There’s a lot of talk these days about digital transformation and the digital experience, what seems to be missing from some of these discussions is the impact the network has on digital experience. In the run/grow/transform model of business most departments want to be a part of the transformation, but the network is typically thought of as something that is needed to run the business. It can’t lead to growth or transformation of the business.
I would like to position things a little differently. If the network isn’t considered, an organization cannot grow or transform. As digital interactions become the norm for key business transactions, organizations need to rethink the network. It’s not just making sure information gets from point A to point B, but about how to extract value from the data as well. By giving network professionals a seat at the table, organizations will be able to grow and transform.
The network is more than just the plumbing or connectivity that enables information to flow across systems; it can reveal insights as applications move to the cloud.
Applications aren’t always developed with the cloud in mind. As more enterprise applications move to hybrid cloud environments, performance may be impacted. The more dependencies on network and Internet health, the more potential for performance issues.
I can’t tell you the number of times I’ve seen (and I’ll admit, created and presented) a simplistic view of the application delivery chain similar to the one below.
The end user requests information, the request travels along the network to the server and is delivered. Easy peasy. In reality, the delivery chain is much more complicated than that. Looking at the network alone, there’s the last mile, the middle mile, the backbone, and the datacenter. The last mile is the ISP being used by the end user; this could be a fiber connection, mobile connection, cable connection, satellite connection, etc. The request doesn’t travel directly from the end user to the datacenter, it makes a number of stops in between. It travels the “middle mile” or backhaul network to connect the request from the local provider to the backbone or core network. Once on the backbone it’s routed to the application, whether in a physical datacenter or the cloud.
To further complicate matters, content for applications doesn’t come from a single source. For some applications, the majority of content is being delivered by third parties such as content delivery networks, ad servers, social media sites, analytics components, and more. This can result in a rather complicated network delivery chain.
The chart below shows the number of unique hosts and connections for three different web sites. For these sites there are up to 100 DNS lookups and 100 TCP connections, which means more than 100 of these leads to more chances for problems. It’s hard enough keeping a few networks healthy, but having to worry about the health of hundreds of networks can be overwhelming.
If the network isn’t healthy the user experience will suffer. Being able to quickly identify network issues can ensure application health and an optimal user experience—whether that is your network or somebody else’s network doesn’t matter.
What does it mean to have a healthy network?
If an application isn’t available there is definitely something wrong, but identifying whether the problem is internal or external is a key component. Is the application inaccessible to all users, or only some? Availability issues could be the result of:
- A misconfigured firewall
- A power outage at a data center
- A fire at a power plant
- A curfew imposed by the government
- Or maybe one of these 10 bizarre stories
There are infinite reasons that could explain why an application may not be available. Being able to quickly identify whether the issue is global or regional will determine if or how to take action.
All network protocols that are relevant to the running of the business need to be accessible and responding in a timely manner.
- Are DNS queries resolving?
- Can I SSH into a server and make configuration changes?
- Is mail being received from SMTP or IMAP servers?
- Are users able to upload and download support files to an FTP server?
- Is the NTP server reporting the correct time and keeping all servers synchronized?
- Are packet loss and latency within expected thresholds?
- Do the APIs return the correct information?
The health of the network can only be determined if you are monitoring and measuring the correct components from a variety of vantage points. Being able to quickly identify whether an issue is related to a specific ISP, a specific geography or a specific component of an application results in a faster resolution and a healthy network. The network may not hold all the answers but it can provide facts on protocols, latency, and much more eliminating the dreaded finger pointing and blame games.