Data Centers exist all over the world. They’re used to stored all sorts of computer data, including websites and web apps, like MIDAS.
Does Data Center location matter?
Yes – The location of the data center that houses your cloud-hosted MIDAS system can have an impact on the quality of your overall experience.
One of the unavoidable aspects of long-distance internet communications is high latency.
What is latency and why is it important?
The term “latency” describes a measure of delay between two events.
In the context of the internet, latency refers to the amount of time it takes for data to perform a round-trip between two points in a network. In the case of your MIDAS system, these two points are represented by your web browser and the server in a data center which runs your MIDAS system.
The amount of time it takes for a unit of data to travel from the server in the data center to your browser is considered the latency of the network. This is usually measured in milliseconds and expressed as ms. This is also frequently referred to as the response time of a server.
We recently added two new data centers to our network, including one in Europe.
In our testing, we saw lower latency leading to an approximately 30% improvement in page loading times for Europe-based users accessing a MIDAS system in our EU data center vs our East Coast US location.
How does a data center’s location help reduce latency?
Being geographically closer to a data center can have several benefits, including:
- Lower Latency: When the distance between a user and a data center is shorter, the time required to transfer data is reduced. This results in lower latency and faster response times.
- Improved Performance: Lower latency can result in improved performance for applications and services hosted at the data center, such as faster loading times for websites, smoother streaming, and reduced lag in online gaming.
- Increased Reliability: By being geographically closer to a data center, users are less likely to experience network congestion, data loss, and other issues that can affect the reliability of their connections.
- Better Compliance: Data centers located in specific regions may be subject to local laws and regulations, such as data privacy and security standards, which can affect their suitability for hosting certain types of data. By being closer to a data center, organizations can more easily ensure compliance with these regulations.
How far should data centers be apart?
Having two data centers physically located next door to each other has little benefit. The whole reason to have data centers spread further apart is to reduce risk of downtime and to increase performance.
If two data centers are built next to each other and a natural disaster – like a hurricane or wildfire – hits that location, both data centers are at risk of suffering damage and outages. Whereas, if data centers are built several hundred – or indeed thousands – of miles apart, the risk of the same natural disaster hitting them all is negligible.
Similarly, in terms of performance, if a user is located in London, England, but a cloud-based application they’re using can be run from one of two data centers both located in San Fransisco, California, it will make very little difference to the client as to which data center their application is run from. However, if the client has a choice between a data center in California or a data center in the United Kingdom, there will likely be significantly lower latency from the UK data center.
So, how far apart should data centers be? They should be far enough apart as to minimize a potential disaster occurring at one location from impacting operations at other locations.
They should also be geographically distributed based upon the location that users will be connecting from. If all users will be connecting from within the US, it makes sense to use data centers around the US. If users will be connecting from all around the world, it makes sense to use data centers spread across the globe.
Our Data Centers
Earlier this year, we significantly expanded our network into new data centers.
We added new client nodes in a data center in Europe (in Amsterdam, Netherlands), and in a data center on the West Coast of the US (in Seattle, Washington).
These two new data center locations are in addition to our existing nodes residing in our East Coast US data center (in Atlanta, Georgia).
We then looked at the location of each organization with an active cloud-hosted MIDAS service. Those which were geographically closer to either our EU or West Coast US locations were seamlessly migrated to those data center locations.
Naturally, we provided customers with advance notice of proposed migrations, and allowed individual customers to opt-out, or to choose a different data center for their MIDAS system.
The client migrations went smoothly, and have lead to noticeable improvements through lower latency for many customers.
Because our cloud-hosted MIDAS systems can now be run out of three different geographic regions (US East, US West, and Europe), we now offer all new cloud-hosted customers a choice of data center where they’d like their MIDAS system to reside.