The statement that BGP neighbors automatically discover each other is factually incorrect. BGP is a path-vector protocol that requires neighbors (peers) to be manually and explicitly configured with each other's IP addresses and autonomous system numbers. Unlike link-state protocols such as OSPF or IS-IS, which use multicast Hello packets on a shared link to dynamically discover neighbors, BGP has no such native mechanism. Since automatic neighbor discovery is not a feature of BGP, it cannot be a reason for its use in the data center. The other options listed are all well-established reasons for BGP's adoption in modern data center fabrics.

A. This is a valid reason. In large-scale data center fabrics (e.g., leaf-spine), BGP is considered more efficient and scalable than link-state protocols because it avoids the overhead of flooding link-state databases across the entire fabric.

C. This is a valid reason. BGP has been the core routing protocol of the global internet for decades. Its maturity, stability, and the widespread operational expertise surrounding it make it a trusted choice for critical infrastructure.

D. This is a valid reason. Multi-Protocol BGP (MP-BGP) extensions allow BGP to carry routing information for various address families (AFI/SAFI), including IPv4, IPv6, and EVPN, making it highly versatile for modern, multi-service data centers.

1. RFC 7938: Use of BGP for Routing in Large-Scale Data Centers.

Section 3, "Why BGP?": This section details the advantages of using BGP in the data center, listing scalability, faster reconvergence (in specific DC designs), support for multiple address families (IPv4, IPv6, L2VPN, L3VPN), and its mature, widely implemented nature. It makes no mention of automatic discovery, as this is not a BGP feature. The configuration of BGP peers is always explicit.

2. Nokia Data Center Fabric Solution Guide.

Section "Underlay and Overlay Routing Protocols": This section typically explains the choice of eBGP for the fabric. It highlights BGP's scalability to handle a large number of leaf and spine nodes, its robust policy control, and its multi-protocol capabilities for services like EVPN. The configuration examples provided in such guides always demonstrate the manual configuration of BGP neighbors, reinforcing that auto-discovery is not a feature.

3. Kurose, J. F., & Ross, K. W. (2021). Computer Networking: A Top-Down Approach (8th ed.). Pearson.

Chapter 5.3, "BGP - The Internet's Inter-AS Routing Protocol": University-level textbooks on networking describe the fundamental operation of BGP. They explicitly state that BGP sessions are established over TCP between manually configured peers. This contrasts with their description of OSPF/IS-IS, where neighbor discovery via Hello packets is a key feature.