Uptime and fast problem resolution are top priorities in enterprise networks. But purchasing dedicated monitoring devices for hundreds of network points is impractical, but you can use a matrix switch to do the same thing.
Uptime and fast problem resolution are top priorities in enterprise networks. But purchasing dedicated monitoring devices for hundreds of network points is impractical, but you can use a matrix switch to do the same thing.
This technology lets network engineers electronically move and share expensive testing and monitoring equipment across a network. Using drag-and-drop functionality in a Web-based GUI, they can connect any device on the network to another without the need for a manual cable-pull.
That compares with the conventional approach in which each network analyzer (or network probe, sniffer, intrusion detection device) is connected one-to-one to a dedicated Layer 2 switch. With a matrix switch, you get the same instant access but with fewer devices, resulting in cost and labor savings.
The solution scales to support thousands of ports and devices, so it’s applicable in even the most expansive and complex networks.
Consider a major financial institution that developed a plan for upgrading its network infrastructure to take advantage of the services supported by a 10 Gigabit core. The plan called for the distribution of eight packet analyzers across the network, with one installed at each Layer 2 switch. System engineers specified three additional analyzers for assurance that all critical aspects of packet analysis were addressed.
Total expense for this these devices was nearly $800,000. Unfortunately, the budget was $400,000.
A workaround strategy featuring fewer analyzers met the budget but required that each device be moved by hand to the network location needed. This meant that when an issue surfaced, engineers would need to physically move an analyzer to the suspected trouble spot, increasing mean-time-to-resolution. There was also the added vulnerability of not having the entire network covered — meaning potential issues could not be dealt with before they became major problems.
A new strategy, centered around a Layer 1 matrix switch, provided a more agile monitoring capability, with complete network visibility — and all at a cost that was under budget. The solution included three packet analyzers cabled to two matrix switches, which featured a GUI accessible via a Web browser. Through the GUI, engineers could observe the entire network and electronically connect the analyzers to points where they were needed, thus ensuring fully optimized network performance.
Equipment, annual warranty and maintenance costs were 77% lower than the original plan. This brought the total cost to less than $200,000, while giving the institution an improved monitoring solution that enabled IT managers to meet stringent network performance requirements.
In another example, a manufacturer of Internet network equipment was looking for a way for its test labs to reduce hit-and-miss trials during system configuration.
One department, operating an automated regression test lab, faced the issue of how to share traffic generators among multiple devices under test in 80 test beds. To increase efficiency and scale, a pool of test beds was created in which a single traffic generator is paired with a matrix switch that supports four connected devices simultaneously. Engineers also created a testing script that automatically configures the matrix switch, creates a pre-set connection topology and recalls each required preset topology for the next test.
Not only can test setups be quickly created and reconfigured through scripting, but traffic generator signals can be broadcast to any number of other ports with no signal degradation. The increased efficiency in equipment utilization means fewer traffic generators are required to test the same number of devices, which resulted in a cost savings of approximately $500,000.
When planning test lab automation and enterprise network monitoring systems, companies should look for:
* Premier matrix switches that provide a user-friendly Web interface. With this GUI, users can manage all switch functions intuitively, in familiar two-click or drag-and-drop fashion. This includes making port patches in both real-time and in batches, setting port properties and presetting configurations to speed the testing process.
* Embedded user interfaces that eliminate the need to install client software, which is prevented in many networks for security reasons.
* Capabilities such as zoning, the ability to restrict user access to a set of ports or entire switches; and port locking, a security feature that lets authorized users lock one or more ports or devices.
With centralized management of device connections, the need to physically re-cable a switch is eliminated. Although the time saved might appear insignificant, imagine the productivity gains over a year if it can all be accomplished directly from one’s desktop.
Similarly, a matrix switch can reduce data center energy costs. Consider the situation of the financial institution profiled above. If it had proceeded with its original strategy to deploy 11 packet analyzers, the heat load on the data center would have increased, as would the cost of cooling.
Implementing the revised plan based on three packet analyzers and a matrix switch lowers the energy costs significantly.
Further, lower heat loads translate into less heat-related wear and tear on equipment, which extends equipment life, cuts down on maintenance repair costs, and also avoids heat-related equipment issues that can impact network performance or even availability.
Holmes is director of sales at Apcon. He can be reached at markh@apcon.com.
