Network Request Process, VIP/RIP Mapping, and Load‑Balancing Strategies

1.1 General Request Process

Modern enterprise or internet systems require traffic planning to achieve transparent multi‑level routing, involving components like client cache, DNS, transport links, CDN, load balancers, and server caches, which together enable high‑availability and high‑concurrency architectures.

1.2 Terminology

RIP : Real IP, the IP of a VM or container.

VIP : Virtual IP, cannot cross data centers, allocated by the system, used for load balancing; includes public and internal VIPs.

Internal network : Inside the data center with strict firewall policies; internal VIPs provide load balancing for inter‑application traffic.

Office network : Personal computers in office areas accessing internal applications via reverse proxy.

Public network : Internet users accessing internal applications through DNS and public VIP.

1.3 Front‑End Network Component Pairing

External access architecture: VIP → jen‑nginx (proxy for main front‑end) → static applications (1‑N).

Testing environment may involve multiple front‑end projects following the pattern: domain → VIP → JEN proxy (routing to different folders) → static application.

1.4 Back‑End Network Component Pairing

Back‑end applications can be accessed via HTTP/HTTPS or JSF private protocol, with a front‑end pre‑service forwarding to JSF micro‑services; load balancing between these layers is typically handled by the consumer.

Method 1: VIP‑based Load Balancer + Nginx

Method 2: Container (K8s) Level Load Balancing

VIP‑provided LB directly attaches machine IPs without grouping or health checks; each deployment requires manually removing machines before upgrade.

Advantages of K8s Load Balancing

Supports machine grouping by data‑center or cluster.

Provides health‑check mechanisms, eliminating the need to detach machines during releases.

2 Load‑Balancing Rules

Two rule types:

Default rule : Domain + port + cluster; one unique rule per port.

Dedicated rule : VIP + port; higher priority; inherits default rule if not set.

2.1 Default Rule Example

VIP 111.111.111 routes domain A (port 80) to backend cluster A and domain B (port 443) to backend cluster B.

2.2 Dedicated Rule Example

VIP + port can point to different backend clusters; a VIP may have multiple ports (e.g., HTTP 80/443, TCP 2000‑2014).

3 VIP‑RIP Mapping

Mapping relationships:

One‑to‑one: one data‑center, one VIP ↔ one RIP.

One‑to‑many: one data‑center, one VIP ↔ multiple RIPs.

Many‑to‑many: multiple data‑centers, each VIP ↔ multiple RIPs within that center.

4 Internal and Public Domain Resolution

Internal domain resolution uses the data‑center VIP; prefer the same‑center VIP for optimal performance.

Public domain resolution uses the public VIP; if no VIP is needed, directly use the public IP.

5 Summary

Use simple, practical component pairings to achieve transparent traffic distribution while avoiding unnecessary complexity; follow VIP‑RIP mapping conventions to limit impact scope and align with disaster‑recovery drills.