ITCore
← All articles
Telephony

Own PBX or a carrier's virtual PBX: what fits a business in Moldova

· 7 min read

The short answer: if you have a small team, the phone isn't where your sales happen and there's no CRM yet, take the cloud PBX your carrier offers and stop there. If calls are the business itself, or you need the phone system wired into a CRM or 1C, your own recording rules, or one line across several branches, an on-premise PBX on Asterisk, FreePBX or 3CX, connected to the carrier over a SIP trunk, usually pays off. Here's how I'd weigh the two, criterion by criterion.

What you're really choosing between

Moldtelecom, Orange and Moldcell all sell cloud (virtual) PBX services. The switch runs on the carrier's side. You get numbers, set up extensions and forwarding in a web portal, and staff call from IP desk phones or an app. Features, terms and prices are the carrier's to quote.

With your own PBX, Asterisk, FreePBX or 3CX runs on your server or on a virtual machine. IP phones (Yealink or Grandstream, say) register to it, and calls reach the outside world through a SIP trunk from the same carriers. You still buy numbers and the line from a carrier, and the contract is still in your company's name. What changes is where the call logic lives and who controls it.

So the cloud PBX vs on-premise PBX question boils down to this: do you hand routing, queues and recordings to the carrier, or keep them in-house?

Five criteria, side by side

Total cost of ownership

A carrier's virtual PBX costs next to nothing up front. No server, no project. You pay a monthly fee that grows with headcount and with every feature you switch on.

Your own PBX flips that. The money goes out early: a server or VM, the setup work, sometimes licences (Asterisk and FreePBX are free, 3CX has paid editions), plus ongoing support. After that, the cost barely depends on how many extensions you run. I'm not quoting numbers on purpose, because they depend on the carrier's tariff and on what has to be configured. The only fair comparison is to add up three years of payments for your own headcount and branches and see which side comes out lower.

CRM and 1C integration

A customer calls, the manager has no idea who it is and goes digging for the record by hand. An in-house PBX fixes that directly. Asterisk, FreePBX and 3CX have modules for Bitrix24, amoCRM and HubSpot, and for 1C we use connectors we built ourselves. The customer card pops up as the phone rings, and the call recording lands in the deal record on its own.

When there's no ready-made module, the PBX can talk to pretty much anything through AMI/AGI and REST webhooks. That's how we hooked one client's PBX into their taxi dispatch software: the dispatcher sees the caller's order history, and if someone hangs up while waiting in the queue, the system calls them back automatically.

A cloud PBX integrates with exactly what the carrier has built. If your CRM or your 1C configuration isn't on that list, you wait or you improvise. Check the list before you sign.

Control and call recording

On your own PBX, recordings sit on your server, and you decide how long to keep them and who gets to listen. More useful still, recording can follow rules. For a five-star hotel we set up the dial plan so that staff-to-guest calls are recorded (for complaints and training) and calls between guest rooms never are. Whether a carrier's portal lets you set a rule like that depends on the carrier, so ask before you commit.

Legally, recording is fine as long as both parties are told, through a voice greeting, and you have a retention policy. That holds wherever the PBX runs. If the carrier stores your recordings, find out up front how long they're kept and how you export them.

What happens when something breaks

Here the cloud looks stronger, and sometimes it is. A PBX on a single box in a back room, with no backup and no second line, loses to the carrier's cloud every time. But a cloud PBX has a weak spot too: your office internet. When the uplink drops, the phones go quiet, no matter where the switch lives.

A well-built in-house PBX holds up better. We bring in SIP trunks from two carriers with automatic failover, set QoS on the routers, use codecs that cope with packet loss (Opus, G.729), and as a last resort push calls to mobiles. For a taxi company, where every minute without phones means lost orders, we added GSM gateways with SIM banks and moved the PBX onto virtualisation with a standby hypervisor and replication. Bottom line: reliability comes from how the system is built, not from where it sits.

Branches

The cloud handles branches well, since there's one switch and people connect from anywhere. An on-premise PBX copes too. It just needs secure tunnels between offices (VPN), which we also set up.

Trouble starts when every branch runs its own little PBX. At a travel agency with five branches, customers couldn't get through and nobody saw the missed calls. We replaced the lot with a single FreePBX system, two carriers' trunks with failover, integrated with Bitrix24. Missed calls went from roughly 100 a day to about 5, and inbound-call-to-deal conversion rose by 34%.

Which one to pick

A carrier's virtual PBX makes sense if:

  • the office is a handful of people and the phone is just how you stay reachable, not how you sell;
  • you have no CRM, or the carrier's ready-made integrations cover it;
  • you don't record calls, or basic recording without rules is enough;
  • nobody is around to look after a server and you'd rather not pay for that.

Your own PBX makes sense if:

  • calls are the revenue: taxis, delivery, clinics, hotels, a sales floor;
  • you need a link to 1C or a CRM the carrier doesn't support, or your own call-back logic;
  • you want recording rules: what gets recorded, what doesn't, where it's stored and for how long;
  • you have several branches and want one queue, short internal extensions and a single view of missed calls;
  • an old analogue PBX left you with dozens of extensions, and you'd like to move them with IVR and hunt groups rather than one number at a time.

Somewhere in between? That's common. Start in the cloud, and when you hit a wall on integrations or recording, move to your own PBX. You won't lose your numbers.

How the switch-over goes

The first question is always about numbers. In 95% of cases we keep them: Moldovan carriers support number porting (MNP) and SIP trunks, and we plan the cut-over ahead so the line doesn't go dead. At the hotel, the main lines were moved overnight, when call volume was lowest.

The second is about cabling. If there's no way to run new cable, phones can work over Wi-Fi. That's what the same hotel did, in guest rooms and for staff, but only after we surveyed the coverage and added access points where the signal was weak. On voice, weak Wi-Fi doesn't mean slow internet. It means words dropping out mid-sentence.

Where we come in

We don't resell telecom services. You get numbers and the SIP trunk from a carrier (Moldtelecom, Orange or an international one), the contract stays in your name, and we connect it. Our part starts there: we deploy and support Asterisk, FreePBX, 3CX or FreeSWITCH, set up IVR, queues, recording and supervision, integrate with 1C and your CRM, and build a turnkey call centre if you need one. What's covered by the retainer and what's a separate project is spelled out on our IP telephony and virtual PBX page.

Two projects in detail: a hotel that swapped its analogue PBX for SIP and Wi-Fi handsets without new cabling, and two taxi fleets running telephony, GSM gateways and 1C under one SLA.

Not sure which you need?

Tell us how many people and branches you have, which CRM you use and what's going wrong with calls today. The first meeting takes 30 minutes and costs nothing. By the end you'll get a straight answer on whether the carrier's cloud is enough or your own business phone system will pay for itself. Book a call with us.

Own PBX or carrier virtual PBX: a guide for Moldova | ITCore