LVS Manager

Deploy a horizontally scaled cluster of high-capacity SRT gateways with LVS Manager to handle thousands of concurrent connections, ensuring consistent stream quality and minimizing the risk of outages or service interruptions during maintenance.

User Story: As a broadcast network engineer, I need to deliver live video streams from a host broadcaster to a large number of clients worldwide using the SRT protocol — reliably, securely, and with fault-tolerance.

What We Do: We deploy a horizontally scaled cluster of SRT gateways behind a Linux Virtual Server (LVS) Load Balancer. LVS Manager orchestrates connection distribution via port-level scheduling and monitors backend health.

Why We Do It: Traditional stream routing methods lack the dynamic balancing and fault recovery needed for modern video delivery. LVS Manager with VRRP failover and automatic sync ensures always-on availability without manual intervention.

Result: The system handles thousands of SRT connections, maintains stable ingress during gateway or node failure, and allows real-time backend changes without stream disruption.

Why LVS Manager: It offers a dedicated, low-level UDP balancer with web-based config, per-port strategy control, auto-failover, and real-time stats — all tailored for persistent SRT sessions.

Maintain seamless streaming sessions over HLS, MPEG-DASH, and RTMP by dynamically routing client traffic to healthy backend nodes with zero disruption during infrastructure scaling or failures.

User Story: As an IPTV or OTT platform operator, I need to deliver live or on-demand video to thousands of users using TCP-based protocols like HLS, DASH, or RTMP. My goal is to maintain session stability and avoid stream disruptions during backend upgrades or outages.

What We Do: LVS Manager manages redundant LVS Load Balancer nodes with VRRP, routing TCP traffic across multiple edge or origin servers. Each client session is routed through a Virtual IP and distributed using session-aware strategies like Source Hashing to ensure backend stickiness.

Why We Do It: TCP streaming protocols are connection-sensitive. Reassigning sessions mid-stream leads to playback interruptions. LVS Manager applies per-port scheduling and failover rules to protect session continuity while allowing live backend updates.

Result: Viewers experience uninterrupted playback, backend nodes can be updated live, and the system dynamically recovers from node failures without stream loss.

Why LVS Manager: It delivers low-level control over LVS with TCP-aware scheduling (like source hashing) with multiple endpoint handling, web-based configuration, automatic sync, and real-time diagnostics — built specifically for persistent TCP streaming sessions.

Aggregate multiple incoming video feeds into centralized ingest points while maintaining control over backend availability and load distribution.

User Story: As a live‑production or ingest‑hub engineer, I must accept dozens (sometimes hundreds) of concurrent WebRTC/RTMP/SRT/RIST contribution feeds from remote encoders, OB vans, and cloud sources. Contributors should hit a single IP address or hostname while the platform automatically spreads load and survives node failures or gateway overloads.

What We Do: We place a horizontally‑scaled cluster of SRT gateways, RIST receivers, and WebRTC servers behind an LVS virtual IP managed by LVS Manager. Port‑range rules and the source‑hash scheduler keep each encoder sticky to one gateway, while health probes and weighted scheduling direct new feeds away from busy or draining nodes.

Why We Do It: Contribution feeds are long‑lived; re‑homing mid‑stream causes packet loss and re‑sync delays. DNS tricks or L7 proxies add latency or cannot track thousands of UDP flows. Kernel‑level LVS provides microsecond switching, and LVS Manager adds the visibility and automation missing from raw ipvsadm.

Result: Ingest hubs scale on demand, accept many simultaneous sources with uniform utilization, no encoder reconfiguration, and hit‑less maintenance windows. Failover of either a gateway or the balancer VIP is transparent to all contributors.

Why LVS Manager: Purpose‑built for SRT/UDP: port‑specific algorithms, weight‑by‑health API, automatic VRRP sync, web UI for live draining, and lightweight kernel forwarding—delivering high throughput with minimal latency and CPU.

Automate reliable, low-latency distribution of live sports streams by balancing persistent SRT connections across multiple edge gateways, ensuring smooth failover and uninterrupted viewer experience during high-traffic events.

User Story: As a live‑sports broadcaster, I receive game feeds from OB vans inside the stadium LAN as well as remote regional encoders and cameras. I want every affiliate or CDN edge to pull the same SRT endpoint while ensuring each feed is routed to an available gateway with zero buffering—especially when viewership spikes at kickoff or overtime.

What We Do: LVS Manager fronts multiple SRT gateways with one virtual IP. Using its configuration scheduling feature, we pre‑load a “Match‑Day” rule set that activates 30 minutes before the event: ports 10000–10100 switch to wlc (weighted least connections) for burst intake, while ports 10101–10150 stay on sh (source hash) to keep remote cameras sticky to their local gateway.

Why We Do It: LAN cameras need stickiness for commentary sync, but remote regional feeds arrive unpredictably and must be balanced evenly. Time‑based rule rotation lets us automate that shift without human intervention and guarantees capacity when the crowd joins.

Result: Thousands of concurrent SRT connections are accepted through a single URL. Local sources stay locked to their gateway; remote feeds spread across the cluster; viewers experience uninterrupted, low‑latency playback even if a gateway fails mid‑match.

Why LVS Manager: Only LVS Manager offers port‑range + time‑scheduled algorithms, VRRP VIP failover, and real‑time health metrics in one UI—so engineering can configure “pre‑game,” “live,” and “post‑game” profiles once and trust the balancer to switch modes automatically.