New Carrier VoIP switch in production

Statistically, we had almost zero outages from 2014 to 2024. If there were any, they were due to a feature upgrade where a technician made a mistake, resulting in one or two brief interruptions per year. However, with our Carrier Switch solution in 2025, which until last night was running in a Windows cluster environment in our data center, the problems accumulated and increased sporadically. Uncontrolled behavior of some services even posed a challenge for our solution's manufacturer, as we had already confronted them with the issue at the beginning of 2025.

He, in turn, was faced with the puzzle of why this was happening, and analytical evaluations did not lead to a definitive solution. However, he confirmed that other customers were also experiencing problems in certain Windows Datacenter Cluster environments, although the causes were sometimes different there as well.

Back then, we were told that the manufacturer was working hard internally to port the entire software stack (which required a lot of work) to Linux, and a release for Linux was indeed available starting in August 2025.

Since we have always had a good relationship with the manufacturer in the past, and have even contributed to certain features, we decided in 2025 to stick with what we know, but we didn't want to take the risk of switching to Linux right away with the first versions, because on the one hand, it involves a certain complexity that increases depending on the scope of the features, and on the other hand, because two or three features that we had in the Windows cluster version were not yet fully available in the new Linux version, or were still in the beta stage.

In January 2026, we were finally able to breathe a sigh of relief, as other telephony providers had taken the plunge in the late autumn, and it was working reliably, even under load, and had met our critical requirements.

In February 2026, we started deploying a Linux cluster in the data center, and we conducted initial tests in a small, not fully configured environment, which were excellent.

The migration.

iQcom (Switzerland) GmbH manages over 15,000 lines with its telephony and internet services. Many customers have been with us for more than 10 years, and thankfully, they have remained loyal even during the recent unstable months (with the exception of one SME).

Of course, over the years, we have accumulated vast amounts of customer data. Just the volume related to our pioneering role, alongside WebRTC (since 2011), in email-to-fax, fax-to-email (PDF), and print-to-fax with the former QuadCom Faxbox, now iQcom Online Fax, is immense. In addition, there are dozens of other customer data silos, and the fact that we are directly connected to 400 telecommunications companies and maintain a Tier 1 connection with 3,100 telecommunications companies worldwide means that the configuration volume is gigantic.

With the new Linux version, new user interfaces based on the latest technology have also been introduced, which naturally need to reflect these settings and files and also ensure the functionality of existing devices, even if some models are no longer supported.

On Wednesday evening at 7:00 PM, the Windows cluster, which had received new hardware just a year ago, was shut down with a celebratory announcement from iQcom's CTO, and the new Linux cluster took over its service with a brief interruption, which also involved an IP and DNS change.

Thank you

We would like to take this opportunity to thank, on the one hand, the solution provider, whose team of four supported us throughout the entire migration process over the past three months, and, on the other hand, our customers, who remained loyal to us despite the occasional outages in recent months.