While subnetting divides, Supernetting (summarization) combines prefixes to keep the global BGP table under management.

Variable-Length Subnet Masking (VLSM) enables the hierarchical subdivision of a Classless Inter-Domain Routing (CIDR) block into subnets of different sizes, matching the actual host requirements of each network segment rather than forcing a fixed-size allocation. The VLSM design process starts from the largest subnet requirement and works downward: given a /20 block (4,096 addresses), the data center VLAN requiring 1,024 addresses gets a /22 (1,024 addresses), the management VLAN requiring 512 addresses gets a /23 (512 addresses), and the P2P interconnect links requiring only 2 addresses each get /30s (4 addresses, with 2 usable). The allocation efficiency is defined as η = Σ(2^(32−mask_i) − 2) / 2^(32−parent_mask), and a well-designed VLSM plan achieves η > 90%, compared to approximately 50% for a fixed-length subnet mask (FLSM) design where all subnets use the same prefix length. The IP subnetting tool implements VLSM by accepting the parent prefix and a list of per-subnet host requirements, then applying the largest-subnet-first greedy allocation algorithm: sort subnets by host count in descending order, assign the smallest prefix that satisfies each subnet's requirement, and track the allocation within the parent block using a binary tree representation where each allocated prefix marks its subtree as consumed.

The binary tree (or "prefix tree") representation of VLSM allocation reveals a fundamental constraint: the prefix tree can become fragmented when subnets of different sizes are allocated and deallocated over time, similar to memory fragmentation in a heap allocator. A /20 block subdivided into 16 × /24s is perfectly non-fragmented — any /24 can be independently allocated or freed without creating unusable gaps because all subtrees are the same size. But a VLSM plan with mixed sizes — e.g., one /22, three /23s, ten /26s, and twenty /30s — creates a fragmented free space map where returning a /23 may leave a hole that cannot be used for a new /22 (too small) or a /26 (too large without further subdivision). The fragmentation ratio Φ = (free_addresses − max_contiguous_free) / free_addresses measures the unusable fraction of free space. For Φ > 0.5, more than half the free addresses cannot be allocated to any practical subnet, and the network operator should consider renumbering or adding a secondary parent prefix. The tool includes a fragmentation visualization that shows the block's allocation bitmap and highlights unusable gaps, enabling capacity planning before the fragmentation reaches a critical level where a new VLAN cannot be deployed without renumbering existing subnets.

Hierarchical route aggregation in a VLSM-designed network relies on the property that a shorter prefix (e.g., /21) summarizes all longer prefixes (e.g., /22, /23, /24, /25) that fall within its address range, provided they are contiguous in the parent block. When routes are advertised to BGP, the router performs longest-prefix-match forwarding within the local router but can advertise the aggregated /21 to its EBGP neighbors, reducing the global routing table size by up to a factor of N (where N is the number of constituent subnets). However, the aggregation is only valid if all subnets within the /21 are reachable via the same forwarding path — a condition called "consistent next-hop." If one subnet within the /21 is behind a different router or has a different forwarding policy (e.g., traffic engineering weights, QoS marking, or security ACLs), the /21 aggregate advertisement bleeds traffic to the wrong destination. The tool's route aggregation verifier checks for consistent next-hop by simulating the forwarding table after each allocation: for each allocated subnet, it tracks the assigned next-hop (derived from the VLAN egress interface configuration) and flags any /21 or larger aggregate where the constituent subnets have non-identical next-hops. The Verifier outputs the recommended aggregate prefixes and the "black hole risk" percentage — the fraction of destinations that would be incorrectly forwarded if the flagged aggregates were advertised without the constituent more-specific routes.