When you are trying to improve your company’s online performance, you have to follow a certain hierarchy that relies on the satisfaction of the previous tier’s requirements. For example, you wouldn’t want to focus your resources on improving the speed of your website unless you were sure it was supported by a reliable backend (more on this in a moment). Without the basic building blocks in place, the rest will collapse.
If you ever took a Psychology 101 class, this may sound reminiscent of Maslow’s hierarchy of needs. Here’s a quick refresher, humans have 5 core needs that motivate them. Each tier of needs builds off of the last one, and you cannot continue to the next tier until the previous one’s needs have been met. Eg: In order to feel “safe”, a person first needs to satisfy their physical needs. And so on…
In this post, we will be outlining the 5 basic “needs” any system administrator, developer, or IT team must satisfy in order to achieve optimal Internet traffic performance.
When it comes to Internet performance, the most basic of needs is uptime. To satisfy this need, you must have reliable hosting providers for your DNS and web hosting. If you have ever registered a domain before, you may remember that your DNS registrar will usually include DNS hosting for free. These free hosting solutions only cover the bare bones of your hosting requirements and are usually supported by networks not engineered to cope with modern threats. We have seen first hand that using a registrar’s DNS services has proven to be unreliable time and time again. We recommend looking at using a third-party solution for both your DNS and web hostingto ensure the most redundancy.
In this case, redundancy is a good thing. We recommend looking at using a third-party solution for both your DNS and web hosting. This way you don’t have all your “eggs in one basket” and can distribute the responsibility for your uptime across multiple networks and providers.
When evaluating providers, make sure they have an SLA (Service Level Agreement). This agreement protects the client in the event of an outage and usually involves the reimbursement of services lost. We also recommend using SLA monitoring for your hosting client to detect latency and service disruptions. Some services, like Constellix Sonar, will alert you when there is latency or downtime so you can be proactive and make the appropriate changes.
As you continue through the following tiers, keep in mind that you should always diversify your vendors. Meaning, don’t use a “bundled” service and have your website or app resting on a single point of failure. Providers that try to sell all-in-one solutions also lack the expertise that providers who only offer a single service have.
Once you have the services in place for reliable uptime, you can start optimizing your site or application for speed. Take a look at every asset that you are using on your site and look for ways to compress your assets such as images, CSS, and scripts. Use a Waterfall Check or Real-User Monitoring tool to identify the different assets that affect load times. We recommend using the Constellix Sonar Lite chrome extension, available for free in the chrome web store. The RUM tool lets you compare the performance of over a dozen websites and view the load times of every request required to render a page.
The extension is developed by our sister company, Constellix. Our team has used the tool in the past to cut our homepage load times in half. You can see the full study here.
At this level, we can focus on performance at a much larger scale. If you have clients in other regions or countries, you will need to make sure your website or app loads as fast is it would if they were in your backyard. This requires the use of a large network that delivers from dozens of locations around the world. If you were to do this using your own infrastructure, it can cost tens of thousands (in some cases millions) of dollars. But thanks to the cloud, you can do so for just a couple dollars a month.
First, you want to find a reliable third-party DNS hosting provider with an Anycast network and location-based routing features. These features allow you to customize your traffic routing on a regional level, so clients in one region will have their queries answered from a server closest to them. These networks are highly redundant and can withstand massive amounts of traffic (including DDoS attacks).
Now that your network is fast and secure, you want to look at content delivery. CDN (Content Delivery Networks) also use Anycast networks and instead of hosting your DNS information, it hosts your web content.
Analytics are crucial to maintaining Internet performance as conditions are constantly changing. However, up until recently the buck usually stopped here for most businesses. Big data analytics are costly, requiring vast amounts of infrastructure to store and present data into something that could be analyzed. Recent advancements in cloud technology have made analytics accessible to even small businesses.
We mentioned one of these analytics services in an earlier tier called Real-User Monitoring (RUM for short). RUM tools are able to show you the real-time performance of your end-users from every country in the world. You can use this information to create unique routing solutions to improve performance for clients in different regions, cities, Internet connections, and much more.
Another tool we recommend is Real-Time Statistics (RTS). These tools use metrics that all DNS providers have, but can take up massive amounts of space. For example, the DNS Made Easy network answers 40 billion queries per day. RTS takes all of these queries, where they are coming from, record type, and time frame; and presents this information in a user-friendly report. Admins can use these tools to predict DNS-based attacks, identify routing latency, and troubleshoot connectivity issues.
The top of the pyramid is a fairly new concept called ITOS, or Internet Traffic Optimization Services. ITOS is a new approach to managing online performance using real-time metrics and intelligent traffic routing. Typical performance management platforms require you to constantly monitor and manually update your records to keep up with performance demands. This is stone-age thinking, which has historically led to service impacting latency, or worse… outages.
ITOS operates at the root of the Internet, manipulating how traffic is managed on the DNS level. All decisions are made based on real user information gathered in real time from tools we mentioned earlier like RUM and RTS.
Once you’ve reached the top of the performance pyramid, don’t think you’re done. Internet performance has to be monitored and maintained. Traffic conditions never stay the same and routing configurations have to keep up with the changes. An automated ITOS solution will monitor your traffic and intelligently route clients to maintain performance. If you want to learn more about ITOS, download this free ebook and learn how you can implement an ITOS solution for your brand.
Also published on Medium.