How Digital Voice Modes Are Linked, and How Large Multimode Networks Are Built

Look at any big digital voice system and you will see a wall of ways in: a BrandMeister talkgroup, a TGIF talkgroup, a Fusion room or two, a D-Star reflector, maybe NXDN, P25, and M17, all reaching the same conversation. This page explains how the modes get linked together, what tools do the linking, how the large multimode networks are built, and why BrandMeister plays by a different set of rules than the rest.

The pieces, and the words for them

Four kinds of thing are in play. Keeping them straight makes everything else easy.

The two ways to connect two systems

There are only two, and the difference is the whole story.

Direct DMR-to-DMR bridge, the kind BrandMeister does not allow A BrandMeister talkgroup connected directly, server to server, to a talkgroup on another DMR network, with no reflector in between. Marked not allowed. Direct DMR-to-DMR Bridge The kind of link BrandMeister does NOT allow BrandMeister master server Talkgroup 31234 Another DMR Network TGIF · FreeDMR · DMR+ Talkgroup 31234 A BrandMeister talkgroup wired straight to a talkgroup on another DMR network, server to server, with no reflector in between. BrandMeister removes talkgroups bridged this way. The allowed method is to link through a reflector.
A direct DMR-to-DMR bridge, server to server, with no reflector.
Two DMR networks linked through a shared reflector, the kind BrandMeister allows BrandMeister and another DMR network each link their own talkgroup to a shared XLX reflector in the middle. The reflector relays audio between them. Marked allowed. Two Systems via a Reflector The kind of link BrandMeister DOES allow BrandMeister master server Talkgroup 31234 XLX Reflector shared meeting room relays audio to all links Another DMR Network TGIF · QUADNET · DMR+ Talkgroup 31234 Each system links its own talkgroup to the shared reflector, not to each other. The reflector relays the audio, so both sides hear the same conversation. BrandMeister allows reflector links (XLX, URF, YSF). This is how one system reaches many networks.
Both systems link to a shared reflector, which relays between them.

How different modes get linked

Two systems on the same mode are the easy case. The harder and more interesting case is linking systems that use different modes, so a DMR user and a D-Star user and a Fusion user can all talk on one conversation. That is what makes a system multimode.

The catch is that each mode codes its audio differently, so a DMR signal means nothing to a D-Star radio as-is. Linking modes needs a transcoding step that converts between the modes' vocoders. Multiprotocol reflectors and bridges do that job, which is how the modes end up in the same room. XLX and URF reflectors, for example, accept DMR and D-Star, and with the Fusion server pieces (YCS, FCS) they take in C4FM as well, so a single reflector module can hold several modes at once. The detail of how the vocoders convert is on the Transcoding and AMBE and Codec 2 pages.

So the reflector plays two roles at the same time. It ties separate networks together (the picture above), and it ties separate modes together (the transcoding). Both jobs happen in the same shared room, which is why a large reflector can present a dozen ways in and put them all on one conversation.

The server-to-server plumbing

When systems are joined at the server level instead of at a reflector, a few common tools do the work.

OpenBridge, for the technically curious. It is built on the MMDVM protocol and carries DMRD (DMR data) packets only. It runs over UDP 62035, and each packet is signed with an HMAC-SHA1 signature so the far end can trust who sent it. There is no login or keep-alive handshake and no automatic routing, so both ends are configured by hand with matching endpoints and matching talkgroup lists. BrandMeister and IPSC2 servers have exchanged talkgroup traffic over it since 2018.
OpenBridge linking a BrandMeister master server to an IPSC2 server A BrandMeister master server on the left and an IPSC2 server on the right, joined by a single OpenBridge trunk in the middle, labeled as applied-for and supervised. OpenBridge: a supervised server-to-server link Allowed by BrandMeister, but only by application BrandMeister master server Agreed talkgroup(s) IPSC2 server DMR+ · FreeSTAR Same talkgroup(s) OpenBridge UDP 62035 · HMAC-SHA1 signed · both timeslots The two operators each configure their own end with the same talkgroup list. Traffic then passes both ways. BrandMeister grants the link case by case. No re-bridging: audio handed over this link may not be passed onward to a third server.
OpenBridge is a direct link between two servers, not a shared reflector. On the BrandMeister side it is granted only by application.

A related but different tool is the DMR Gateway (G4KLX), which runs on a hotspot or repeater and lets one device reach several networks at once. That joins one radio's gateway to many networks, not one network to another. Keep the two straight: HBlink and OpenBridge join networks, the DMR Gateway joins one radio's gateway to many networks.

BrandMeister's rule, and why it exists

BrandMeister does not allow the first kind of connection to other DMR networks. Its own documentation states that it will not provide bridges to other DMR networks (DMR+, FreeStar, FreeDMR, and so on), and talkgroups found bridged that way are removed by the administrators. It does allow reflector links through XLX, YSF, WIRES-X, and NXDN, and it allows OpenBridge links on a supervised, case-by-case basis. Cross-mode reflector bridging is fine.

The rule in one line: link to a reflector, yes. Wire one DMR network straight into another, no.

OpenBridge is worth a closer look here, because it is a direct server-to-server link, which sounds like the very thing BrandMeister refuses. The difference is control, not the wire. An OpenBridge link to BrandMeister is applied for through a support ticket, approved case by case, limited to your own or your group's IDs, and forbidden from re-bridging. BrandMeister says such bridges "are not guaranteed to be provided" and that it "may deny requests for any reason." The re-bridging rule is the sharp one: any connection that shows talkgroups being re-bridged to another server "will be disconnected without notice." So the sanctioned, supervised link is allowed; the quiet, unmanaged one is not.

It is easy to read all of this as territorial, but it lines up with sound network management. Uncontrolled network-to-network bridging causes real, well-known problems.

There is also a scaling reason, and it is the one that has grown over time. Direct network-to-network bridging does not scale, because the number of possible links grows far faster than the number of networks. Two networks have one possible link between them. Three have three. Nine have up to thirty-six. Every one of those links is another place for a loop to form or a number to collide.

The reflector model scales the opposite way. A reflector is a hub, so each network links to it once. Nine networks means nine links, not thirty-six, and the hub keeps the routing under control so loops cannot form. Hub-and-spoke grows in a straight line while a full mesh grows like a curve. It is the same reason the telephone system and the internet use central exchanges instead of wiring every party directly to every other party.

The protocol capability and the policy are two separate things. OpenBridge and HBlink can technically connect almost anything. Whether BrandMeister enables a given link to another DMR network is a decision its administrators make, and for the unsupervised kind the answer is no.

How large multimode networks are built

Put all of the above to work and a pattern appears. The same era and the same tools produce two very different designs, and the choice between them decides whether a system can include BrandMeister at all. Two real networks show the split.

Model A: keep the networks separate (Colorado Digital)

Colorado Digital runs the XLX303 reflector and deliberately keeps BrandMeister on its own set of modules. In their own words: "As a rule we coded in BM routing on Modules A through H. This allows Modules I, J, K, L, M and beyond to route other places such as other XLX Reflectors such as XLX720 or other networks such as the TGIF DMR Network."

So BrandMeister sits on modules A through H, cross-mode systems like WIRES-X and AllStar ride along on those same modules (allowed, since they are different modes, not another DMR network), and the other DMR networks live on separate modules further down. BrandMeister never shares a room with another DMR network. That segregation is the practical form of the no-re-bridging rule, and the result is that Colorado Digital keeps its BrandMeister presence.

Model B: bridge everything together (SouthEast Link)

SouthEast Link takes the opposite approach. Its bridge system ties talkgroup 43389 across many DMR networks at once (AmComm, DMR+, HAM United, QRM, TGIF, Extended Freedom) and then out to the modes as well: Fusion through YSF, FCS, and YCS rooms, D-Star through URF and XRF, plus NXDN, P25, M17, WIRES-X, AllStar, and Echolink. In their words, "all of these are bridged together, so talk to your buddies no matter what popular digital mode is used."

Because that design bridges multiple DMR networks together on one talkgroup, which is the re-bridging BrandMeister prohibits, BrandMeister cannot be a full member. Their own site says a repeater currently switches between BrandMeister and the bridged networks, and that "BM will be dropped in the future and HAM United will be full time." So a bridge-everything system ends up dropping BrandMeister, because the two designs cannot coexist.

That is the whole lesson in one comparison. Segregate the DMR networks onto separate rooms and you can keep BrandMeister (Colorado Digital). Bridge them all together and you cannot (SouthEast Link). The policy is not about who is allowed in. It is about whether the architecture re-bridges DMR networks or keeps them apart, and the operators treat it as an engineering choice.

Why the numbers are different on each network

One more thing that confuses people: the same system often has a different talkgroup number on each network it appears on. Each network assigns its own number, and there is no rule that they match. Two things drive it.

Reading a cross-reference table or a dashboard

Once you know the rule, both a table and a live reflector dashboard are easy to read.

You can see both patterns in the Linked & Bridged Talkgroup / Reflector Cross-Reference.

The short version

Modes are linked by transcoding them together in a shared reflector room. Networks are linked either directly, server to server, or by each joining a shared reflector. The tools that do it are HBlink3, OpenBridge, and IPSC2 at the server level, and XLX, URF, YSF, FCS, and YCS reflectors at the meeting-room level. Large multimode networks are built one of two ways, by keeping the DMR networks on separate rooms or by bridging them all together, and that single choice decides whether BrandMeister can be part of it. Systems meet at reflectors, each network keeps its own number as its door into the room, and BrandMeister will join a reflector but will not run an unsupervised line into another DMR network. That is a management choice, not a grudge.

Return to n6jet.com  —  73 de N6JET

Sources and further reading: Bridging to BrandMeister (official docs) · Open Bridge (BrandMeister Wiki) · BrandMeister and IPSC2 traffic exchange (2018) · FreeSTAR — IPSC2 server · DMR+ overview (Kings Point ARC) · Colorado Digital — XLX303 · LMARC — SouthEast Link DMR · Linked & Bridged Cross-Reference