SberDevices has integrated Matter over Wi‑Fi into its SberBoom 2.0 smart speakers, tackling certification and device variety hurdles to enhance interoperability and control in smart homes.
SberDevices has added Matter over Wi‑Fi support to its SberBoom 2.0 smart speaker line, turning the device into a home hub that works with both Zigbee and Matter. The company said the integration was not straightforward because Matter had to be wired deeply into the back end rather than handled as a thin application layer. The result, according to the company, was a system that moves certificate generation into the back end, hands controller duties between a phone and the speaker, and uses Sber’s own scripting language to handle device typing. Matter itself was designed to reduce the fragmentation that has long defined the smart home market, where interoperability has often depended on the vendor ecosystem chosen at purchase.
Matter commissioning is a multi-step process built around secure discovery, onboarding payloads, PASE, certificate validation, network provisioning and CASE session establishment. Silicon Labs’ commissioning guide and Google’s developer documentation both describe the commissioner as the device that initially discovers and provisions a Matter device before it joins the operational fabric. In Sber’s implementation, that role is split more deliberately: the smartphone running the Salyut app performs the first pairing near the device, while the SberBoom 2.0 speaker takes over as the long-term controller. That change is meant to reduce friction in apartments and homes where Bluetooth Low Energy range would otherwise force a user to carry the hub from room to room.
The company said it could not rely on Google Home Mobile SDK, because it is tied to Google Mobile Services and would have limited the pool of compatible phones. Instead, SberDevices based its work on the open code available from the Connectivity Standards Alliance repository and used that as the starting point for its Android pairing flow. The company said the demo code would connect and control a test device, but production use required substantial hardening and much tighter control over configuration. That included moving operational credential generation from the client library into Sber’s back end so the system could keep a stable record of homes, devices and secrets.
One of the earliest engineering problems was controller configuration. The initial builder-based setup for the Matter controller accepted a wide range of parameters, but it did not enforce enough validation, so missing or malformed values only surfaced as opaque failures later in the commissioning flow. The company said it refined the configuration model after comparing notes with colleagues working on Apple’s Matter framework, which offers clearer contracts for custom certification flows. That helped define which parameters had to be passed into the controller before pairing could begin.
Certificates were the second major obstacle. Matter certification depends on cryptographic artefacts such as the Device Attestation Certificate, the Product Attestation Intermediate and the operational certificate issued to a node after commissioning. Industry guidance on Matter certification stresses that these artefacts are central to proving device identity and trust, and SberDevices said it had to generate and store them on the back end rather than relying on demo defaults. The company created a network-manager microservice to hold fabric data, generate node identifiers and issue Node Operational Certificates, which also makes it possible to support pairing from a web application that does not have direct access to the home network.
SberDevices also had to solve an implementation detail that caused commissioning to fail intermittently. The company said that although Matter uses unsigned 64-bit identifiers, values in the upper half of the range broke the cryptographic libraries it was using, so the back end had to restrict generated identifiers. It also had to reconcile differences between Android and C++ cryptography formats, and then adapt the key material so the Matter library would accept it. Once that was fixed, the company said it could complete commissioning reliably enough to proceed with the handover of admin rights from phone to speaker.
That handover is another part of the flow that SberDevices implemented itself. After the smartphone has paired a device, the system updates the access-control list so the speaker becomes the administrator and the phone is removed from that role. The company said that simple calls to set the admin subject were not sufficient, so it instead reads the ACL, adds the speaker as an administrator, removes the phone and writes the updated list back to the device. Once that is done, the speaker can interrogate the device for supported capabilities and pass the result back to the back end.
The deeper challenge, however, was not commissioning but device variety. Matter defines a clear hierarchy of nodes, endpoints and clusters, yet SberDevices found that real products often describe themselves less consistently than the standard suggests. The company said it encountered devices that reported the wrong device type, custom clusters, unexpected attributes and even reused Zigbee cluster identifiers inside Matter implementations. That is why it built a template layer based on its Fortuna scripting language, which it originally designed for low-memory home-hub scenarios. Fortuna compiles scripts to bytecode in advance, runs from fixed buffers and lets Sber map a discovered Matter device to the closest supported behaviour without hard-coding a new C++ rule for every model.
According to SberDevices, the template system now lets the hub read a device during commissioning, inspect its endpoints, clusters and feature map, and then generate a small script tailored to that model. The company said this is especially important for lighting, where Matter devices may support different combinations of on/off, level control, colour temperature, colour and power measurement. By translating those variations into a common internal model, Sber can keep the controller generic while absorbing vendor-specific quirks in a narrow adaptation layer. The company said that is the practical basis for its next step, Matter Bridge support.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
