Overview | Denial of Service | Interface

Feature List

• Web-based Management Interface
• Primary/Failover server capabilities allows a backup server to take over if a server is to crash
  • Servers can be configured with a default role. This means when a crashed server recovers, the server that took it's place can be configured to stay active or go back to standby.
• Load Balancer redundancy
  • With 2 load balancers one can be configured as standby and will only become active if the primary load balancer is to crash
• All services and servers are ordered, color coordinated, and identified with names as well as IP's
• Servers can be taken down and brought up within a service, without the compromising the entire service.
  • Persistence, weights, and scheduling may be configured without any service downtime.
• Any TCP/IP-based service can be load balanced (http,https,ftp,mail,www,etc..)
  • Tested single port services
    • httpd
    • ftp (not passive)
    • DNS
    • smtp
    • telnet
    • netstat
    • finger
    • proxy
    • nntp
    • spinbox
• All changes are made directly to the routing table, applying all changes instantly
• Scheduling Algorithms
  • Round Robin
    • Connections are distributed evenly among the servers in the pool
  • Weighted Round Robin
    • Connections are distributed according to configurable server weight in the pool of servers. This is extremely useful if the servers in the pool have different processing capacities.
  • Least Connections
    • Connections are distributed to the server with the least number of established connections in the pool of servers.
  • Weighted Least Connections
    • Connections are distributed in a similar fashion as the least connection scheduler, but servers in the pool have a performance weight which allows different servers to handle different amounts of established connections
  • Locality Based Least Connections.
• All scheduling algorithms may be configured to be persistent to allow secure services which require communication with the same server for authentication.
• Custom monitoring software notifies via email/pager of any possible problems, including disk and memory usage, number of processes, DoS,DDoS, and SYN attack alerts, and connections.
• Logging of all load balancer user activity in Common Log Format
• Server comes configured for VS-NAT with the option of using VS-TUN or VS-DR depending on your application
• If your setup consists of a primary and failover load balancer, all changes are automatically repliacted between the two systems.

Different Load Balancing Implementations

	VS-NAT	VS-TUN	VS-DR
Servers That May Be Used	Any IP:Port TCP/IP-based service	Server must support IP tunneling	Non-arp device
Server Network	Private	LAN/WAN	LAN
Max Virtual Servers	Unlimited
Max Real Servers	Around 25 Real Servers with throughput of 400k	Scalable	Scalable
Total Bandwidth	10 mbps	1 gbps	1 gbps
Server Gateway	Load Balancer	Router	Router
Advantages	Only one public IP is needed for each virtual service. Real servers can run any operating system that supports TCP/IP. The real servers run on a protected network	Servers can be geographically remote or on another network. Higher throughput than NAT (no rewriting of packets, each server has own route to client). load balancer only schedules. For http, requests are small (GET /index.html), can direct 100's of servers. Total server throughput of Gbps, director only 100Mbps	All of the advantages of the tunnelling method but doesn't need tunneling servers
Disadvantages	Due to the fact that the load balancer has to redirect packets from the realservers through the load balancer, Traffic limited to approx 9mbit.	Server must tunnel and not ARP	Servers are on the same network as the load balancer. lo device must not respond to ARP