How to control CDN traffic gracefully in cloud development?

In the community , There are a lot of feedback from small partners using cloud development “CDN Flow consumption is like running water ” The situation of .

One day I woke up ：

Have been loaded with high quality pictures “ scared ” I dare not use it ：

So here comes the question , How to manage gracefully in cloud development CDN Traffic consumption ？ This article will talk with you in detail ！

01、 Pay as you go and pay as you go

For the sake of understanding , Let's take a look at the traffic billing mode of cloud development first .

Simply speaking , Pay by volume Namely ： How many of you ran out , I'll give you the quantity ; But it doesn't limit how much you can run out at the same time , That is, there is no limit to the peak flow .

If you need to run out at the same time 100M The amount of , Then I'll drive you to the peak 100M, Run at the same time , The final cost is 100M The money .

and Pipeline payment It's about limiting how wide it's going to be for you “ road ”, According to this “ road ” Wide and narrow charges of , For example, you choose 1M The road of the Internet , Then according to 1M I'm going to charge you at a reasonable price ; When use , Your quantity can only reach 1M The speed of , It can't be any higher , This maximum speed is called bandwidth .

But even when the low peak doesn't have much , You also have to work for this “ road ” pay . Or the above , You need to run out at the same time 100M, But the speed will only give you to 1M, So these quantities are 100 You can walk in seconds , The final cost is 1M The road rent of * The length of the lease .

And in real projects , There will always be peaks and lows in business , It's almost impossible to keep the flow at a constant rate all the time , This creates a ceiling with speed 、 But the waste of resources that are still charging when idle .

Cloud development as Serverless Cloud native integrated back end service , The traffic provided is paid by volume , Pipeline payment is not supported , Don't limit the flow rate at the same time , therefore , If developers use it unreasonably, it will lead to too fast flow , This will lead to rapid excess flow .

After understanding the difference between pay as you go and pay as you go , Let's get to the point , say something CDN.

02、 What is? CDN Traffic ？

When we use cloud development , You can see it almost everywhere CDN The word .

CDN also called Content distribution network , Generally speaking, it is to store your main memory （ Origin station ） Documents in , Copy to storage points everywhere （CDN node ）, When a user accesses this resource , Directly from the nearest storage point （CDN node ） Get it .

Natural support for cloud storage and static website hosting of cloud development CDN Speed up , So the traffic that your users download files through the client is CDN Traffic .

above CDN Traffic is only applicable to all kinds of channels to speed up the downloading of files from the public network , For example, the browser loads the website , Client downloads resources , External system request file , Opening files through various accesses to temporary addresses is consuming CDN Traffic .

But if it's an intranet access file , I will not go CDN Flow consumption , For example, in the cloud function through fileID Access files, etc .

03、 What is? CDN Back to source traffic ？

When we have file updates in our storage , Stored in CDN How to deal with the old file of the node ？ Introduce a knowledge point here —— Cache time .

The cache time here is actually the storage point of file copies in different places （CDN node ） Effective time of , For example, the default is two hours , So the effective time of each copy of the file in each storage point is two hours , When you receive a request after this time , Storage point （CDN node ） Will discard old files that are out of date , To the primary storage （ Origin station ） Request the latest file , And the traffic generated by this request is called CDN Back to source traffic .

The cache time cannot be too long or too short , If CDN Cache interval too short , that CDN Data on nodes will often fail , Cause frequent back to source , The load of the source station is increased , And then affect the overall transmission efficiency ; If the cache interval is too long , It will bring serious business problems such as untimely data update .

Cloud storage of cloud development provides very fine-grained cache time settings , You can target a file 、 A path or even a file suffix is used to set it separately .

in addition , There are also priority policies in multiple cache rule settings , Deployment becomes more flexible .

Cloud storage does policy computing in a back to front configuration mode , For example, a cloud storage domain name has the following cache configuration ：

Now request resources under this path /test/abc.jpg, The matching method from back to front is as follows ：

Match the first all files , hit , The cache time is 2 minute .

Match the second , Not hit .

Match the third , hit , The cache time is 300 second .

Match the fourth , hit , The cache time is 400 second .

Match the fifth , hit , The cache time is 200 second .

04、 How to reasonably control CDN？

We need to grasp a principle ：

Reduce the size , Make good use of cache .

Reduce the size It means , When we develop applications , Multimedia files needed , If there are no special requirements （ For example, the original pictures of the camera are sent ）, It needs to be compressed as much as possible . Only by reducing the size of business resources , Can fundamentally reduce the consumption of traffic resources .

in addition , When users upload multimedia resources , You can still use platform or framework capabilities to compress resources before uploading them , Ensure that all resources are optimized and stored , This will reduce a lot of load when you request to download .

Make good use of cache It means the same user 、 Try not to request the same resource more than once , To use the client's local caching capability properly , Cache all fixed resources . When the user enters again , Using cached resources directly .

I'd like to give you another golden sentence ： The cache works well ,PV The effect is just UV The amount of .

And in specific use , We still need to reasonably grasp it according to our own business situation . For example, your resources change very frequently , It's not suitable for cache optimization ; And your resources can't be compressed , Keep it the same size , It is not suitable for compression optimization .

05、 Conclusion

The above strategy is for you hold Unable to live CDN Friends of traffic , If you think CDN Consumption is like running water , Users are active but not much better , Maybe you need to check where you waste . Thrift is a traditional virtue of the Chinese nation , Cloud development is good , We also need to use it sparingly ～

author ： Yunyun senior