Announcement

Collapse

The Infinity Blade Forums Have Moved

We've launched brand new Infinity Blade forums with improved features and revamped layout. We've also included a complete archive of the previous posts. Come check out the new Infinity Blade forums.
See more
See less

What exactly is a redirect?

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

    #31
    I have done u/l's overnight but most of the time I am a typical American and want the 'instant gratification' of having it right now...not tomorrow morning.

    Actually, I get to u/l 3 times: 1) the .ZIP archive to our website, 2) the raw files to our server, and 3) the compressed files to our redirect site. This is after (usually) re-archiving the original .ZIP/.RAR file that I downloaded so the embedded folder info is consistant in every .ZIP archive on our website and using Brad G's UT3-Vise utility to compress the files for the redirect server. Getting to play with Brad's toy is one reason for compressing.

    Comment


      #32
      So in other words your 3 hour saving's is lost doing all that compressing and what not. LOL j/k

      Comment


        #33


        hehe

        I was thinking about that while I was typing but the only time lost is the creation of the .uz3 files. The rest of it happens anyway. So, yah, I lose some of the 3 hrs, but not all of it, or even most of it

        Comment


          #34
          so by putting the compressed files on the redirect server it doesnt save the user any dl time? i was just going to do compression on all my maps and stuff and re upload them to my redirect to hopefully save players some time. am i wasting my time trying to do this? or does the compression actually help the dling of the files too (even if just a little bit i'll do it).... thx!

          Comment


            #35
            Generally it should save at least a little time. Of course the easy way to tell is to compress (some of) your files and compare the compressed sizes with the raw sizes. Naturally the smaller files will d/l faster.

            Comment


              #36
              Originally posted by UnrealGrrl View Post
              so by putting the compressed files on the redirect server it doesnt save the user any dl time? i was just going to do compression on all my maps and stuff and re upload them to my redirect to hopefully save players some time. am i wasting my time trying to do this? or does the compression actually help the dling of the files too (even if just a little bit i'll do it).... thx!
              Basically what you are doing is recompressing already compressed files. It's not likely you'll gain much more by compressing files that are already compressed maybe 2% or a little more but hardly worth the time and effort to me and certainly the few seconds it will save a player are negligable and I dare say unnoticable in the end. Don't forget that when a player downloads a compressed file their game has to read that compressed file. In other words it has to decompress it. This process will take some time and cpu cycles away from the time you thought you were saving the players by compressing the files. Now JonahHex has found that with a slow upload speed compressing the files does tend to save him a little bit of time getting the files up to the server or redirect and I'm sure in that case it's a help but I doubt you'll actually be saving the players who come to your server much if at all. Give it a real world test though and judge for yourself. There's certainly no harm at all in compressing your files and if you don't mind the extra steps and time involved then by all means do it. Personally I find it a waste of time, time I could better spend playing or looking for good maps to put on our server.

              Comment


                #37
                As I pointed out in an earlier post in this thread, my actual space savings from compressing the UT3 map files is 16.5%, not merely "2% or a little more." 16.5% of a 3 min d/l is just less than 30 secs. Obviously, some of the 150+ files that are on my redirect server are compressed more than the 16.5% and others less than this, but on average, users that d/l from my redirect server take 2.5 min for every 3 min that it takes users getting the same maps from a non-compressed redirect server. How much of that savings is lost due to uncompressing the maps is hard to say. Your actual savings will depend upon the specific map files you are offering -- it's your call as to whether this is worth the effort.

                On a good 1.5Kb/s DSL connection, a user can d/l about 10MB/min. On this connection, the 3 min d/l I refer to above gets the user about 30MB...about the size of an "average" UT3 custom map.

                Comment


                  #38
                  JonahHex, my 2% figure comes from my own personal experience, the number and size of our maps and from several others that have bothered to test it for themselves and reported their findings both here on Epic's site and all around the net. It's great that you are seeing an average way above everyone else but I wouldn't call it typical. In fact on a per map basis I doubt you'd be able to make a claim of more than about 2% like the rest of us. Even at 16.5% that translates to just a few seconds which again will be somewhat offset by the decompression of said file by the client. In my case I further offset that potential savings by using a reliable high speed redirect server with very good bandwidth. In the end it really would depend on the choice of maps and the particular system used to push the files as for what type of time savings if any at all would be achieved. Since most internet connections enjoy faster downloads than uploads (asyncronous) you are right that the potential for time savings is greatest for the admin actually uploading the files though I can't really see the argument for speeding up the client downloads to a noticable degree.

                  My point isn't that admins should not compress files but rather that they consider for themselves if it's actually worth all the effort in the first place when there is an almost equally viable alternative by using a good redirect. In my case it's decidedly not worth the effort for a mere 2% savings to the client. If they are using a stopwatch and testing my clans server then testing your server they might be able to glimpse a very small difference but the average game player isn't going to be timing their downloads to the second with a stopwatch and will likely never percieve this difference. Now if they were on a non-stop downloading spree of all 150+ maps on our server then yes they would certainly benefit from compression but they are not getting our maps all at one time nor is it possible without ftp access and it's doubtful that they would hang out while 150+ maps were played in order to get them all. If they did I'd say the last thing on thier mind is time being wasted. Your calculations are based on total map size before and after compression. In order to see the same benefit your clients would first have to download all your maps. The plain and simple truth is that on a per map basis, which lets face it is how our clients are going to get these maps, they will not likely see a benefit from compression.

                  Of course there are many variables that come into play here so there is no easy one size fits all answer to the question posed. For instance if an admin had access to a broadband syncronous internet connection they would then offset the upload penalty they might see with an asyncronous connection making the upstream time savings far less perceptable to even the admin uploading maps. An admin might prefer larger WAR maps on their server which would benefit more from compression, more so than lets say your average DM map. My clans server defaults to DM so the majority of our maps are smaller than a typical WAR servers maps would be.

                  On a side note, I've never recieved even one complaint that our redirects are slow. In fact quite the opposite, I've heard from many players that they like our server partly because of the speedy downloads as compared to many other top servers. As I aluded to earlier the choice of redirect hosts is perhaps the most important part of getting files to the clients fast. A good redirect alone can almost completely negate the need for compression. Couple that with the fact that the clients game/machine does not have to decompress the incoming files before game play begins and you've got the formula for speed. If you don't have the luxury of a good redirect then that of course is another one of those important variables that must be considered when deciding whether or not to compress your files. Also something cool that I can do for my clients that yours would find a bit more difficult is to send them directly to my redirect via Http. If for some reason their in-game downloads become slow I can give them my redirect Http addy and they can go there to get the files lightning fast then just pop them in their CustomMaps folder and they are ready to play. This comes in handy before matches too because they can get whatever map(s) they need before actually getting on our game server. If you did this for your clients they would first have to download a decompression program, install it then learn to decompress the files before they could use them at all. I'm not going to ask players to go to all that trouble when it isn't required and serves no tangible purpose.

                  Comment


                    #39
                    JD8, I am trying to make the same point: it depends on the maps, the capability of the redirect server, the number of maps, the admin's u/l speed, and the users' d/l speeds. My 16.5% is based on the actual total size of all ~150 files on my redirect server and the same uncompressed files...not just a sample or a rough estimate (like my statement that 30MB is about the average size of a UT3 map). Consequently, I would say that the info that both of us are providing is useful, even if it is possibly confusingly contradictory...that is point WE are making...the actual decision must be made based upon what each admin is actually using: my approach may not be appropriate for your situation and vice versa.

                    Comment

                    Working...
                    X