If you have not read the articles about our NetRule product reviewed by NetworkWorld, you can link to them here: NetworkWorldFusion (“Network-modeling Tools”, October 18, 2004).
The review made a number of good points, including that NetRule is an easy-to-use, competent tool that can accurately predict the impact of many network topology changes. We offer the following information to help you interpret some of the other points put forth in the review.
Run Time: There is a very significant difference between the run times of the tools when used to evaluate any latency or delay, and this was not reported in the review. In particular, NetRule can evaluate a large enterprise network in just seconds or a few minutes in the worst cases, while ITGuru can take many hours of run time to model even one hour of traffic for a large enterprise network. This is because NetRule uses mathematical formulas to calculate results in full detail for all of the traffic, while ITGuru uses discrete-event simulation to play out every event as it accumulates all of its latency and delay results. (This is true even if using ITGuru's "hybrid" feature to ignore some traffic). This is important because a fast run time lets you model the full traffic levels, explore more alternatives, and complete your studies in less time. For example, one of our customers recently evaluated a model with 3 billion packets/hour originating from 8,500 users and traversing 27,000 links, and the evaluation took under 4 minutes. The companion project using OPNET required more staff and time and still chose to scale back the problem in order to handle it.
Modeling Enterprise Networks: The review suggested that NetRule "lacks some of the features and capabilities required for modeling enterprise networks," and "does a decent job modeling smaller-sized networks." In fact, NetRule's focus on fast performance and streamlined detail shines best when applied to large enterprise networks, and our user's have built a number of large, enterprise models, some with over 100,000 users.
Data Import: The reviewers never tested NetRule's features for data import, yet suggested that they were "labor intensive." In fact, preparing a csv file for import into NetRule involves simple spreadsheet activities such as deleting or moving columns. The reviewers did invest the time to build the full model in ITGuru, and found that “creating manual configurations for the Foundry gear was a painful and time-consuming process." NetRule users routinely import large amounts of data and build full enterprise models. We can only suggest that the reviewers might have found a more pleasant and timely experience had they tried NetRule for the full model. In addition, our next release will import router configuration files.
Configuration Errors: The review said that "The tool does not check for configuration errors or device inconsistency, nor does it check for IP address conflicts or protocol errors. It seemed that the model assumes that users do not introduce input errors. That is a large, and probably unreasonable, assumption." The review has mixed up two separate issues: performance modeling versus configuration management. In fact, NetRule is based on a supremely reasonable tradeoff, which is to support fast building of logically-consistent models while leaving the physical router configurations or IP address spaces to the configuration management tools, which can do that job better than any of the modeling tools. Questions of utilization, throughput, delay, and cost, which NetRule handles, are quite distinct from questions of whether a given router actually has an open port for a T3 line. NetRule's focus on performance issues and the simplification of configuration detail is a major reason that NetRule users can get results in days that would take weeks to configure, debug, and simulate under our competitor's excruciating approach.
Rating Scores: NetworkWorldFusion admittedly uses subjective scores for rating, and does not tie them to objective data. For example, the review did not report any comparative run times or variance between predicted and real utilization or latency, yet gave scores for "performance" and "accuracy.” Conversely, the review reported that NetRule installed in the least time and space, yet rated NetRule below the others for Installation. Such rankings are best taken with a grain of salt.
Conclusion: The review said that ITGuru is the best choice for enterprise network modeling, although it did allow that NetRule might be ok for small networks or LANs. However, NetRule can handle, and has handled, many huge enterprise network models, and has done so more efficiently and economically due to its advantages in simplicity, speed, and cost. Just yesterday an aerospace customer mentioned in his email, "The value of NetRule [vs. OPNET] to our project is the ability to implement the true network topology with full-scale end-to-end traffic modeling where large complex models can be constructed in a week and run in minutes." Simple logic then suggests that NetRule is the best choice for those models it can handle, since otherwise you must invest more time and money in ITGuru.