address of their directly connected link to the other peer by default.
Since you are using loopback for resiliency, you should update loopback as the source address.

eBGP peers use physical IP
You often configure eBGP peers using their physical IP addresses.
This is because eBGP peers often only have a single link between them, with no alternate paths.
Also, unlike iBGP, eBGP peers must be directly connected.

Next-hop-self option
Basically if you have an eBGP route that is being introduced to an iBGP network, you must make sure that the next hop address is reachable in the routing table. If it is not, you would need to do something such as the configuring BGP next-hop-self to make sure that network gets added to the routing table.
The fact that the next-hop IP is not reachable simply excludes the route from the BGP "best path" selection process.
So when the external subnet is not reachable via the IGP(like OSPF), iBGP should be configured with "next-hop self" option.