In a recent post, Facebook announced plans to increase the rank of webpages that load faster in its News Feed. Since this is an important approach that has far-reaching effects, let’s delve into the details to understand the implications and angles.
According to Facebook, “Webpages that are particularly slow could see decreases in referral traffic.” This should be already by itself a matter of concern and definitely draw the attention of all involved stakeholders. The same blog also mentions that “Factors such as the person’s current network connection and the general speed of the corresponding webpage will be considered.”
So, how can we prevent the general speed of the corresponding webpage from becoming an issue?
First, there are many best practices that site owners can follow to improve the load-time of their web pages. In the above-linked post, Facebook offers some very good examples that can also be directly accessed here.
However, moving down the pipe, there are more aspects that other stakeholders (especially those responsible for managing the serving infrastructures) should consider. Namely, what are the ramifications and potential effects of the underlaying infrastructure?
Multiple elements usually found on the path to a site can impact load-time. Just to name a few: carrier-grade NAT (CG-NAT), proxies, application delivery controllers (ADC), server load balancers (SLB), firewalls and next-generation firewalls (NGFW) and the list can go on. It is not uncommon for such elements to face performance issues and become potential bottlenecks on various key performance indicators (KPIs) that a network must sustain, such as overall throughput, concurrent sessions, connections per second just to name a few. Of similar importance, but often overlooked, are the resiliency and stability aspects of these products. These can only be validated under long-duration, stressful conditions and then applying negative-testing scenarios. These aspects eventually have a direct impact of the user quality of experience (QoE), which is of paramount importance.
Therefore, it is important to perform a thorough characterization of such infrastructure elements before placing them in production environments, and make sure they are properly sized for the production environment.
Our recommendations for such performance characterization are not limited only to the pre-deployment phase, but also for change management to avoid unwanted degradations down the road. Change management includes uploading a new firmware version or software patch, changing configuration, and adding new services.
A well-planned process of validating infrastructure elements can prevent numerous issues with performance and scalability and eventually help get you to the fast lane and meet customer QoE expectations.
It is no secret that application performance has a direct effect on business results. Below are just some examples of the results that few companies learned from various studies they performed:
*source: Steve Souders @ Velocity Conference
Based on several years of testing and validating numerous devices, Ixia has put together a series of test methodologies to assess the impact that various network elements might introduce on performance and security. These resources are available for free to everyone interested, and few examples include the Application Delivery and Network Security Black Books (other complementary ones here).
Ixia helps you stay prepared with the test solutions you need to validate the scale and performance expected from modern requirements. Designed with future testing requirements in mind, Ixia’s application and security test solutions include hardware and software optimizations that ensure superior test performance.