Q-Balancer
The Q-Balancer or (QB) provides redundant connectivity by balancing multiple Internet and private links from various providers. This technology is also known as a WAN load balancer.
Q-Balancer is built around our DeanSoft Designed OS which has proven to be very powerful and robust. The hardware platform for our products has no mechanical parts, which reduces the risk of hardware failures. Our product's entire configuration resides in Flash memory on the device to enhance robustness and performance. Physical redundancy is available with rack mount models.
Link Load Balancer vs router
A Q-Balancer permanently monitors its links to verify their availability by polling far into the Internet. QB keeps real time metrics like the response time trip (RTT) and bandwidth usage of each link. Those metrics are used for algorithm decision making and to verify logical routing errors on the ISP network. The QB can support many links in parallel to their maximum usage for optimizing available bandwidth to users.
For the services accessed from the outside and for remote users (VPN, SMTP, WEB SERVER, terminal server, etc.), Q-Balancer Load Balancers add intelligence to the DNS service to handle DNS requests and give the selected link IP address for optimized performances and 100% availability.
Q-Balancer algorithms are very granular and can be applied on all combination of source IP/destination IP/protocol/port and link ensuring optimal control of all traffic flows.
Installation
 |
The physical installation of the Q-Balancer is the simplest on the market. All you need to do is plug in two Ethernet cables and traffic will automatically start to flow through the unit. Afterwards, all that will be needed is to create balancing rules so you can tailor the load balancing to your needs. The Q-Balancer is completely independent from the other systems, as well as Internet service providers. You do not need any special configuration in your ISP routers. The QB does not use BGP (Border Gateway Protocol) and does not require an ASN (Autonomous System Numbers) either. Since it operates at the data link layer (layer 2) of the OSI model, no network or sub network changes are required. |
Primary Link (inline operation)
|
When installing a Q-Balancer, an organization usually already has infrastructure in place and one telecommunication link. This first link is called the primary link. The firewall and all servers are configured with the IP addresses of the primary link. The strength of the Q-Balancer QB is that it keeps the primary link concept and manages to load balance traffic on all links transparently. This means the firewall and servers never know they are being balanced. No reconfiguration or IP address migration is required. The QB must be perceived as a "black box" that is transparent to the rest of your network and manages all links according to advanced algorithms. |
High Availability mode
Robust and flexible, the Q-Balancer integrates real time High availability, allowing the addition of a second load balancer in a fault tolerant configuration. The Link Load Balancer's system and configuration reside in FLASH memory on the device, thereby reducing the risk of failure. |
Configuration
Besides being able to adapt to any type of link, the Q-Balancer uses advanced load balancing algorithms on different protocols to manage traffic on multihomed networks. Algorithms can be used both for incoming and outgoing traffic. Another feature of the Q-Balancer is its granularity. Algorithms for load balancing can be applied to entire protocols or specific to source IP/ports. Balancing rules can use different algorithms:
- Fast ISP response time.
- Redirect to transparent proxy
- Weighted least traffic by packet
- Weighted round robin IP persistent
- Weighted round robin by connection
- Weighted round robin by packet
Inbound requests can be distributed to multiple servers. The Q-Balancer supports persistent sessions (or persistence) under a variety of protocols. This is required for compatibility reasons. The most common protocols are FTP, IPsec and PPTP.
Traffic segmentation
Quality of Service with a single link requires the use of expensive devices with varying results. Multiple links preserves redundancy and give more option to the network administrator. For the most efficient Quality Of Service, DeanSoft recommends traffic segmentation.
A feature of the Q-Balancer is its granularity. Algorithms for load balancing can be applied to entire protocols or specific to source IP/ports. Traffic segmentation approach is to separate various traffic types on different links. This way, corporate services (like email, EDI, web hosting, VoIP) can be on the primary link during normal operation. Bandwidth intensive protocols are on the second or third link and will not saturate corporate services. In case of a link failure, IP services are merged on remaining links.
A similar approach is recommended for VPN remote users. With home high speed internet, a few remote workers can easily saturate a corporate link. Traffic segmentation can be done for both incoming and outgoing traffic.
Another important feature for Quality of Service is Band Width Management. This function allows the restriction of bandwidth for file transfers (for example FTP and HTTP downloads).
Finally, Q-Balancer can be used in conjunction with specialized QoS devices like Packeteer for greater management and control on IP traffic, priority, and link selection.