Now we are able to handle a month 8000000000 Disqus PV, handling 45,000 requests per second. In the comments sent to many different people, we learned some 2014 Nike KD 7 653997-001 Mens Basketball Shoes Orange Blue White Factory Outlet things. Air Jordan Outlet Disqus is known to use Django to handle most of the web traffic. At the time of any web framework, to be in development speed vs Nike Basketball performance, quick start make a choice between vs customization. And Disqus tend to quickly develop and easy to use, Nike Air Jordan 6 Women taking into account the performance and flexible customization. So why slow web framework? On the surface, the first impression of web framework is slow, because in your application, there are a lot of you do not need a code, which is Nike Air Jordan 5 Women a normal impression. In Nike Air Jordan 5 Women practice, more often than not due to the slow expansion of the framework and language selection caused. It should slow due to your request in your network in the result of communication with other services. In our scenario, these 'other services' is PostgreSQL, Redis, Cassandra and Memcached. Slow database query and network latency often drag on the performance of a robust framework like Mens Nike Free 3.0 V2 Shoes Black Blue Django a. In order to avoid these delays, people use a variety of caching Mens Nike Free 3.0 V2 Shoes Grey Green techniques. The most common way is to use the built-in cache Django libraries. Common applications cache as follows: data = cache.get ('stuff') if data is None: data = list (Stuff.objects.all ()) cache.set ('stuff', data) return data if you are familiar with Django, This should be a very common pattern. This form of caching is very simple and straightforward, and in most situations. With Memcached, it is fast enough, but in response to a request is still a lot of work to be done. It processed 45,000 requests per second that we have to deal with things very slow to do caching. However, reaching 45,000 requests per second when there is still a lot of work to be done. We may return JSON, rendering HTML templates, HTML or perform simple analytical Django middleware. The problem is, we want to make the work faster return, let Django handle it good: only unique data processing. Inside the 45,000 requests per second, how many it is unique? In it returns the number of requests is not the same with the next return? You really need to do repetitive tasks in return the same results when it? We need to return the entire HTTP cached so we do not duplicate work. Introduction VarnishVarnish is God horse? Varnish is load balancing between work and Django backend as HTTP caching layer. This means that it can return the entire Nike Air Max cached HTTP, not only for those who do not hit the Django server request. Before, Varnish makes us a black box. We install and minimally Nike Blazers configure it, to be honest, it works great. But I think we can do more. It took me some time to learn more about Varnish and we need them something. Over time, we can make thousands of requests per second without hitting Django server. Today, inside the 45,000 requests per second, only 15,000 requests hit our application server. The remaining are Varnish received, it runs very fast and effective. Because this is very useful for us, but also some 2015 Nike Free 5.0 very good learning experience, this topic has become some of my recent lecture topics. Recently, I had a speech in Chicago on DjangoCon Assembly. This lecture is for people who are not familiar Lebron Slide 2 Elite with Varnish, with hopes to inspire and drive 541100-300 Black Green White Style Nike Lebron X 10 Outlet them to learn more. For me, I was very excited this speech, because the subject is rarely mentioned application developers. This is one I hope to be able to hear the speech a few years ago, hoping to make people understand how HTTP works, and how to use tools such as Varnish to manage its interactions. (HTTP for Great Good) Before that, I attended VUG7 (Varnish User Group) held in New York, and in-depth presentation of some techniques to solve our problems. The speech Air Max 2011 Dark Black Blue was related to many Varnish configuration language we use. (Caching is Hard: Varnish @ Disqus) tl; dr learn Varnish, it will not solve all your problems, but it's worth you invest the time to learn it and assess its value. If this type of thing is very interesting to you and you like me like to one week at least five days at the computer roaring, speed and contact us, we recruit people! (Editor's Note: The original comment is also wonderful, worth a visit.) Supplementary information: Disqus is a third-party social comment system, which offers commentary hosting services for website owners. WordPress, Blogger, Tumblr and other third-party blog platform offers a third-party review Disqus plugin. By Disqus main goal is to provide powerful third comment system, the current relative isolation of different sites, isolated commenting system, connected to a large network with a community of property. By Disqus comment system with reply comments notice, comment and share hot text sharing and other social functions, site owners can effectively improve the website user activity and traffic. Users use Disqus, in comments on different sites, no need to repeat registration account, simply use a third party platform Disqus account or the account can be convenient for comment, and all comments are stored and kept in the background Disqus account, at any time convenient to view, review ʱ?? Also, when a user replies own comment, you can choose to use the mailbox to receive relevant information, to ensure Mens Nike Free 3.0 V2 Shoes Grey Green that all comments can keep abreast of subsequent behavior. Meanwhile, Disqus will social dating function well into the comment system, when a user sees on a website have similar views with their comments, can be concerned about the comments of reviewers, after attention, All comments after the reviewers will be displayed in your account background. (From Baidu Encyclopedia)Django-based Disqus how to support 8 billion monthly PV