1,平台的服务器整体架构
2,项目结构
3,业务结构
4,技术实现
5,业务举例
登录页:
登录:
其它的服务只是业务不同,这里省略...
6,应用的启动过程资源准备
7,json的封装
public static void CrossDomain2(HttpServletResponse response, HttpServletRequest request, Object msg) throws IOException {
response.setContentType("application/json");
final String call = request.getParameter("callBack");
final ObjectMapper mapper = new ObjectMapper();
String resp = mapper.writeValueAsString(msg);
if (!Strings.isNullOrEmpty(call)) {resp = call + "(" + resp + ")";}
PrintWriter out = response.getWriter();
out.write(resp);
out.flush();
out.close(); }
8,cookie域的算法
//构造cookie
final long time = System.currentTimeMillis();
String userId = String.valueOf(uid);
String USERINFO = userId + "," + userName + "," + time;
String u = Base64.encode(USERINFO + "," +
buildUserInfoSign(String.valueOf(userId), userName, String.valueOf(time)));
u = URLEncoder.encode(u, "utf-8");
String key = URLEncoder.encode(Base64.encode("USERINFO"), "UTF-8");
Cookie cookie = new Cookie(key, u);
//构造cookie加密部分
public static String buildUserInfoSign(String id, String name, String time) {
String userInfoSrc = id + name + time;
String key = ConfigurationUtils.get("login.cookieKey");
String userInfo = userInfoSrc + key;
String md5UserInfo = MD5Utils.digestAsHex(userInfo);
return md5UserInfo;
}
//解析cookie
public static int getUserId(HttpServletRequest request) {
int userId = 0;
String value = getCookieValueByKey(request);
if (!Strings.isNullOrEmpty(value)) {
String val[] = value.split(",");
userId = Integer.parseInt(val[0]);
}
return userId;
}
public static String getCookieValueByKey(HttpServletRequest request) {
String value = "";
Cookie[] cookies = request.getCookies();
String name = "USERINFO";
try {
name = Base64.encode(name);
name = URLEncoder.encode(name, "utf-8");
} catch (Exception e) {
log.error("USERINFO解码出错!{}", e.getMessage());
}
if (cookies != null && cookies.length > 0) {
for (Cookie cookie : cookies) {
String nameStr = cookie.getName();
if (nameStr.equals(name)) {
value = cookie.getValue();
break;
}
}
}
if (!Strings.isNullOrEmpty(value)) {
try {
value = URLDecoder.decode(value, "UTF-8");
value = Base64.decode2Str(value);
} catch (Exception e) {
log.error("URL解码出错!{}", e.getMessage());
}
String val[] = value.split(",");
String sign = buildUserInfoSign(val[0], val[1], val[2]);
if (!Strings.isNullOrEmpty(sign) && !Strings.isNullOrEmpty(val[3]) && sign.equals(val[3])) {
return value;
} else {
return "";}
}
return "";
}
9,http的理解和封装
public class HeadlessServlet extends HttpServlet implements InjectionProxy
里面加了一个对自定以的注解的过滤;
protected boolean filters(final HeadlessServletRequest request, final HeadlessServletResponse response) throws ServletException, IOException {
final Parameter parameter = this.getClass().getAnnotation(Parameter.class);
if (null == parameter)
return true;// filter methods
if (!Strings.isNullOrEmpty(parameter.method()) && !request.getMethod().equalsIgnoreCase(parameter.method())) {
notFound(response);
return false;
}
// filter parameters
boolean flag = true, emptyCheck = false;
try {
final Map<String, String[]> parameterMap = request.getParameterMap();
for (String name : parameter.value()) {
if (name.endsWith(":")) {
name = name.substring(0, name.length() - 1);
emptyCheck = true;
}
flag = parameterMap.containsKey(name);
if (emptyCheck) {
// must has parameter.
flag = flag && parameterMap.get(name) != null;
flag = flag && parameterMap.get(name).length > 0;
flag = flag && !Strings.isNullOrEmpty(parameterMap.get(name)[0]);
}
if (!flag)
break;
emptyCheck = false;
}
return flag;
} finally {
if (!flag) {
response.sendInvalidParameters();
}
}
}
10,db设计
11,老平台小结
老平台采用的架构很简单,一个Guice+Servlet,简单的把一个url映射到一个特定的Servlet, 这里借用Guice-servlet,省去了原生的servlet配置文件复杂的嫌疑,简单的使用Guice容器, 把所需资源按照分类和层级一一放到容器中初始化,简化了实例化的过程,提高了效率;业 务也不复杂,很容易从其它的游戏平台移植过来,业务的技术难点有两个,一个是跟其它服 务器的交互和通信,另外一个是缓存的使用;对于跟其它服务器的交互,主要是基于http 请求,如果对这一部分比较熟悉,只要文档给的比较规范,开发和调试起来还是完全没有问 题的;然后是缓存的问题,这里主要使用的Redis,一个key-value数据库,通过把经常用的 数据从db中查询出来,放到redis里面,需要的时候从redis里面取得,这大大提升了性能, 如果redis不稳定,可以直接从db里取得,保证了程序的健壮性;暂未进行过并发性测试, 目前能支撑2000w用户的使用已经算是一个还不错的架构。 |