Engineering
Migration guides, data-plane deep-dives and engineering explainers on running a carrier-grade BNG, LNS and CGNAT on commodity x86. Written by the engineers who build and operate Netavo BNG on our own network.
Map Junos MX PPPoE/IPoE access, RADIUS, address pools and L3VPN VRFs onto Netavo's single config.json.
ReadKeep FreeRADIUS and billing while moving subscriber termination from a self-built stack to a supported software BNG.
ReadWhen RouterOS PPPoE and NAT stop scaling, you can swap just the NAS for Netavo and leave PHPNuxBill, Splynx and FreeRADIUS exactly as they are.
ReadPlan a staged migration from an ASR9000, 7750 or MX to a software BNG, with checks for subscriber policy, routing and failure behaviour.
ReadOpen NAT is two independent properties, mapping and filtering. Here is what a console checks, what the CGN controls, and what the CPE controls.
ReadSeparate shared-IP reputation, geolocation errors and service selection, then fix the right problem at the right egress pool.
ReadUse subscriber-to-port-block history to answer public IP and port attribution queries without collecting a log for every flow.
ReadA public IPv4 has a fixed number of ports. How you carve them into blocks sets your address-sharing ratio and port-exhaustion risk.
ReadA CGNAT subscriber cannot host a server on a shared address. Here are the ways to give them inbound reachability, and how to choose per use case.
ReadSeparate the subscriber WAN from the routed LAN prefix, then validate RA, SLAAC and DHCPv6-PD against your access protocol and CPE.
ReadCalculate the L2TP and PPPoE packet budget, diagnose size black holes, and use MSS clamping alongside working path-MTU discovery.
ReadThe two ways to terminate broadband subscribers differ in how they identify a line, what they cost in overhead, and who they trust. Here's how to choose.
ReadUse RFC 5176 CoA for live rate and walled-garden changes, and Disconnect-Message for explicit session teardown.
ReadA RADIUS failure should stop new logins, not existing sessions. Here is what survives, what amplifies the outage, and how the BNG should behave.
ReadHigh availability is three separate failures with three different answers. Here is what survives each on a software BNG, and how to test it.
ReadCompare core, BNG and access replication for IPTV, with clear boundaries between multicast routing, IGMP membership and snooping.
ReadPolicing drops, shaping queues. One adds latency, the other does not. Here is the trade-off, and where Netavo sits today.
ReadHow packets get from the NIC to the forwarding code decides a software BNG's throughput. Here are the three approaches, what each costs, and when to use them.
ReadNetavo reaches interface line rate at IMIX on commodity x86. What actually bounds a software BNG, and the numbers behind the claim.
ReadThe NIC, the NUMA layout and how you pin cores matter more than headline core count. A practical guide to sizing a box for subscriber termination.
ReadNIC RSS can't see inside PPPoE or a single L2TP tunnel, so every session lands on one core. Here's the software-spray dispatch that fixes it, and the hardware path that avoids it.
ReadEvaluate BNG CUPS through placement, fleet scale, state ownership and failure recovery before committing to a distributed architecture.
ReadModel residual IPv4 state, address commitments and support costs to see which CGNAT investments IPv6 can actually remove.
Read