What’s the difference between cloud hosting and traditional shared hosting? Ah, the eternal question! We provided an overview in a previous article, but now we are going to look into performance.
In a Code Project review, Microsoft MVP Kevin Griffin ran load tests on Everleap servers and reported impressive results. But he used a basic ASP.NET MVC test site, and we wondered how load tests on a more complex application would look. So we ran load tests against nopCommerce, an e-commerce storefront with a SQL database back end.
We tested on Everleap running on 1 to 3 servers, a DiscountASP.NET account with SQL add-on, Arvixe BusinessClass ASP, and GoDaddy Ultimate. You can read more details about our methodology on our website.
In this test, we recorded the average response time with a sustained load of 50 users on identical installations of a nopCommerce v3.5 demo store site.
You can see that the 50 user load had an average response time of 2.52 seconds on Everleap running on 1 server. At DiscountASP.NET, 7.37 seconds. Under the same load, Arvixe BusinessClass ASP had average response time of 20.32 seconds, and GoDaddy Ultimate, 26.43 seconds. Both of those speeds would be classified as errors in the next test we’re going to show you, because the majority of users are not going to wait more than 20 seconds for a page to load.
In this test we steadily increased the number of users and observed when the site experienced its first error. “Error” is defined as either a server error (e.g., 500 error) or a response time of 20 seconds or more. The duration of the test was one minute.
For this test, Everleap running on a single server handled 175 users consistently hitting the nopCommerce installation before generating an error. DiscountASP.NET was able to manage 110 users before an error. Arvixe BusinessClass ASP began to see errors at 35 users, and GoDaddy Ultimate at 20.
The positive effect scaling and load balancing had on the results when we increased the number of servers at Everleap is pretty clear. With 2 servers (which are available to every site at no extra charge) Everleap handled 20 times the traffic of the GoDaddy Ultimate account and more than 11 times the traffic of the Arvixe account. And as you can see, adding a third server to Everleap increased the client load by 62% over that of two servers. That’s pretty dramatic, and that’s a good example of what scaling can do for you.
The test results show that the Everleap platform outperformed the traditional hosts included in the tests, in both speed of response and in ability to maintain a high traffic load. They also showed the dramatic effect that scaling has on the ability to serve a site quickly to a large number of users.
These were stress tests of course, and the vast majority of e-commerce sites will never have to accommodate 100 users placing orders simultaneously. But if you’re running an e-commerce site you certainly want that to happen. You’re working every day to make that happen. And when it does happen, you want to be prepared. A traditional shared host is going to have a hard time keeping up with a quickly increasing or fluctuating user base.
Increasing or fluctuating user bases aside, modern web applications are growing increasingly resource hungry every day. You don’t need a ton of traffic to cripple a complex app on a shared server, as both tests show. And bear in mind as well that these were demo storefronts, so an actual store with more products and a larger database would also add to the load.
Traditional hosting certainly has its place. We host thousands of businesses at DiscountASP.NET, and the traditional shared hosting service there is working well for those businesses. But if you’re experiencing growth and need to scale an existing site, or have plans to launch a large site or application, you may want to take a look at Everleap.
You can try Everleap free, for 30 days. See for yourself how the flexibility of the cloud can work for you.
No responses yet