Bulwark Sizing Planner
Tell us your deployment and get a recommended starting configuration — CPU, RAM, storage, network and server layout — for Bulwark Voice Suite. Transcoding and video are the biggest drivers, so those matter most.
Users / extensions 250
Usage profile sets busy-hour concurrency
Media handling transcoding = the main CPU cost
Concurrent video participants
Override concurrent calls (blank = auto)
IVR / auto-attendant flows
ACD queues
Recorded minutes / day
Recording retention (days)
Features
Enterprise
2 × active-active
Design concurrent calls
88
Peak bandwidth
65 Mbps
Per server
⚙ CPU
8 coresXeon Silver / EPYC
▤ Memory
16 GB
▦ OS / app disk
60 GB SSD
⇅ Network
1 GbE ×2
Shared recording storage
⛁ Recordings
180 GB90 days retention
How these numbers are worked out
These are engineering estimates for a recommended starting configuration, not a guarantee —
real capacity depends on your codecs, call patterns and hardware and should be confirmed with a benchmark
on the target servers.
Key assumptions: busy-hour concurrency of 20% (office) / 35% (mixed) / 60% (contact centre) of extensions;
roughly 40 transcoded calls per CPU core and far more for passthrough/relayed media (then
network-bound); video at ~2.5 Mbps per HD stream on the SFU; ~1 MB per recorded minute;
10 GbE once sustained traffic passes ~700 Mbps. High availability is active-active across up to 4 nodes,
with each node sized to carry the load if one fails.