LinkedIn Connectivity Problems? What We Found….
by December 15, 2010

Filed under: Performance Monitoring

Apparent Networks

Our legacy blog entries tell the AppNeta story from the early days…Read on!

With the growing frenzy over cloud-based services and the recent announcements about managed cloud services, it is clear that we are looking at the future of business technology, one that merges critical business services with the cost efficiency, speed and easy integration of the internet.

But, as one article aptly noted, “The cloud just might be the biggest thing to hit the internet in years, but that doesn’t mean it’s a cinch to keep it running perfectly.”

Companies of all sizes are making this move to the cloud –20% – 35% of end users are already adopting SaaS solutions- whether they be cloud in Amazon EC2, Saas like SalesForce.com or hosted Exchange. It is critical to not only monitor the availability and performance of those hosted servers, they need to ensure that their employees and customers can access them continuously and efficiently.

Let’s take a look at a timely example. If any of you have trie

d to access your LinkedIn account in the past three days, you know how important it is to manage the performance of  your network, web-based applications, and business dependent websites, from any remote locations.

Many of us here at Apparent noticed that LinkedIn was experiencing frequent outages, so we decided to take a look at what was going. This is what we have found!

LinkedIN Network

LinkedIn apparently was expanding its infrastructure with a new major datacenter in Los Angeles. While this should improve website performance long term, the network changes and application deployments impacted the end user experience from a variety of locations – the performance of LinkedIn varied DRASTICALLY from Boston, to London, to New Jersey and Maine. One interesting detail is that each of these remote sites was accessing LinkedIn.com from different circuits. And when LinkedIn was able to fix the AT&T circuit, it caused a 50% increase in latency from all other locations accessing LinkedIn.com on Verizon.

While we don’t want to pinpoint LinkedIn, it is critical for all organizations to have direct insight and understanding of their networks and their web-based applications, from any location and via any cloud-service provider.

As the move to the cloud expands across all business services, from hosted email, to CRM, to backup and recovery we are all becoming more and more vulnerable to service degradation and service failure – and many companies, like LinkedIn, cannot afford this impact on end-user experience.