We have a network that was set up 7 years ago. Since that time we have performed many changes to that original (star) topology. For instance, we have run additional CAT 5 cables to connect additional PCs and peripherals. We have replaced hubs with switches. We have also migrated the PC servers to blade servers on our new IBM Iseries. What we do not know at this point is the effect of all these changes on network performance.
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
We see on the Internet that numerous software programs are available to monitor services. However, we seek specifically to obtain a tool that will tell us if our router, switches, hubs, cabling, etc., are fully functional and efficient so that their configuration does not produce bottlenecks, excessive latency and other general throughput issues.
Can you recommend a product that diagnoses/analyses hardware configuration?
My unqualified answer here is AppareNet. Read no further if you have an aversion to sales pitches. But it's "just the facts from" my research perspective.
On-demand network performance analysis is one of the primary functions we designed AppareNet for. The situation you describe (creeping changes with potential unforeseen behaviors) is ideal for that particular capability. You don't necessarily want to instrument your network from a device-centric standpoint - you simply want to confirm that network performance is what you expect it to be, and on an on-going basis, without incurring significant overhead in making the assessment. Basically point-and-shoot on demand when needed (on-going monitoring using the same proprietary approach is also available).
The other key design intention is troubleshooting and diagnosis. The expert system uniquely identifies type of performance issue and location (at Layer 3) if present. In many cases, configuration issues typical of the network you describe (ex. duplex conflict) can be discovered without querying any devices directly. This allows you to focus your attention on suspect interfaces, cabling, and configurations.
Another significant advantage is the analysis of NIC/driver performance. After duplex conflict, poor driver performance is your number one performance issue. NICs can appear to be performing but at seriously degraded levels. Upgrading to the most recent versions is not always an assurance of best performance either – you are simply trusting the manufacturer to know better. Recent problems with certain GigE drivers underline the fact that everyone is simply guessing until they actually do performance analysis in the real world.
If our technology is not the answer to your question, the alternative is the classic (what I refer to as "old world") network management systems. They are typically expensive, either in capital costs (Nagios is free) or in maintenance/configuration overhead. And they don't really offer network performance analysis as such. They are often strong on the device configuration management though. If you are interested, check out an article in the June 2003 hardcopy issue of Network Computing called "Network Management on $1.19 a Day." They name Ipswitch's WhatsUp Gold as Best Value and I tend to agree relative to the criteria I laid out.
Dig Deeper on Real-Time Performance Monitoring and Management
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.