Dedicated Servers in Gear of War 3

-



Why dedicated server?
Overview of datacenters



latency --- 150ms is playable, 50~90 is best

avg is 75ms

Finding the best match

How much hardware do we need for day one?

Hardware is not most expensive part
-bandwidth is most one

VOIP traffic is p2p to reduce bandwidth.

Through LSP

G4WL challenges




Modification to the existing UE3 server
- detecting empty and resting
- server shutdown whitelist
- auto config.
-memory less test
-perf counters, events, logging


Memory was not as a big deal as performance
  - run under 150MB/intance
 -  memory was cheap

Major performance wins
-strip out the renderer
-WinServer to R2



Major components of infrastructure
- master DB
( all components handshakes with DB, fixed size, parameterize stored procedure only )
- master service ( write to matter db, configuration setting of the machine )




GOW3 process
- all communication is asynchronous with db.
-db status
- performance counter


Launcher service
- runs one per server

Developer environment
Client/Server environment
- run multiple clients and servers on same machine

HTTPS webservice is better tan direct DB access- better caching of static data in the DB of offset the DB load.

The future is to cloud
- pay what you use
- more volume upfront for day one demands
-tier1 build in to the purchase
-freeze VMs on machines debug later


Comments

Post a Comment

Popular posts from this blog

glTF loader for Android Vulkan

-
Image
This post is going to introduce how we integrate with an existing glTF loader library to make it be able to show glTF models in our Vulkan rendering framework,  vulkan-android . tinygltf In the beginning, we don't want to make our new own wheel, so choosing  tinygltf as our glTF loader. tinygltf is a C++11 based library that would help us support all possible cross-platform project easily. tinygltf setup in Android Studio In  vulkan-android  project, we put third party libraries into third_party folder. Therefore, we need to include tinygltf from third_party folder in app/CMakeLists.txt as below. set(THIRD_PARTY_DIR ../../third_party) include_directories(${THIRD_PARTY_DIR}/tinygltf) Model loading from tinygltf We are going to load a gltf ASCII or binary format model from the storage. tinygltf provides two APIs , they are LoadBinaryFromFile() and LoadASCIIFromFile() respectively based on the extension name is *.glb or *.gltf. TinyGLTF loader will return a tiny...
Read more

Drawing textured cube with Vulkan on Android

-
Image
Vulkan is a modern hardware-accelerated Graphics API. Its goal is providing an high efficient way in low-level graphics and compute on modern GPUs for PC, mobile, and embedded devices. I am personally working a self training project, vulkan-android , to teach myself how to use this new APIs. The difference between OpenGL and Vulkan OpenGL: Higher level API in comparison with Vulkan, and the next generation of OpenGL 4 will be Vulkan.   Cross-platform and Cross-language (mostly still based on C/C++, but people implemented diverse versions and expose similar API binding based on OpenGL C++, WebGL is a good example).  Mainly used in 3D graphics and 2D image processing to interact with GPU in order to achieve hardware acceleration. Don't have a command buffer can be manipulated at the application side. That means we will be easily see draw calls being the performance bottleneck in a big and complex 3D scene. Vulkan: Cross-platform and low-overhead. Erase the boundary bet...
Read more

How to train custom objects in YOLOv2

-
Image
This article is based on [1]. We wanna a way to train the object tags that we are interested. Darknet has a Windows version that is ported by AlexeyAB [2]. First of all, we need to build darknet.exe from AlexeyAB to help us train and test data. Go to build/darknet, using VS 2015 to open darknet.sln, and config it to x64 solution platform. Rebuild solution! It should be success to generate darknet.exe. Then, we need to label objects from images that are used for training data. I use BBox label tool to help me label objects' coordinates in images for training data. (python ./main.py) This tool's image root folder is at ./Images, we can create a sub-folder ( 002 ) and insert 002 to let this tool load all *.jpg files from there. We will mark labels in this tool to help us generate objects' region to mark where objects are. The outputs are the image-space coordinate in images and stored at ./Labels/002 . However, the format of this coordinate is different from YOLOv2, YOLOv2...
Read more