
Start with the purpose of the addresses
Renting IPv4 subnets allows planned use of an address block under the project’s and agreement’s conditions. Before choosing a prefix, describe why operations need addresses: identifying environments, separating customers, publishing services, or organizing application egress. Each purpose creates different routing, control, and monitoring requirements.
The number of robots or processes does not automatically determine IP count. A group of tasks may share an egress address, while a single environment may require several addresses for previously defined architecture and continuity reasons. Document the relationship rather than assume one address per worker.
Create an inventory of current use, expected growth, and justified reserves. Separate addresses for production, testing, infrastructure, and contingency. Avoid reserving large quantities without a usage plan, but do not size solely from consumption on a quiet day. Consider peak periods and the time needed to expand.
The aim is a block that accommodates demonstrated needs and remains understandable to administer. This analysis makes proposals easier to compare because it shows whether a size difference serves a real requirement or merely adds resources with no defined purpose.
Convert the prefix into a total count
IPv4 uses 32-bit addresses. For a /p prefix, the total address count is two raised to the power of 32 minus p. A /24 therefore contains 256 addresses and a /16 contains 65,536. VLSM.io’s IPv4 subnet offering considers the range from /24 through /16, subject to availability and project requirements. A smaller number after the slash means a larger block. This calculation describes the mathematical set size; it does not independently determine how many addresses will be delivered for a particular use or network design.
Do not automatically subtract two addresses for every service mode. In a conventional IPv4 subnet, network and broadcast affect the number used by hosts, but other contexts exist, including /31 point-to-point links and individual /32 routes.
A pool delivered through a proxy is not necessarily equivalent to a LAN configured across every address in the block. Request the actual delivered count, technical reservations, and association method for your application. The proposal should distinguish prefix total from usable capacity under the contracted design, rather than hide those differences in one number.
| Prefix | Total addresses |
|---|---|
| /24 | 256 |
| /22 | 1,024 |
| /20 | 4,096 |
| /16 | 65,536 |
Choose delivery according to existing infrastructure
Delivery through HTTP or SOCKS IP Proxy serves applications that can configure an intermediary. VXLAN and GRE participate in connectivity designs and require compatible network configuration. cross-connect involves physical interconnection in an appropriate environment. A VM or bare-metal server provides an execution location and computing resources associated with the project.
These options are not synonyms and cannot be selected by name alone. Identify where the application is hosted, which network permissions you have, and whether your team needs to control routes or only configure access through a proxy interface.
When choosing tunnels, confirm endpoints, return paths, packet sizing, and configuration ownership. Encapsulation adds information to traffic and must be considered in path behavior. Do not assume GRE or VXLAN alone provides encryption. If confidentiality between endpoints is required, include an appropriate mechanism in the design.
For proxy delivery, confirm actually supported protocols, authentication, and UDP handling where needed. The SOCKS5H label found in clients represents a proxy-side name-resolution option, not an additional transport layer that replaces the other design decisions. Test the chosen mode with the real application.
203.0.113.24198.51.100.20203.0.113.24Addresses and bandwidth are different dimensions
A larger block does not automatically increase transmission capacity or a service’s connection limit. Size traffic in bits per second, packets per second, simultaneous connections, and new connections per unit of time separately. Also consider application CPU, memory, and processing.
VLSM.io describes scalability from 1 Mbps through multiple 100 Gbps according to the project; that range does not mean every address or subscription has the entire capacity available. The proposal should specify what is contracted, how it is measured, and which components may constrain performance during representative workloads.
Compare the usage profile with the proposed architecture. Many small lookups may require many connections, while a few large downloads may consume more bandwidth. A monthly average does not reveal short bursts or concentrated task periods.
Record response-size distribution, duration, and concurrency to build a useful estimate, then validate it with controlled, authorized load. Address count remains important, but it does not replace this analysis. A balanced project avoids purchasing many IPs while the real bottleneck remains processing, latency, or coordination of work across queues and destinations.
Reserve operational capacity to manage the block as well. As the set grows, a reliable inventory of assignments, owners, and changes becomes more important. A spreadsheet without update control may work for a pilot and become inadequate during expansion. Define a source of truth for address assignment and release, with change history. The goal is not bureaucracy but the ability to quickly identify which service used an origin at a given time. That information helps support and reduces errors when environments are created, migrated, or retired.
Define routing and resource responsibilities
An agreement to use a block should not be confused with an ownership or registration transfer at an Internet registry. Clarify who provides the resource, who administers announcements and related records, and which procedures apply when use begins or ends.
Where BGP routing is involved, align requirements with the provider and responsible network team. Do not assume every subdivision of the block will be accepted globally as an independent announcement. Routing and filtering policies belong in the feasibility analysis for the chosen path, including its return traffic.
A Route Origin Authorization, or ROA, relates a prefix to an autonomous system authorized to originate it and may specify the permitted maximum length. It addresses routing-origin validation, not permission to query third-party applications. Likewise, correct registration does not guarantee every destination will accept traffic.
Include route checks, record consistency, and operational contacts in delivery. For application use, separately maintain permissions, limits, and data-handling rules. Each layer needs evidence appropriate to what it actually controls; one valid routing artifact cannot substitute for all the others.
Prepare expansion, acceptance, and exit
Before contracting, consolidate required quantity, delivery mode, protocols, capacity, and deployment schedule. Define a pilot proving access to the delivered addresses, the return path, and application operation. Include component failure and recovery according to the required continuity level.
Keep written criteria for determining when the block is ready for production. Approval should represent an observed result in your environment, not merely delivery of an address list or connection credentials. This creates a shared understanding between your application team, network team, and the service provider.
Plan how to expand and how to end use without leaving forgotten dependencies. Access lists, proxy settings, DNS, and internal records may need updates during migration. Identify owners and deadlines for removing old associations.
When discussing requirements with VLSM.io, present inventory and usage indicators to compare prefixes from /24 through /16 according to actual need. Correct sizing combines address space, delivery, and operations. That combination provides a more predictable foundation for growth than selecting the largest available block and discovering how to administer it only afterward.
Institutional references
Reference documents on the topics covered in this article.
- IETFRFC 4632: Classless Inter-domain Routing (CIDR): The Internet Address Assignment and Aggregation Plandatatracker.ietf.org
- IETFRFC 3021: Using 31-Bit Prefixes on IPv4 Point-to-Point Linksdatatracker.ietf.org
- IETFRFC 7348: Virtual eXtensible Local Area Network (VXLAN)datatracker.ietf.org
- IETFRFC 2784: Generic Routing Encapsulation (GRE)datatracker.ietf.org
- RIPE NCCBGP Origin Validationripe.net
- IANAIPv4, IPv6 and ASN registriesiana.org

