Question

Now I want to make app to show pic from server.But found it is so slowly to make the pics come out. Is there any way or code to make the http request be more faster?

Answer 1

Approach the problem from a different angle. Think it as an entropy/economy thing.

There is two distant point in the problem you are describing and they want to transfer a data between them. Lets say this is going to cost 100 units to realize in ideal conditions. And again lets assume it is not possible to further lower this cost. This cost is where the energy required to transfer is minimum

Now assuming that transfer rates are not under our control. Here are some theoretical “seemingly” improvements which are actually just different trade-off sets.

1. Forward Caching / Caching: Preload/download all images whenever possible so they will be ready when the user requests them. Install everything static with the first time installation.

   Trade-off: You spent most of your 100 points on disk space and pre-process power this may make your app go little slower always but the performance will be great once they are loaded on disk. Effectiveness decrease as images you want changes frequently

2. Compression / Mapping: If your bottleneck is at transfer rates compress/map the images as much as you can so they will be transferred with low cost but when they arrive you will use much processor power once they arrive at the app.

   Trade-off: CPU power is used a lot more then before but while in transfer they will be moving fast. Side that compresses uses a lot more memory and side that decompress also uses more memory and CPU. Only good if your images are slow moving because they are huge then you will see this trade-offs benefits. If you want to try this install 7z and check advanced zipping settings and try really huge maps.

3. Drawing algorithms: If your images are drawings instead of bitmaps/real pictures. Send only vector graphics format, change all your pictures(raster images to be technical) to vector images. Will greatly reduces the number of bytes to carry an image to app but it will need more processor power at the end.

   Trade-off: Aside from not all pictures can be converted to vector graphics. you are going to be using more CPU and memory but there excellent libraries already built in which are very optimized. You can also think this as “mathematical compression” where a single infinite line can not be stored in any computer in universe a simple one line mathematical expression such as x = y + 1 will make it happen.

4. Write your own server: If you are sure the bottleneck is at the communication time with the service provider(in this case a http server which is most probably very efficient). Write you own server which answers you very quickly and starts sending the file. The likely hood of this improvements is to low to even talk about a trade off.

5. Never be forced to send repeating information: Design and tag your information and pictures in such a way. You will only send non-repeating chunks where the receiving side will store any chunk received to further improve its cache of information.

   Trade-Off:Combination of 1,2 and 3 you this is just another form of disturbing 100 points of cost. You get the point.

6. BitTorrent Ideology:If the bottleneck is your servers bandwidth there is a simple formula to see if using your user’s bandwidth is logical and have positive effects. Again it is probably only effective if your data set is very large.

   Trade-Off: This is an interesting option to discuss about trade-off. You will be using your users bandwidth and proximity to compensate for your servers lack of bandwidth. It requires a little bit more CPU then conventional way of acquiring data(Maintain more TCP connections)

As a closing note: There can not be a function call that can improve and make the cost of transferring information from 100 points to 95 points. In current technology level it seems that that we are really close to effectively transfer. I mean compression, mapping and various other techniques are pretty mature including network transfer methodologies but there is always a room for improvement. For example currently we think sending data with the light speed is the absolute maximum as they are electrical signals but quantum entangled observation technique denies this limit where two entangled particles theoretically send and receive information in infinite speed across universe(??). Also If you can develop a better and more efficient way of compression that will be awesome.

Anyway, as you have asked a question which does not provide much information where we can talk. I would strongly recommend thinking like an engineer, creating a test environment pointing out to the main cause and attacking it with all you got. If you can define the problem with better mathematical expression or pointing out the bottleneck we can answer it better than generic information theory.

Also one last note. I am not saying information transfer is not going to be any more efficient it may go up %1000 percent tomorrow, I am just saying this field is pretty mature to get any huge improvements without working on mathematics and theory for years. Like any other field of research.