By Mehdi Daoudi
While just understanding the value of website performance monitoring is an asset in and of itself, there remains confusion within the IT world as to the benefits and drawbacks that result from active and passive monitoring strategies.
Despite what many believe, the issue is not whether Web professionals should be utilizing either approach, but rather how the two can be used in congruence to provide an efficient and accurate look at a sites' performance. So, what's the difference between active and passive monitoring?
In order to understand how these very different approaches to monitoring work together, it is first necessary to understand exactly what they are and what they cover.
Active (or synthetic) monitoring takes a proactive approach to ensuring a site is working at optimum efficiency (e.g. availability, download speed) from specific geographies and Internet service providers (ISPs). Maintaining an active monitoring strategy requires software-based agents (backbone, last mile, wireless and private nodes) distributed throughout the world in a controlled setting to simulate the user experience. This method measures and validates key business procedures and functions in a website (shopping carts, customer relationship management record retrieval, Web lead registrations, conversion goals, etc.).
Since active monitoring happens continuously and on a fixed schedule, site owners can create a controlled baseline for performance, which allows for alerts to let them know about availability and performance at the first sign of trouble, even when no users are on the site. This enables them to diagnose and troubleshoot problems before they affect the end-user experience, minimizing the negative impact on their enterprises' bottom line.
In addition to early warnings, synthetic traffic generated from backbone ISPs eliminates the effect of variables outside of one's control, such as browser extensions, badly configured home connections, etc. This allows active monitoring to also be used for service-level agreement (SLA) management and verification, as well as benchmarking a site's performance against the competition.
Hosting Got You down?
Check out news and tips in Website Magazine's Web Hosting Panel on the 'Net at wsm.co/hosting.
Performance data collected is impacted not only by an infrastructure, third parties and content delivery networks (CDNs), but also determines if code changes or architectural/infrastructure changes had the desired effects, or if they caused errors and/or performance degradations. This helps a company identify and optimize for these variables, which allows workers to test changes before rolling out any to end-users.
These advantages are all well and good, of course, but that doesn't mean that active monitoring is without drawbacks. Most notably, while it can simulate user experience, that's not a perfect substitution for the real thing because ultimately the end-users can experience something different (and perhaps worse) due to external variables.
Additionally, since active monitoring relies on node infrastructure around the world to create those simulated experiences, it's limited by geography and ISP locations. Utilizing that infrastructure to monitor every page and navigation path can also become costly. This forces one to choose what pages he or she wants to focus on (generally the most heavily trafficked ones), which could mean missed performance problems for less visited pages or not understanding and optimizing for true click-through paths, or the best representations of true end-user behavior on a site.
For this reason, passive monitoring (done predominantly through Real User Measurement or RUM) comes into play as it picks up where active monitoring leaves off. RUM comes in two types: by devices installed in the data center, or by JavaScript embedded on the page. As the name suggests, JavaScript-based RUM measures the performance of all of the pages as experienced by real end-users, giving businesses a picture of user experience across all geographies and devices.
Whereas active monitoring lacks insight into the business metrics of a site, JavaScript-based RUM is also able to gauge the amount of time and/or money that users spend there, an undeniably important indicator of success for any sites that rely on ecommerce or advertising revenue. In doing so, passive monitoring accounts for many aspects that active monitoring is unable to cover, though it clearly cannot handle everything a company needs to optimize its website. Passive can't detect or measure downtime where it occurred. In fact, if an enterprise relies only on passive monitoring it may not know when its site is down at all. Passive monitoring also misses opportunities to optimize slow-performing sites, because many visitors won't stick around for a site owner to measure the full download time of pages. In addition, passive monitoring can't benchmark a site's performance against competition; it's unable to provide filmstrips, screenshots, ping or trace route monitoring; and it can miss performance problems during light- or no-traffic time periods.
When it comes to active and passive monitoring, electing to utilize one or the other is going to result in an incomplete picture of a site's overall performance. If a company focuses exclusively on active, then it will be left in the dark with regard to certain external factors which lie beyond its control, but ultimately still drastically affect how the site performs in the real world. Yet, companies that take a decidedly passive approach with RUM, will likely find themselves reacting to problems as they happen rather than getting out ahead of them.
That's why organizations around the world are continually realizing that this is not a matter of "active versus passive," but rather "active plus passive." The tools are now available to get a complete look at your site's performance, and to optimize it as much as possible.
:: Mehdi Daoudi is the co-founder and CEO of Catchpoint Systems ::
