CloudFlare railgun is a “network optimizer” that accelerates the dynamic portions of a website. No one has a monopoly over the definition of the term: network optimizers, but it is certainly not limited to compression.
In this blog post we are going to take a look at the technology behind Railgun. We will also see how Datapath.io’ network optimizer can be used as an alternative to CloudFlare Railgun to accelerate dynamic content delivery.
Why do we need a Dynamic Content Delivery Solution?
CDNs work well for static content like images, pdf files and static videos. However once dynamic content enters the picture the concept of caching content at edge locations ceases to be of any use.
Dynamic content delivery is usually the realm of network optimization solutions. Dynamic content has to be fetched from the origin server every time. This means optimizing the entire network path from the actual web server to the end user.
Social media websites, e-commerce, online media, online gaming, VOIP and Adtech all have elements of dynamic content.
What is a Network Optimizer?
A network optimizer allows NetDevOps and network engineers to improve performance over the public internet. They do this by inserting a set of performance metrics into the best path selection algorithm of BGP. Network performance metrics like latency, congestion, jitter and packet loss are crucial to determing end-user QOS and experience.
Datapath.io: Alternative to Railgun
The key difference between CloudFlare Railgun and Datapath.io network optimizer is the technology being used behind the scenes.
Railgun compresses the dynamic portions of a website to make dynamic content delivery faster.
Datapath.io on the other hand creates a real time performance map of the internet. We then use this map to route internet traffic over the fastest and best performing internet routes. This type of optimization works both for static as well as dynamic content. No ttl issues result, because every request is optimized regardless of the source or the destination.
This results in performance improvements of up-to 60% in latency, congestion, jitter and packet loss.
Using compression for the delivery of dynamic content still leaves the issue of CDN ttl’s. Railgun will still wait for the ttl to expire before going to the origin server for newer content.
Railgun is also limited by the range of internet real estate it can optimize. Railgun can only optimize dynamic content delivery from the origin to one of its own edge locations. Datapath.io on the other hand optimizes latency, bandwidth and packet loss to every individual network prefix. There are upwards of 600,000 of these. This basically means that Datapath.io optimizes performance over the complete internet.
Railgun is also limited by the type of network performance metrics it can optimize. This results in its in ability to serve some web services. One example is online gaming latency experienced during online games. Another would be the latency for real-time bidding requests between SSPs, DSPs and ad networks during the ad serving process. Jitter and latency experienced by VOIP services also fall into this category. These performance problems are a result of the underlying routing policies of the internet and as such compression cannot make any noticeable improvements.
These services can only be improved by using a solution that can address performance bottlenecks like latency, congestion, packet loss and jitter.
Download the AWS Optimization Whitepaper to learn more.