If you've looked at cloud computing, you've run into the "as a service" alphabet soup. The big ones are Software as a Service (SaaS), Platform as a Service (Paas), and Infrastructure as a Service (IaaS). They just indicate how "close to the metal" the service brings you.

For a lot of purposes, you just want a tool to do a job, and you don't care about the implementation details. That's why SaaS is so popular. It doesn't matter what operating system or hardware it's running on. Setting it up is simple. You just want it to do the job right, and that's what SaaS gives you.

Network monitoring as a service

If you want to track how well your network is doing and catch problems, SaaS makes good sense. When there are network issues such as overloading or disconnection, you want to know quickly. Extended downtime can lead to lost business and reduced productivity. The need for reliable notification means the monitoring should come from outside your premises, so it will keep working even if the network doesn't. A cloud-based network monitoring service keeps working, no matter what problems the network has.

A more traditional approach to monitoring is a dedicated system that's external to the network. This is an attractive option when the highest level of security is necessary. The network's on-premises agent for the monitor accepts communication only from one machine, which doesn't share any facilities with other software. This is an expensive solution, though. If the monitoring device can't keep up with a growing network, it will have to be replaced with a more powerful one. If it goes down, monitoring stops.

A dedicated system inside the network's facilities is the most secure approach. It shares many of the problems of an external device, though. It may have to work beyond its capacity as the network grows. In that case, replacement with a more powerful device will eventually be necessary. If the network fails at the router level, an in-network monitoring device could be unreachable, and the only indication of a problem will be the lack of response.

Benefits of SaaS

Cloud-based network monitoring doesn't depend on the availability of a particular machine. A cloud service will keep running even if there are some hardware glitches. It's considerably less expensive than a dedicated device.

SaaS is scalable. It can use greater or lesser amounts of processing power and network capacity as the need arises, so you just have to choose your plan level and the number of devices to monitor. If your needs change, it's simple to adjust your service.

You can access the SaaS service from any location to see how your network is performing. You can configure it to alert you when your network is down or seriously misbehaving. If it's under a serious denial-of-service attack, you'll still have access to the monitoring console to see what's happening.

Security considerations

A network needs an on-premises software agent to give the monitor access. To do its work properly, the monitor needs detailed access to network traffic. For this reason, you have to choose your monitoring service carefully. Giving  inside access to your network requires choosing a service that maintains the highest level of security.

If your organization's security requirements are unusually strict, using a cloud service for network monitoring may not be an option. A dedicated external device offers more security, since only one machine accesses the network, and the on-premises agent can refuse communication from any other machine on the Internet. An on-premises device is the most secure, but a network outage can cut it off from reporting the problem. The right tradeoff depends on the organization's requirements and obligations.

SaaS monitoring promises a high level of availability; make sure the service you choose delivers on that. You want outages to be as rare as possible, and you want notification when they're scheduled.

For many situations, SaaS-based network monitoring is the most effective approach. Network monitoring from a reliable provider gives you early detection of problems, reduced downtime, and increased productivity at a reasonable cost.