Apium Cloud OS Control Panel — Dashboard, Cascade, Tunnel Services, Active VPN Tunnels & Real-time Load
Apium Cloud OS Control Panel (26.11): Corporate Gateway switcher, Cascade routing, tunnel service status (MS-SSTP, L2TP/IPsec, strongSwan, sing-box), and live session telemetry.
Cascade Modes
4 Live Modes
Health Probes
Auto-Failover (15s)
Platform Isolation
Cascade vs Corporate
Kernel Controls
HTB/TBF + nftables

1. Two Completely Isolated Platform Operating Modes

Unlike consumer panels that mix all users and rules in a single database table, Apium Cloud OS cleanly separates two architectural operating modes on the node:

  • Cascade VPN Mode: Designed for high-throughput multi-protocol ingress (VLESS Reality, Hysteria 2, Windows SSTP, L2TP/IPsec) -> kernel traffic shaping (HTB/TBF) & group nftables profiles (Basic, Restricted, Paranoid) -> multi-hop Upstream Relay balancing over Hysteria 2 and VLESS.
  • Corporate Gateway & Site-to-Site Mode: Uses an independent database, isolated user/role/quota/audit tables, dedicated sing-box, strongSwan, SSTP/L2TP, and nftables configs, X.509 CA certificates, AD/RADIUS (AAA) authentication, corporate subnet IP allocation (e.g. 10.20.0.0/24), segment ACLs (DEV/ERP/DB), and SIEM / Prometheus event export.

2. Four Cascade Routing Modes (Without Dropping Active Sessions)

Each user group or subscription tier can run its own routing policy independently, and switching policies takes effect without restarting active client connections:

Apium Cloud OS Routing Modes

Cascade Engine

How traffic flows from the entry gateway to exit nodes across different workload requirements.

ModeMechanismKey GuaranteePrimary Use Case
Direct (Zero Hop)Local NAT / MASQUERADE exit directly from the entry nodeMinimum possible latencySingle-region gateways & direct low-latency links
Auto-Failover (15s)Continuous RTT & reachability probing every 15 seconds99.99% uptime; instant switch to backup nodeMission-critical channels where ISP/upstream drops occur
Load BalanceEven or Priority-weighted session distribution across exit nodesHorizontal bandwidth & CPU scalingHigh-concurrency fleets & multi-exit redundancy
Selected NodeHard geo-pinning of a group or tier to a specific exit serverPredictable dedicated exit IPBanking, payment gateways, streaming & IP allowlists

3. Split Tunneling: Proxy Clients vs Native OS & Router Tunnels

Split tunneling in Apium Cloud OS works at two distinct layers depending on the client type:

  • App-Level Split Tunneling (VLESS Reality / Hysteria 2): Domain, geosite, geoip, and CIDR-based routing rules send local/trusted traffic direct while encapsulating only target destinations or cascading sensitive flows through Upstream Relays.
  • L3 / Subnet Split Tunneling (Windows SSTP & Router L2TP/IPsec): In Corporate Gateway mode, remote employees and branch routers (MikroTik, Keenetic) receive targeted corporate subnet routes (10.20.0.0/24) and static IPs—keeping general internet traffic off the corporate hub while enforcing strict segment ACLs for internal DEV/ERP/DB access.
  • Unified Accounting & Kernel QoS: Regardless of whether a user connects via VLESS, Hysteria 2, SSTP, or L2TP, all bytes aggregate into a single user profile with automatic session termination on quota/expiry and per-group HTB/TBF bandwidth caps.

4. How This Differs From 3X-UI, Marzban, Hiddify, Remnawave & S-UI

Cascading & Routing: Apium Cloud OS vs Other Platforms

Deep Comparison

Why raw JSON outbounds in proxy panels do not equal an enterprise cascade & split-tunneling architecture.

Capability3X-UI / Marzban / RemnawaveHiddify / S-UI / MarzneshinApium Cloud OS
Multi-hop Upstream Cascade◐ Manual Xray JSON outbound◐ Manual sing-box / node list✓ Native Hy2/VLESS Relay UI + API
15s Auto-Failover Probing✕ No built-in RTT failover✕ Client-side / manual✓ Server-side 15s Health Failover
Per-Group Live Mode Switch✕ Core reload / global config✕ Core reload required✓ Direct / Failover / LB / Geo Pin live
Corporate Mode Isolation✕ Single shared DB & config✕ No corporate mode✓ Separate DB, CA, RADIUS & ACLs
Kernel Shaper & Firewall✕ Bytes quota only✕ Bytes quota only✓ HTB/TBF Mbps + 3-tier nftables

What Other Panels Require Manually

  • Hand-writing Xray/sing-box outbound routing JSONs and restarting cores on every edit.
  • Writing custom cron/bash scripts to detect dead upstream VPS nodes and rewrite configs.
  • Manually configuring strongSwan, accel-ppp, RADIUS, and nftables for routers and corporate subnets.

What Apium Cloud OS Delivers Out of the Box

  • Hitless per-group switching between Direct, Auto-Failover (15s), Load Balance (Even/Priority), and Selected Node.
  • Complete isolation between Cascade VPN and Corporate Gateway (separate DB, CA, ACLs, SIEM/Prometheus).
  • Unified traffic accounting, HTB/TBF shaping, and split tunneling across all 4 protocols.

5. Architecture Diagram: Cascade & Failover

The diagram below illustrates how client devices, the Apium Cloud OS entry server, and distributed exit nodes across multiple countries interact in real time:

How It Works

  • 1. Single Entry Point for All Clients: Windows PCs (native SSTP), hardware routers (MikroTik/Keenetic via L2TP/IPsec), and mobile/desktop apps (VLESS Reality / Hysteria 2) connect to one stable Apium Cloud OS entry server IP/domain.
  • 2. Kernel Enforcement & Group Policy: Before forwarding any packet, the entry server applies group nftables rules, tc HTB/TBF bandwidth caps, and quota accounting.
  • 3. Continuous 15s Probing & Mode Execution: Every 15 seconds, Apium probes RTT and reachability of all upstream country nodes (e.g., NL, DE, FI). In Auto-Failover, if NL degrades or goes down, traffic seamlessly switches to DE or FI. In Load Balance, sessions distribute across nodes (Even or Priority). In Selected Node, a group stays pinned to a specific country, while Direct exits locally via NAT.

Key Advantages

  • Zero client reconfiguration: Changing exit countries or recovering from a dead foreign VPS happens transparently on the server without touching client devices or routers.
  • Protocol upgrading across hops: Even legacy L2TP/IPsec routers or clientless Windows SSTP sessions get encapsulated over encrypted Hysteria 2 / VLESS upstream links between countries.
  • Hitless per-group isolation: Different departments or subscription tiers on the same Apium server can simultaneously use different countries and failover modes without dropping active sessions.