Full Layer 7 Load Balancing
Layer 7 load balancing is load balancing at the application protocol level. The benefit of this is that the device understands the type of data it is processing, in this case HTTP and HTTPS. This allows for many features that are not possible at lower level load balancing such as URL switching, Cookie based persistence, Request scanning etc
High Throughput
jetNEXUS can maintain 15,000 concurrent connections and fulfill over 2000 requests per second. This is a throughput of about 120meg +. Many Layer 4 devices claim that they handle over 4 million concurrent connections. This is because they simply rewrite the packets for direct return. In a real environment 4 million concurrent connections is not realistic unless you are load balancing over 400 performent web servers in which case the throughput would most likely be well over 2 gig !! (Assuming each server producing 5 meg). Many layer 4 devices also quote their throughput metrics based on pages that are 2k or sometimes less in size. We do not believe this is very realistic as such jetNEXUS metrics are quoted with a page size of 64k.
Connection Management
Connection management is an advanced feature that can optimize the performance of TCP/IP as well as manage the level of stress placed on the web servers. By maintaining two sets of connections web servers are abstracted from slow end user dialogue.
Connection Pooling
The conventional behaviour of a proxy request is that a connection from one TCP/IP device to another is opened, the data is passed and the connection is terminated. To open a connection involves a handshake every time that can add additional unnecessary overhead. jetNEXUS is capable of opening a set of connections to some web servers and reusing them thus reducing the effects of TCP/IP slow start.
Connection Capping
jetNEXUS will allow you to protect your web environment from overload by limiting the number of incoming connections. This can help to stop the embarrassing time out error when the environment is under too much stress. Once the limit is reached jetNEXUS can respond with a pre-prepared page such as 'Please come back later' rather than the user having to wait for a minute only to experience a time out error.
Multi Channel
jetNEXUS Accelerator will support multiple virtual acceleration engines on the same physical hardware.
Multiple Load Balancing Strategies:
Round Robin
Round Robin
is the most basic type of load balancing strategy. All web servers simply take turns with the incoming requests.
Least Number of Connections
The least number of connections is the most popular non persistent load balancing method. This method takes into account the servers current work load and will allocate new requests to the least busy server.
IP Sticky - Persistent
IP sticky is the most basic persistent load balancing technique. This method will simply direct the same user back to the same web server based on their IP address.
Cookie Sticky - Persistent
This is the most advanced and popular method of persistent load balancing. The client is issued a cookie on the first connection with the internal server ID. Subsequent requests will read the cookie and redirect the request to the appropriate web server.
Weighted Load Balancing
Weighted load balancing takes into account the fact that some servers have more capacity than others. The weighted option allows you to define the relative performance of particular web servers. This weight will influence the number of request the servers is sent to process.
Security Filtering
Layer 7 load balancers maintain two separate sets of connections. One set to the end user and the other to the web servers. This allows a great deal of control over the type and volume of traffic sent to the back to the web servers. In addition to this invalid requests can be terminated at the jetNEXUS appliance. | 
