Skip to content
Carrier-grade software BNG · in production

The broadband gateway, rebuilt as software

GridM BNG terminates PPPoE and IPoE subscribers at line rate on commodity x86, speaks RADIUS to the billing you already run, and scales out node by node. No chassis, no forklift upgrades, no lock-in.

PPPoE + IPoE250K+ sessions per nodeRADIUS + CoACGNAT + IPv6

GridM BNG · Console

cluster: delhi-core · 2/2 nodes online

Live

▼ Download

62.2 Gbps

▲ upload 4.44 Gbps

Subscribers

23,312

PPPoE + IPoE · 0 faults

Live network throughput

15m1h6h24h
Download Upload Subscribers

Per BNG

N+1 · session sync

bng-n1 · 11,868 subs31.8G 2.22G
bng-n2 · 11,444 subs30.5G 2.23G
IPv6 Adoption17.7%
NAT Port Pressure62%
DDoS Protection98 mitigating

IPDR flows today

1,842,032,315

Flows / last 5 min

12,252,753

Terminating subscribers for operators like

RealtelAeroFiberNorthWaveCityLinkSkyMeshVertexNet
The problem

Legacy BNG chassis are where budgets go to die

The box that terminates your subscribers was designed for a world without cloud, APIs or commodity 100G. Operators pay for that history every year.

Forklift upgrades

Chassis hit end-of-life or a capacity ceiling, and the only path forward is another seven-figure box with months of lead time.

Licensing that fights you

Feature unlocks, scale licenses and support renewals are bound to the chassis and renegotiated on the vendor's terms, not yours.

Closed, CLI-only boxes

No real APIs, no central console, no automation. Every change is a maintenance window and every integration is a professional-services quote.

Scaling by chassis, not by need

Capacity comes in giant, expensive steps. You overprovision for years or run hot with no headroom, and redundancy doubles the bill.

GridM BNG

Everything the chassis did. Everything it never could.

One horizontally scalable software package: session termination, AAA, addressing, QoS, CGNAT, HA and observability.

Subscriber session termination

PPPoE and IPoE (DHCP) termination that scales to hundreds of thousands of sessions per node.

AAA and policy

Standards-based RADIUS that drops into the billing and CRM stack you already run.

Addressing, IPv6 and CGNAT

Stretch scarce IPv4 with integrated CGNAT while rolling out native dual-stack IPv6.

QoS and plan enforcement

Enforce every plan, fairly, at line rate, without external shaper boxes.

Line-rate data plane

A kernel-bypass forwarding plane that turns commodity NICs into carrier capacity.

High availability

Carrier availability from redundant software, not redundant chassis.

Drop-in architecture

Sits exactly where the chassis sat. Behaves like a cloud.

GridM BNG terminates subscribers between your access network and IP core, exactly like the hardware it replaces. Your OLTs, your RADIUS, your upstreams: nothing else changes.

  • Interoperates with any OLT, DSLAM, switch or wireless access layer
  • Same RADIUS and billing stack, with CoA for live plan changes
  • N+1 cluster with session sync: node loss is a non-event for subscribers
  • Scale from one 1U server to a terabit POP with ECMP scale-out
How it deploys
OLT / GPONFTTH accessDSLAM / SwitchDSL + ethernetWireless / RFWISP accessSUBSCRIBERSPPPoE · IPoE · QinQbng-node-01bng-node-02bng-node-NGridM BNG clusterN+1 · session sync · ECMP scale-outcommodity x86 · bare metal / VM / K8sRADIUS · BillingAAA + CoA · your stackIP core · InternetBGP · upstreams · peering
One console, cluster-wide

Run your whole network from one beautiful screen

There is no CLI to learn and no boxes to log into one by one. Every session, pool, plan, CGNAT setting and security policy is managed from a single web console that spans the whole cluster in real time. Preview a change, apply it everywhere, and read the audit trail after.

  • Central, cluster-wide control
  • Preview then apply, everywhere
  • Role-based access + audit log
  • Everything built in, nothing to bolt on
Tour the console

Illustration mirrors the shipping console. Identifiers are fictional.

250K+

Concurrent sessions per node

1.2 Tbps

Scale-out forwarding, per cluster

99.999%

Availability design target

~60%

Typical TCO reduction

Illustrative platform figures. Ask us for reference architectures and tested numbers for your hardware.

Side by side

Chassis economics vs software economics

The BNG role does not need proprietary hardware anymore. Here is the short version; the full comparisons go deeper.

Factor
GridM BNG
Legacy chassis
Capacity step
Add a commodity node
Buy another chassis
Pricing
One plan, every feature, no tiers
Hardware + feature + scale licenses
Operations
One central web console, cluster-wide
CLI, box-by-box logins
In production today

One live cluster, measured this week

23K+

subscribers online

60+ Gbps

peak served throughput

1.8B+

IPDR flow records / day

12M+

flows analyzed / 5 min

Figures from a single production GridM BNG cluster (two nodes, N+1) at an FTTH ISP in Delhi.

“We moved our subscriber termination from hardware to GridM BNG and stopped thinking about the BNG as a constraint. Sessions, CGNAT, IPv6 and per-subscriber visibility run on servers we already know how to buy and fix, and the console shows us in seconds what used to take a CLI session and a prayer.”

Network Operations

Realtel Network Services, FTTH ISP, Delhi

FAQ

Questions network engineers ask first

Straight answers on architecture, migration, licensing and interop.

A software BNG (also called a virtual BNG or vBNG) implements the Broadband Network Gateway function, subscriber session termination, AAA, addressing and QoS, as software on standard x86 servers instead of a proprietary hardware chassis. GridM BNG is a cloud-native software BNG: it terminates PPPoE and IPoE sessions at line rate on commodity NICs and runs on bare metal, VMs or Kubernetes.

See GridM BNG on your traffic

Book a live demo with a network engineer. We will walk your topology, size the hardware and show real sessions terminating on the console, not slides.

Runs on your hardware · One plan, every feature · Migrate in slices