Celluster™ IP-Less / IPAM-Less

Stop plumbing IPs. Start flowing intent.
Addresses • Identity • Intent
IPs, IPAM and DHCP are plumbing. Intent is the workload.

Cloud today still treats every workload as an IP endpoint: carve subnets, allocate IPs, manage DHCP, wire CNI, debug conntrack. Most of this has nothing to do with the business logic you are trying to run.

Celluster flips the stack: workloads run as Cells identified by intent, zone, and lineage. IPs become an implementation detail of the substrate, not something every team has to juggle.

See How IP-Less Works What This Means for SREs
Cartoon of a frazzled IP plumber juggling CIDR blocks and pipes versus calm intent-driven Cells.
The IP plumber vs intent-driven Cells. Stop patching pipes. Let the substrate carry meaning.

From IP Endpoints → Intent-Native Cells

Addresses are implementation. Intent is the interface.

In traditional stacks, every workload is born as an IP address. Before it can do anything useful, somebody must:

  • Carve CIDR blocks and subnets.
  • Run DHCP / IPAM workflows to assign addresses.
  • Attach CNIs and overlays to make those addresses routable.
  • Mirror the same intent again in firewall rules and policies.

Celluster starts from the opposite side:

Cells are defined by intent, zone, and tenant, not by IP.
The substrate is free to use whatever addressing is efficient underneath, while the intent graph stays stable.

The IP Plumber Problem

Every new service becomes another pipe to maintain.

If you are an SRE or network engineer, your job probably looks like plumbing:

  • Expanding subnets when teams run out of IPs.
  • Debugging “it worked in dev” because address plans differ across environments.
  • Tracking down ghost entries in ARP, DNS, or conntrack after rollouts.
  • Coordinating IP address changes with security, platform, and app teams.

None of this work is product value. It is just the cost of representing workloads as IP endpoints.

Every IP you assign is another thing to migrate later.

Celluster: IP-Less at the Edge, Not in the Kernel

We are not deleting IP. We are deleting IP from your backlog.

Celluster maintains whatever addressing it needs inside the substrate. But to operators and developers, the surface area is:

  • intent: what this Cell is and how it should behave.
  • zone: where it belongs (cluster, region, cloud, slice).
  • tenant: who owns it.

Routing and reachability are derived from this semantic graph, not from manual IP accounting.

Think of it as: “No IPAM in your workflow. Plenty of IP where it belongs.”

IPAM & CNI vs Intent-Native Cells

What changes, what stays, and who feels it.

Aspect Traditional (IP / IPAM / CNI) Celluster Cells
Identity Workloads known by IP, hostname, or interface. Workloads known by intent + zone + tenant; IP is internal detail.
Onboarding Request subnet, carve CIDR, allocate IP, wire CNI, update security rules. Declare intent; Cells are born with correct reachability and policy from the start.
Change IP changes ripple into DNS, firewalls, ACLs, configs, scripts. Change intent or zone; Cells migrate with preserved semantics. Addressing churn is absorbed by the substrate.
Operations IPAM tickets, IP conflicts, CNI upgrades, overlapping CIDRs. Operators reason about who can talk to whom, and why, not about addresses.
Multi-cloud Gateways, VPN meshes, NAT, duplicated address plans. Intent graph spans zones and clouds; Cells discover peers via semantics.

Goodbye IPAM Queue

The IPAM bureaucrat meets intent-only Cells.

Cartoon of an IPAM bureaucrat stamping forms versus happy intent-only Cells walking through a gate.
No IPAM • No CNI • No DHCP. Just intent. Cells carry an intent badge instead of an address form.

In many organisations, getting an IP address is still a ticket. With Cells, the ticket is replaced by a declaration:

  • intent: api.orders.v1
  • zone: eu-prod
  • constraints: latency, tenant, risk band

The substrate takes care of fit, placement, reachability, and any addressing details required to make it real.