The cloud is an amorphous technological construct with varying degrees of performance and availability, depending on the location and the network of the end user. That’s where the Cedexis Radar cloud app performance and benchmarking service comes into play, providing users with visibility into cloud availability and performance metrics.
Marty Kagan, co-founder and CEO of Cedexis, explained to eWEEK that his company is now expanding the scope of its monitoring business to provide a more comprehensive view into cloud services delivery. The Cedexis Radar service is now being enhanced with improved data visualizations and alerts that enable users to compare multiple aspects of cloud availability and performance, including page load times.
The way Cedexis Radar works is by way of a network that is connected to more than 80 cloud and content delivery network (CDN) providers.
“We collect the data from all of our customers, so it’s all based on end-user measurements,” Kagan said.
For example, he noted that NBC News is a Cedexis customer and, as such, anyone who visits that site will trigger the Cedexis Radar code, which will measure the performance of the Website. The actual code is just a piece of JavaScript that runs in the end user’s browser. The idea is to get insight into what the actual end-user experience is for a given Website.
Given the vast diversity in end-user devices as well as Internet connectivity options, there is a lot of data that Cedexis can look at. Kagan explained that Cedexis can isolate different components of a download, including the TCP connect time, the DNS lookup speed, as well as rendering time across browsers and devices. The core area of interest for Cedexis right now is network latency and throughput.
Across different cloud vendors, there are myriad options for servers and applications that could be running. From an availability perspective, Kagan said those details are abstracted in order to get a big picture view into a given cloud.
“We essentially ignore server time as we work under the assumption that server response times are trivial when compared against network latency,” Kagan said. “There may be differences across servers, but it shouldn’t have an impact on what we’re measuring.”
Availability Zones
Cedexis’ core business is all about being able to compare different clouds, from the perspective of different end users coming from different Internet service providers. For example, Cedexis Radar can visually compare and contrast what the experience is for a particular cloud deployment for end users assessing the cloud from AT&T Wireless versus Verizon.
Cedexis Expands Cloud Availability Visibility
The visibility that Cedexis Radar provides is granular such that users can see how performance varies based on the specific data center that a cloud service is being deployed from. All cloud and CDN vendors have multiple points of presence in order to serve content closer to consumers and to provide additional redundancy and availability.
For example, Kagan noted that if a service is deployed in Amazon’s Virginia-based data center, that might work well for U.S.-based end users, but the Cedexis Radar service can show what the impact would be if the service was shifted to a different Amazon data center, based on what ISP the end user is using.
In the Amazon example, Amazon already provides its users with basic availability information about its cloud. Amazon recently reported a service outage in its Virginia data centers. Kagan noted that big outages are not what really concern him as he’s looking at what the actual end-user experience is like from specific Internet service providers.
“If something is slow, I want to know if it’s just me or if it’s my peers as well,” Kagan said. “That’s valuable data, and it’s what unique about our platform.”
Cedexis also has a load balancing service called Openmix that will send traffic to the most available platform, based on policy and availability intelligence, Kagan said. One thing that Cedexis does not do today is provide visibility into why a given cloud service is running slow.
“Openmix provides real-time load balancing as a service, but it skips over that answer of why a cloud is running slow,” Kagan said. “Users will be able to determine that they are getting—for example, a poor connection to Amazon Virginia and a better connection to Amazon in London— but we don’t have a whole lot of insight into why that is occurring.”
Kagan said that long term, Cedexis will address the issue of cloud availability slowdown specifics, but today the business use case is all around the fact that Openmix can enable a customer to take action in real time and the Radar service provides the visibility.
“So we don’t know why, for example, if everyone is having problems connecting to a certain provider, but we can solve the problem by shifting the traffic around,” Kagan said.
Sean Michael Kerner is a senior editor at eWEEK and InternetNews.com. Follow him on Twitter @TechJournalist.