One product.
Every service included.
Netvyn is a single broadband network gateway. Subscriber access, authentication, bandwidth control, address translation and filtering are not add-ons or premium modules — they are what the gateway is. You buy one product and every service below is in it.
01 What is included
Nine services, one configuration, one commit
Each of these is normally a separate purchase from a separate vendor with its own change window. In Netvyn they are parts of the same gateway, applied to the same subscriber session.
BNG / BRAS
Subscriber termination and session management for PPPoE and IPoE on one gateway.
Included AccessPPPoE
Session-based access with full protocol handling and per-MAC session protection.
Included AccessIPoE
DHCP-driven access with Option-82 circuit identity, alongside PPPoE on the same gateway.
Included Authentication & accountingRADIUS
Authentication, authorization and accounting against the RADIUS you already operate.
Included Traffic policyQoS
Per-subscriber shaping nested under the real contention points in the access network.
Included Address translationNAT
Stateful translation with the mapping behaviour real-time applications depend on.
Included Address translationCGNAT
Shared public IPv4 with deterministic port blocks and one record per allocation.
Included SecurityFirewall
Stateful filtering bound to the session, with address groups your billing system drives.
Included Access network & addressingVLAN & IP management
Single-tag and QinQ access models, with address pools that extend while live.
Included02 BNG / BRAS
Subscriber termination
A broadband network gateway your operations team can actually run
The BNG is where every subscriber session begins, gets its identity, gets its address and gets its plan. Netvyn owns that whole boundary in one platform, so session state has one home instead of four.
Specification
- Category
- BNG / BRAS
- Access types
- PPPoE, IPoE
- Address families
- IPv4 + IPv6
- Session limit
- Licence-sized
- Redundancy
- active / standby
Unified session table
PPPoE and IPoE sessions live in one table with one identity model, so reporting and support do not fork by access type.
Policy bound at connect
Bandwidth, filtering and translation policy attach as the session comes up, and can be changed while it stays up.
Accounting your billing trusts
Start, interim and stop records with per-session byte and packet counters, in the format your AAA already ingests.
Controlled teardown
Disconnect a subscriber by username, address or session id — individually, per interface, or in bulk.
Live visibility
Session counts by access type and state, per-subscriber detail, and resolved policy with its source.
Redundancy
An active/standby pair replicates session, translation and accounting state so a failover is not a mass re-dial.
How it works
-
01
Session request
A subscriber’s CPE initiates PPPoE, or the access node relays DHCP for IPoE.
-
02
Identity
Netvyn presents the username, MAC or circuit ID to your RADIUS and waits for the verdict.
-
03
Address
The accepted session draws an address and, where used, an IPv6 prefix from the right pool.
-
04
Policy
Plan and filter policy bind to the session; accounting starts.
<Netvyn-BNG> show sessions Access Sessions State ------- -------- ----- PPPoE 14 220 up IPoE 5 780 up <Netvyn-BNG> show subscribers IP Profile Down Up Source 100.64.10.14 res-100 100M 100M radius <Netvyn-BNG> clear session user "subscriber-id"
03 PPPoE
Access
PPPoE subscriber aggregation at ISP scale
PPPoE is still how most broadband subscribers connect, and it is where an access network is most easily destabilised by CPE that misbehaves. Netvyn handles the protocol completely and protects itself while doing it.
Specification
- Identity
- username
- Alternate
- MAC address
- Auth methods
- PAP, CHAP, MS-CHAP
- Address family
- IPv4 + IPv6 PD
- Protection
- per-MAC / per-VLAN
Complete protocol handling
Discovery and link negotiation are handled in full, including the keepalive that detects a dead peer instead of leaking a session.
Authentication choice
PAP, CHAP and MS-CHAP are all available; Netvyn proposes the strongest method you allow and negotiates down only within that set.
Session protection
Per-MAC and per-VLAN limits, half-open session reaping, and rate limiting on discovery keep one bad CPE from becoming an outage.
Address fallback
If RADIUS returns no address, a configured pool takes over — so an AAA hiccup does not stop subscribers connecting.
Dual-stack
IPv4 and IPv6 negotiate on the same session, including a delegated prefix for the subscriber’s own network.
Diagnosable
Discovery counters and per-VLAN tracing tell you whether a failing subscriber is a tagging problem or an authentication problem.
How it works
-
01
Discovery
The CPE finds the gateway and a session is offered and confirmed.
-
02
Negotiate
Link parameters are agreed, including the maximum frame size for PPPoE.
-
03
Authenticate
Credentials go to RADIUS; the reply carries the subscriber’s policy.
-
04
Address & go
IPv4 and IPv6 configuration completes, policy binds, accounting starts.
[Netvyn-BNG] ppp interface add access0 access1 [Netvyn-BNG] ppp auth-protocol chap,pap [Netvyn-BNG] ppp lcp-echo interval 30 failure 3 [Netvyn-BNG] ppp ipv6 enable [Netvyn-BNG] commit [Netvyn-BNG] save # confirm <Netvyn-BNG> show ppp <Netvyn-BNG> show sessions
04 IPoE
Access
IPoE access without the PPP overhead
Where you do not want a session protocol between the subscriber and the gateway, IPoE gives you DHCP-driven access with the subscriber identified by the access network itself.
Specification
- Identity
- Option-82 circuit
- Alternate
- MAC address
- Addressing
- DHCPv4 / DHCPv6
- Prefix delegation
- supported
- Coexists with
- PPPoE, by VLAN
Identity from the network
The access node’s circuit and remote identifiers become the subscriber name, so identity does not depend on CPE configuration.
MAC-based authentication
Where circuit information is unavailable, the client MAC authenticates against RADIUS in the format your realm expects.
Coexists with PPPoE
Different VLAN ranges on the same port can run different access types, so a migration is gradual rather than a cutover.
Lease control
Lease time, subnet and DNS are configured centrally and handed to every subscriber on that access domain.
Same policy pipeline
IPoE subscribers get the same bandwidth, filtering and translation treatment as PPPoE subscribers.
Dual-stack
IPv6 addressing and prefix delegation are available to IPoE subscribers as well.
How it works
-
01
Request
The subscriber’s device sends a DHCP request; the access node adds its circuit identity.
-
02
Identify
Netvyn presents that identity, or the MAC, to RADIUS.
-
03
Lease
An address is leased from the pool selected for that customer or access domain.
-
04
Policy
Plan and filter policy bind to the session; accounting starts.
[Netvyn-BNG] dhcp mode ipoe [Netvyn-BNG] dhcp interface add access2 [Netvyn-BNG] dhcp auth option82 [Netvyn-BNG] dhcp lease-time 3600 [Netvyn-BNG] ip dns 203.0.113.53 198.51.100.53 [Netvyn-BNG] commit [Netvyn-BNG] save
05 RADIUS
Authentication & accounting
Your RADIUS stays your RADIUS
Netvyn does not want to be your subscriber database. It authenticates against the AAA you already run, takes each subscriber’s policy from the reply, and sends usage back in the records your billing system already reads.
Specification
- Policy source
- Access-Accept
- Accounting
- start / interim / stop
- Mid-session
- CoA + disconnect
- Failover
- ordered server list
- Integrity
- signing on by default
Policy from AAA
Rate, pool and filter policy arrive in the accept reply, so a plan lives in your billing system rather than in gateway configuration.
Mid-session changes
An upgrade, a speed boost, a suspension or a disconnect can be pushed to a live session without dropping the subscriber.
Server failover
An ordered server list with configurable timeout and retry count, plus per-server latency and failure counters.
Message integrity
Request signing is on by default, which is both current best practice and increasingly a hard requirement of modern RADIUS servers.
Usage records
Interim accounting at your chosen interval keeps usage current for quota and fair-use enforcement.
Operator login
Staff logins to the gateway itself can also be delegated to a RADIUS realm, with the role returned by the server.
How it works
-
01
Request
Netvyn sends the subscriber identity and access details to your RADIUS.
-
02
Accept
The reply carries the plan: rate, pool, filter policy, session limits.
-
03
Account
Start, interim and stop records report usage back for billing.
-
04
Change
Your systems can push a new plan or a disconnect to the live session.
[Netvyn-BNG] radius server add 198.51.100.10 secret <shared-secret> [Netvyn-BNG] radius nas-id netvyn-bng-01 [Netvyn-BNG] radius interim-interval 300 [Netvyn-BNG] radius message-authenticator on [Netvyn-BNG] commit <Netvyn-BNG> show radius statistics
06 QoS
Traffic policy
Bandwidth control that matches what you sold
Every subscriber has a plan, and every access network has places where that plan competes with others. Netvyn enforces the plan per subscriber and nests it under the real contention points, so the network behaves the way your product promises.
Specification
- Shaping
- hierarchical
- Direction
- up + down
- Rate source
- RADIUS or static
- Queueing
- per-profile
- Mid-session
- supported
Per-subscriber shaping
Separate upstream and downstream rates with a configurable burst allowance for short peaks.
Hierarchical limits
Subscriber limits nest under an aggregate that represents a real shared resource, so oversubscription is enforced where it exists.
Queue behaviour you choose
Different queueing disciplines are available per plan, including modern fair-queueing for latency-sensitive traffic.
Rate from anywhere
A rate can come from the authentication reply, from a static assignment, or from the service default — and the resolved source is visible per subscriber.
Tuning without downtime
Queue tuning applies to new sessions immediately and migrates existing sessions gradually, with no forwarding interruption.
Honest speed tests
Overhead compensation accounts for framing so a customer’s speed test reads the number you sold them, not the number minus protocol overhead.
How it works
-
01
Define
A queue behaviour and a plan rate are defined once, as reusable profiles.
-
02
Bind
The plan attaches to a subscriber from RADIUS, or to an address range statically.
-
03
Enforce
Upstream and downstream traffic is shaped per subscriber under its parent aggregate.
-
04
Adjust
A new rate can be pushed mid-session, or the profile retuned live.
[Netvyn-BNG] qos queue-profile add res-fq fq_codel qsize 4096 [Netvyn-BNG] qos subscriber-profile add res-100 \ rate-down 100M rate-up 100M queue-profile res-fq [Netvyn-BNG] commit [Netvyn-BNG] qos assign add 100.64.10.0/24 subscriber-profile res-100 [Netvyn-BNG] commit ; save <Netvyn-BNG> show qos bindings
07 NAT
Address translation
Translation that does not break your subscribers’ applications
Subscribers on private addressing need translation, and the way that translation behaves decides whether video calls, gaming and peer-to-peer applications work. Netvyn defaults to the behaviour that keeps them working.
Specification
- Placement
- on the gateway
- Extra appliance
- none
- Table size
- memory bound
- Timeouts
- per protocol
- Publishing
- destination NAT
Application-friendly mapping
The default mapping model is the one interactive and peer-to-peer applications expect, rather than the strictest option.
Fragment handling
Fragmented traffic is translated correctly, so large-datagram applications do not silently freeze.
Port forwarding
A business customer’s service can be published on a public address and port, with ownership verified so one customer cannot claim another’s address.
Tunable timeouts
Per-protocol idle timeouts let you balance table size against how long an idle connection survives.
Policy-driven
An ordered policy list decides which subscriber ranges are translated, which bypass translation, and which get static treatment.
Diagnosable
Counters separate a missing mapping from an expired one from a fragment problem, so a one-way flow has an answer instead of a theory.
How it works
-
01
Classify
A subscriber’s traffic is matched against the translation policy list.
-
02
Translate
The private source address and port are rewritten to a public pair.
-
03
Track
The connection is tracked so replies are translated back automatically.
-
04
Expire
Idle connections age out on per-protocol timers you control.
[Netvyn-BNG] nat enable [Netvyn-BNG] nat udp-mapping endpoint-independent [Netvyn-BNG] nat policy add nat-rule 100.64.0.0/10 pool pub1 [Netvyn-BNG] commit # publish a customer service [Netvyn-BNG] nat dstnat add 203.0.113.5 proto tcp dport 8080 \ to 100.64.0.9 to-port 80 [Netvyn-BNG] commit
08 CGNAT
Address translation
Carrier-grade NAT you can still answer a subpoena with
CGNAT is how an ISP keeps growing after IPv4 has run out. The usual price is a logging problem: thousands of records per subscriber per hour. Netvyn assigns each subscriber a fixed block of ports, so one record covers the whole session.
Specification
- Allocation
- deterministic block
- Log volume
- per allocation
- Sharing ratio
- tunable
- Traceability
- single record
- Deployment
- inline or fronting
Deterministic allocation
Each subscriber gets a fixed, reproducible block of ports on a public address — the mapping is computed, not looked up in a log.
One record per block
Compliance logging volume is proportional to subscribers rather than to their connections, which is the difference between kilobytes and terabytes.
Traceability on demand
“Who held this public address and port at this time” resolves from a single self-contained record.
Tunable sharing ratio
Ports per subscriber is your dial between address efficiency and headroom for connection-hungry households.
Per-subscriber visibility
Block utilisation per subscriber distinguishes a heavy user from a genuinely undersized public pool.
Works with your existing edge
CGNAT can front an existing gateway, or run as part of the same Netvyn instance that terminates the subscribers.
How it works
-
01
Allocate
A subscriber address is assigned a fixed port block on a public address, and that allocation is logged once.
-
02
Translate
Every connection from that subscriber uses a port inside its own block.
-
03
Return
Replies to that public address and port map back to exactly one subscriber.
-
04
Trace
A request for a past address and port resolves from the block record alone.
[Netvyn-BNG] nat pool add pub1 203.0.113.10-203.0.113.20 \ ports 1024-65535 block 512 deterministic [Netvyn-BNG] nat policy add nat-rule 100.64.0.0/10 pool pub1 [Netvyn-BNG] commit ; save <Netvyn-BNG> show cgnat port-blocks summary <Netvyn-BNG> show nat statistics
09 Firewall
Security
Filtering as part of the subscriber’s policy
A separate filtering appliance is another device to size, patch and correlate. Netvyn applies filtering as policy on the session it already owns, and lets your billing system move a subscriber into a restricted group without touching a rule.
Specification
- Rules
- stateful, ordered
- Grouping
- named address lists
- Driven by
- RADIUS / API
- Anti-spoof
- source verification
- IPv6
- separate rule set
Stateful rules
An ordered rule list that understands connection state, so return traffic does not need a mirror rule.
Named address groups
One rule can cover thousands of addresses, and membership changes without editing the rule.
Billing-driven quarantine
Your systems can tag a subscriber into a group for the life of their session — walled garden on suspension, released on payment.
Anti-spoof
Source verification stops a subscriber sending traffic from an address that is not theirs.
Separate IPv6 policy
IPv6 filtering is its own rule set, so an IPv4 rule can never accidentally govern IPv6 traffic.
Per-rule counters
Every rule reports its own hit count, so you can prove which rule is doing the work.
How it works
-
01
Define
Rules and named address groups are defined once for the gateway.
-
02
Bind
A filter policy attaches to subscriber sessions, from configuration or from RADIUS.
-
03
Enforce
Traffic is classified against the policy on the way through.
-
04
Adjust
Group membership changes take effect immediately, mid-session.
[Netvyn-BNG] firewall enable [Netvyn-BNG] firewall urpf strict [Netvyn-BNG] firewall filter add accept state established [Netvyn-BNG] firewall filter add drop src-list quarantine [Netvyn-BNG] commit ; save # immediate, no commit needed [Netvyn-BNG] firewall address-list add quarantine 100.64.7.12
10 VLAN & IP management
Access network & addressing
Access topology and addressing, managed together
How subscribers are tagged and how they are addressed are the two things that make an access network hard to change later. Netvyn keeps both declarative, so a new region or a new wholesale partner is configuration rather than redesign.
Specification
- Single tag
- 802.1Q
- Double tag
- QinQ
- IPv4 pools
- live extend
- Overflow
- chained pools
- IPv6
- prefix pools
Single and double tagging
One VLAN per subscriber, one per access node, or an outer tag for the segment and an inner tag for the line.
Ranges mapped to service
A VLAN range binds to an access type and a policy set, so residential, business and wholesale traffic separate cleanly on shared ports.
Pools that grow live
Extending an address pool takes effect on commit, with no disturbance to subscribers already holding addresses from it.
Overflow pools
A pool can chain to a successor, so a busy region degrades to a second range instead of refusing connections.
IPv6 prefix pools
Carve subscriber /64s and delegated prefixes from named blocks, selected by product or by RADIUS.
Utilisation alerting
Pool usage is exposed so exhaustion is a monitored threshold rather than a support queue.
How it works
-
01
Tag
Physical ports carry tagged sub-interfaces for the access network.
-
02
Map
Each VLAN range binds to an access type and a service policy.
-
03
Address
Subscribers draw from the pool selected for their product or region.
-
04
Grow
Pools and prefix blocks extend live as the subscriber base grows.
[Netvyn-BNG] interface access0 add vlan v100 100 [Netvyn-BNG] interface access0 add vlan v10 100 10 [Netvyn-BNG] pool res range 100.64.10.10 100.64.10.250 \ gateway 100.64.10.1 prefix 24 [Netvyn-BNG] pool res next overflow [Netvyn-BNG] ipv6 pool res 2001:db8:100::/40 assign 56 [Netvyn-BNG] commit ; save <Netvyn-BNG> show pool
See the gateway on your network
One demo covers the whole product — your access model, your RADIUS attributes, your address plan. Thirty minutes with a network engineer, no sales deck.