Cedexis handles over 50 billion transactions a month to measure and optimize website and mobile-application performance. At the core of our platform, we use patent-pending technology to determine what data matters, and then make it actionable, in real time.
Real User Visibility Into Every CDN & Cloud
The test objects are probed using the Radar client to collect actionable, comparative data about the providers and platforms measured. Over 800 organizations have joined the Cedexis Radar community by deploying the Radar tag.
When someone visits their site, the browser instantiates the Radar client. The Radar client waits until two seconds after the browser’s onLoad() event and then downloads instructions from Cedexis. These instructions specify which platform to measure next during the session, picked from among the community platforms and any private platforms specific to that community member.
The Radar client initiates up to three measurements for the platform it is instructed to measure. At each step in the chain, the Radar client “cache busts” the local browser cache by appending a random number to an HTTP GET query string.
Test objects are deployed to every major CDN and cloud provider and any private platforms which community members would like to measure. These are used as the basis for all Radar measurements.
Small Test Object
A “small” object of 43 bytes, which fits within one TCP packet, is used for Resource Timing-based latency measurements.
As soon as that first download completes, the Radar client downloads the small object again, reusing the TCP connection and the browser’s cache for DNS resolution. The HTTP Response time is calculated using the Resource Timing API by subtracting startRequest from startResponse and we consider it to be an accurate representation of basic latency from the browser to the provider.
Note that as our test object is small enough to be delivered in a single response packet, our Response Time is also considered the equivalent of Time To First Byte (TTFB).
HTTP 100 KB Throughput
HTTP 100 KB Throughput
The large object is downloaded, again reusing the TCP connection and browser DNS cache, to measure Kbps based on the elapsed time between the Resource Timing requestStart and responseEnd events of the download of an uncompress-able, 100KB object over time. Not all providers are measured for throughput.
Finally, the Error Rate is calculated based on HTTP ColdStart. How often did I attempt to download a single packet and how often did the download fail. All measurement types enforce a 4 second timeout.
Page Load/Resource Time
Page Load/Resource Time
Beyond probing providers, Radar also collects Navigation Timing and Resource Timing data from the browser for each page load. This kind of data is usually called “Real User Measurement” or “RUM” and is useful for understanding per-page performance in the browser. It is not useful for making comparative choices between providers.
The Radar Tag can also identify what performance metrics relate to different sections of a website buyer’s journey. This data allows correlation between web page performance and conversion rates such as Time on Site, Pages Viewed, Bounce Rate, Conversion Ratio, and Shopping Cart Size. Such data allows Cedexis customers insight into where to focus their web performance improvement efforts to maximize business KPIs.
Frequently Asked Questions
Q: Does the Radar tag and data collection impact the performance of my website?
The Radar tag has no impact on page load time or performance, it is engineered to be non-blocking and is triggered only after the page is fully downloaded to the browser. The time offset from full page download is configurable in the Cedexis portal under “Radar configuration”.
Q: Where should the Radar Tag be deployed on my website or mobile app?
Q: How much bandwidth does the Radar Tag use?
Radar Community performance measurements of CDNs and Clouds uses less than 100 Bytes of traffic. When you elect to also collect Resource Timing metrics from the browser, the payload is 10 -15 KB depending upon the number of objects on your page data is collected on.
Q: Does Cedexis benchmark TLS maps, and load balance TLS traffic?
Yes, Cedexis Radar benchmarks the TLS maps of all major CDNs, and load balances TLS services. As a Global Traffic Management service, Cedexis does not actually carry TLS traffic, so we do not have to load TLS certificates on our platform.
Q: What is an Openmix decision?
An Openmix load balancing decision is a DNS or HTTP query response that includes an IP, URL, URI or ordered list of such values, defining where to route a user. Openmix DNS decisions are subject to ISP DNS resolver TTL, so not all end user DNS queries will be sent to Cedexis.
Q: How resilient is the Cedexis Openmix Platform?
A: Cedexis Openmix is deployed across more than 80 global datacenters, with our platform separated into four independent anycast networks. PoPs scale to multiple servers using hardened load balancing technology and we continually monitor the load and capacity. We maintain significant capacity headroom to cover customer expansion and traffic surges. All our applications are built as individually scalable services, allowing us to right-size our platform to deal with changing traffic patterns attributed to the wide range of use cases we support.
Q: How does Cedexis counter DDoS?
A: Our DDOS mitigation strategy is multilayered: Most authoritative DNS providers rely on a single anycast map across 15-20 locations. By deploying four independent anycast networks, and isolating each one from the others, we achieve significantly more fault tolerance and fault isolation. The Openmix platform itself is maximally hardened, exhaustively PEN tested, and designed for robust sandboxing (and rate-limiting) of individual applications. As a last resort, we have the ‘fallback’ mechanism, which enables us to ‘fail up’ (ie, return something rather than nothing) when the load on a particular application exhausts the resources of an individual server, giving us an order of magnitude boost in emergency capacity.
Q: What is the TTL used on Openmix applications?
Openmix customers may set whatever TTL they desire, for each individual Openmix application. The most common practice is 20 seconds, which balances data freshness with lower query volumes.
Q: What happens if Cedexis Openmix fails to respond?
Our objective is to always provide a response, if for some reason a real time data supported decision is not possible, a static option is provided as a “fallback” response.
Check out the pre-built API integrations to APM, CDN and Cloud vendors. We also support a Customer integration point, which allows customers to ingest any type of structured data available, via a RESTful interface. To learn more about our powerful scripts, visit our Developers Exchange.
The cookie settings on this website are set to "allow cookies" to give you the best browsing experience possible. If you continue to use this website without changing your cookie settings or you click "Accept" below then you are consenting to this.