JetNexus

JetNexus Load Balancer

jetNEXUS is a range of high performance, scalable, and robust appliance and software products that can transform the performance of your network infrastructure and therefore your business.

jetNEXUS L7 Load Balancer

The jetNEXUS web accelerator range consists of three core products designed from the ground up for high performance, high availability web environments.

They provide a robust front end for your web servers offering features such as SSL offload, compression, load balancing, connection management and full W3C logging.

Implementation is fast, simple and cost effective.

• jetNEXUS Web Accelerator
• jetNEXUS L7 Load Balancer
• jetNEXUS Integrated Application Accelerator

The jetNEXUS application load balancer integrates into the common management platform to perform high performance layer 7 load balancing.

Core Platform Management Features

The Modular design provides future proofing as well as offering a sensible entry point into the enterprise web architecture.

Basic load balancer only solutions are in many cases not providing the flexibility for the applications of today let alone the next three years.

Features

The jetNEXUS Load balancer is integrated into the jetNEXUS platform; as such all the management feature such as logging, email alerting and failover etc are included (click here for a list)

• Full layer 7 load balancing
• High throughput
• Connection management
• Connection pooling
• Connection capping
• Multi Channel
• Multiple load balancing strategies including persistent cookie based
• Weighted load balancing
• Security filtering

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 for fill over 2000 requests per second. This is a through put 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. further information

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 simple 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 get requests can be terminated at the jetNEXUS appliance