|
| DevCentral > Weblogs > - Two Different Socks
|
 |
 |
 |
|
posted on Wednesday, September 23, 2009 4:15 AM
Understanding the various types of load balancing
When someone says “load balancing” the immediate reaction is usually to think of pools of servers and applications being load balanced to provide high-availability for massive sites like Amazon or Google or Facebook. But there’s a couple of other types of load balancing that deserve to be recognized because although they sit in the shadow of “load balancing” they are often invaluable assets to network and application architects attempting to ensure availability and adherence to service level agreements.
 Link Load Balancing
Link load balancing is the act of distributing network traffic over at least two network (ISP) links. Link load balancing is used to provide availability through failover in the event one network link becomes unavailable or to optimize delivery of data by intelligently choosing the ISP link with the best performance characteristics (latency, congestion, etc…) for the application in question. Link Load Balancing generally operates at Layer 2 through Layer 3.
Global Load Balancing
Global Server Load Balancing (GSLB) is used to distribute application requests across multiple data centers (cloud or traditional) to assure availability and meet performance requirements. Disaster Recovery (DR) scenarios often include the use of GSLB to redirect users to a secondary (or tertiary) data center in the event of a disaster, but GSLB is also increasingly used to manage performance through the use of GeoLocation technologies and more advanced evaluation of business information involving power consumption and costs and other business-related metrics. GSLB can operate at Layer 4 through Layer 7.
Local Load Balancing
Local Server Load Balancing – often referred to as simply Server Load Balancing (SLB) despite its increasing inaccuracy in the case of cloud computing and virtualized architectures – is used in a local environment to direct application requests to the appropriate application instance based on configured performance and availability metrics, among other variables. The most sophisticated of the load balancing types, local load balancing offers the most flexibility in configuration options and in the variables that can be used to determine where to direct requests. SLB solutions can operate at Layer 4 through Layer 7.
Firewall Load Balancing
Firewall load balancing is a specific implementation of local load balancing used to load balancing multiple firewalls to provide scale, resiliency, and performance. Firewall load balancing has unique requirements that are not always met by SLB even though the core functionality for this type of load balancing is almost always provided by an SLB device. Firewall load balancing solutions require extremely high performance even under massive load, and generally operate at layer 4 (TCP) or below, though some provide options at higher layers of the stack as well.
         
Related blogs & articles:
|
|
|
|
|
|
|
|
|
