If you’ve dabbled with network monitoring services before, you’re probably wondering how RUM differs from a traditional web check.
RUM captures real user experiences directly from their browsers. Whereas, web checks simulate user experiences because they are monitoring from data centers, not actual browsers.
Why does this matter?
Because there are dozens of network hops between those data centers and your actual users! Dozens, maybe hundreds of ISP’s and networks that your users depend on that you aren’t monitoring.
Now just because RUM gives you data on these “last mile” connections, doesn’t mean you can do something about them.
But what it can do is give you insight into whether a long load time is your fault or an upstream provider.
#2 There are a lot of upstream providers
When I say “upstream” that means any network, data provider, or hosting service that your users depend on to reach your site.
For any given user, they will have to cross a handful, if not a dozen of these upstream networks to reach your site. That’s dozens of chances for connection failure or latency.
RUM makes it easy for you to pinpoint latencies caused by a particular provider and then make the necessary updates to your routing configurations to avoid the issue.
Synthetic monitoring captures most of the same data, but it’s stored in traceroutes. Traceroutes are text files that list all the network hops between a monitoring node (location) and your system. Each hop is labeled by an ASN (Autonomous System Number) which maps to an upstream provider.
Unless your monitoring service has visual traceroutes (we do!), this is a tedious task.
Visual traceroutes still use synthetic monitoring, but they can give you valuable insight into how your users are reaching your site.
Say see long load times from New York. You can run a traceroute from our monitoring node in New York to your website. We’ll actually run five separate traces, which will usually return a few different paths.
As you can see, the orange line is the most used path with the fastest time. Most of the other traceroutes follow a similar path but veer off (and sometimes return) at different hops.
Both synthetic and RUM can be used to identify upstream provider latency, but only RUM can be integrated with a managed DNS provider to actually influence routing decisions.
#4 Limited data sources
Synthetic monitoring captures data from nodes that are located in data centers. A typical monitoring network will have anywhere from 50 to a few hundred nodes.
Sounds like a lot, but those 50-100+ nodes will almost always be less than the number of people visiting your website.
RUM data comes from actual website visitors. Depending on your website traffic this could mean thousands, if not millions of data points a month.
The bottom line, the more data sources you have that more accurate your data will be.