TriLoad Balancer - Application Load Balance System
 
Load Balance Technology Overview  
System Management  
 
Detail
High Scalability
 
  • Builds application cluster using, can dynamically add and remove servers
  • Single Load Balancer support many application cluster at the same time
  • High Availability
     
  • Monitors the application server status, automatically remove or add the server according their health condition
  • Alarms on the admin console when there is server down
  • Main/Standby of Load Balancer
  • Core Forward Technology
     
  • DR (Direct Routing)
  • NAT (Network Address Translation)
  • IP Tunnel (IP Tunnel)
  • Schedule Algorithm
     
  • RR (Round Robin)
  • WRR (Weighted Round Robin)
  • LC (Least Connection)
  • WLC (Weighted Least Connection)
  • Other schedule algorithms
  • Remote Admin
     
  • Configure dynamically
  • Realtime Monitor
  • Role-based access control
  • System Specification

    Hardware

    Shape / Height 1U rack-optimized
    CPU  Intel Pentium4 processor up to 2.6GHz with 533MHz front-side bus speed
    Cache  512KB L2 Cache
    MEM( Standard / Maximum) 512MB/4GB
    HD (Standard / Maximum)  80GB/160GB IDE
    Network Adapter  2 * 10/100/1000M Ethernet Card
    Power  200W
    Cooler  5

    Software

    Linux with kernel 2.4.18
    Admin Console 

    Service

    Installation and Training 
    Standard one year service  

     

    Load Balancer Forward Technologies

    Direct Routing
    How it works:
    • Client sends request
    • Load Balancer receives the request, set the MAC(physical network) address of the packets to the MAC address of the real server chosen, and forward them out
    • The destination real server will process the packets received and sends the response direct to the client (do not via Load Balancer)
    • Client receives the response

    Features:

    • Implements using data link layer technology, has lest system overhead
    • No need to modify routing rules of real servers
    • Only process and forward client request, all response packets return directly to client. It is most effective for most the applications and with minimum network latency
    • Load Balancer and real servers are at the same physical network segmant

     

    Network Address Translation
    How it works:
    • Client sends request
    • Load Balancer receives the request, change the destination address from service IP to the IP address of the real server chosen, and forward them out
    • The destination real server will process the packets received and sends the response back to Load Balancer
    • Load Balancer changes the source IP from the IP address of real server to the service IP, and sends the response back to client
    • Client receives the response

    Features:

    • Implements in IP layer
    • The listening port of real server can be different than the service port
    • Load Balancer can be at the different network segmant of real servers

     

    IP Tunnel
    How it works:
    Works almost identical to direct routing, the only difference is that this technology forwards client request to real server through the IP tunnel established between Load Balancer and real server.
    • Client sends request
    • Load Balancer receives the request, forward it to the real server chosen through IP Tunnel
    • The destination real server will process the packets received and sends the response direct to the client (do not via Load Balancer)
    • Client receives the response

    Features:

    • Implements in IP layer, has lest system overhead
    • No need to modify routing rules of real servers
    • Only process and forward client request, all response packets return directly to client. It is most effective for most the applications and with minimum network latency
    • Load Balancer and real servers can connect each other across WAN

     

    System Management

    Service Management




    Monitor


     

    Contact Us
    WatchMouse website 
monitoring