EXCLUSIVE: Zeus Extensible Traffic Manager 7400

In the world of application traffic management Zeus Technology is one of the key players in this market and for good reason. Claimed as one of few software solutions on the market, Zeus' Extensible Traffic Manager (ZXTM) has always stood out thanks to its superb range of features. Originally only offered as a software product, Zeus made the move to appliance-based solutions around eighteen months ago and now provides a family of four platforms. In this exclusive review we take a closer look at its new flagship - the ZXTM 7400.

ZXTM offers an extensive range of features to web based applications and web sites including Layer 7 traffic management and high-performance load balancing along with extensive security measures. The appliance runs the latest ZXTM 4.1r2 software which includes a few minor improvements and fixes. The last significant release was version 4.1 which introduced Zeus' request rate shaping feature which, as its name suggests, allows you to set a limit on the number of request types allowed through the appliance. It's aimed at stopping problems such as service abuse, floods and dictionary based attacks.

One thing we've always liked about the Zeus solution is its easy installation and the 7400 is no exception. Just connect the dedicated management interface to your network, point a web browser at the default or DHCP assigned address and follow a brief, seven-step questionnaire. This takes you through assigning hostnames, entering IP addresses for the four public interfaces, providing DNS and gateway details, licensing the box and then moving on to the main web interface for general management and monitoring.

The interface is very easy to navigate and opens with a list of all your appliances which are referred to as traffic managers. Their icons provide an at-a-glance status readout of each one and quick access to general settings. For the latter you can view and modify network settings for all ports, set up email alerts and SNMP trap destinations, secure administrative access with different user accounts and backup the appliance's configuration.

The Services tab is where all the action takes place as this is where you define what you want the appliance to handle. Zeus employs the concept of virtual servers which are assigned to multiple physical servers known as nodes. These are grouped together in pools and the appliance uses their designated virtual server to intercept, manage and load balance their web traffic. It's worth noting that although some may see Zeus' asking price as quite steep it does include support for an unlimited number of virtual servers. In most cases you'll use one virtual server to manage each specific protocol such as HTTP and FTP and create rules that tell the appliance how this traffic should be handled.

Virtual server creation is dealt with by a wizard that steps you neatly through defining the service. It really is very simple as you just choose a name for your virtual server, pick a protocol from the drop down list and add your nodes. We successfully created pools for collections of Windows servers running general web services, FTP and web mail which were then assigned to our virtual servers. The appliance functions as an application proxy and as far as our test clients were concerned it was completely transparent as we merely pointed their web browsers at its service ports to access the required service.

Using virtual servers means you can add more physical servers to a pool as and when required without taking a service offline. A drain function also allows you to safely remove back end servers as it stops new connections associating with a specific node. Once your service is up and running the appliance performs load balancing across all nodes in the pool and offers a range of schemes to choose from. You can go for a simple round-robin mode which distributes inbound requests to each node in strict rotation or modify this behaviour by applying different weightings to each node in the pool. The least connections and fastest response modes are self explanatory whilst the perceptive mode measures system responses and distributes traffic accordingly. Alternatively, you can create a priority list which determines a strict order that each node will be used in.