diff --git a/Makefile b/Makefile
index 3a60486..068ff55 100644
--- a/Makefile
+++ b/Makefile
@@ -16,7 +16,8 @@ LIBOBJS = 	libs/ini.o \
 			libs/utils.o \
 			libs/ws.o \
 			libs/sha1.o \
-			libs/list.o 
+			libs/list.o \
+			libs/scheduler.o
 			
 PLUGINSDEP = libs/plugin.o
 
@@ -30,6 +31,9 @@ httpd: lib $(SERVER_O)
 	$(CC) $(CFLAGS)  $(SERVER_O)    -o $(BUILDIRD)/httpd httpd.c $(SERVERLIB)
 	cp antd $(BUILDIRD)
 
+relay: lib $(SERVER_O)
+	$(CC) $(CFLAGS)  $(SERVER_O)    -o $(BUILDIRD)/relay relay.c $(SERVERLIB)
+	cp relayd $(BUILDIRD)
 lib: $(LIBOBJS)
 	$(CC) $(CFLAGS)  $(DB_LIB) $(SSL_LIB)  -shared -o $(LIB_NAME).$(EXT) $(LIBOBJS)
 	cp $(LIB_NAME).$(EXT) $(LIB_PATH$)/
diff --git a/libs/scheduler.c b/libs/scheduler.c
new file mode 100644
index 0000000..0b04f1b
--- /dev/null
+++ b/libs/scheduler.c
@@ -0,0 +1,301 @@
+#include "scheduler.h"
+
+/*
+private data & methods
+*/
+static antd_scheduler_t scheduler;
+
+static void enqueue(antd_task_t* task)
+{
+    // check if task is exist
+    int prio = task->priority>N_PRIORITY-1?N_PRIORITY-1:task->priority;
+    antd_task_queue_t q = scheduler.task_queue[prio];
+    antd_task_item_t it = q;
+    while(it && it->task->id != task->id && it->next != NULL)
+        it = it->next;
+    if(it && it->task->id == task->id)
+    {
+        LOG("Task %d exists, ignore it\n", task->id);
+        return;
+    }
+    antd_task_item_t taski = (antd_task_item_t)malloc(sizeof *taski);
+    taski->task = task;
+    taski->next = NULL;
+    if(!it) // first task
+    {
+        scheduler.task_queue[prio] = taski;
+    }
+    else
+    {
+        it->next = taski;
+    }
+}
+
+static int working()
+{
+    return scheduler.status;
+}
+
+static void stop()
+{
+    pthread_mutex_lock(&scheduler.server_lock);
+    scheduler.status = 0;
+    pthread_mutex_unlock(&scheduler.server_lock);
+    for (int i = 0; i < scheduler.n_workers; i++)
+        pthread_join(scheduler.workers[i].pid, NULL);
+    if(scheduler.workers) free(scheduler.workers);
+}
+
+static antd_task_item_t dequeue(int priority)
+{
+    int prio = priority>N_PRIORITY-1?N_PRIORITY-1:priority;
+    antd_task_item_t it = scheduler.task_queue[prio];
+    if(it)
+    {
+        scheduler.task_queue[prio] = it->next;
+    }
+    return it;
+}
+
+static antd_task_item_t next_task() 
+{
+    antd_task_item_t it = NULL;
+    pthread_mutex_lock(&scheduler.server_lock);
+    for(int i = 0; i< N_PRIORITY; i++)
+    {
+        
+        it = dequeue(i);
+        if(it) break;
+    }
+    pthread_mutex_unlock(&scheduler.server_lock);
+    return it;
+}
+
+static int available_workers()
+{
+    int n = 0;
+    pthread_mutex_lock(&scheduler.server_lock);
+    for(int i=0; i < scheduler.n_workers; i++)
+        if(scheduler.workers[i].status == 0) n++;
+    pthread_mutex_unlock(&scheduler.server_lock);
+    return n;
+}
+
+static void work(void* data)
+{
+    antd_task_item_t it;
+    antd_worker_t* worker = (antd_worker_t*)data;
+    while(working())
+    {
+        pthread_mutex_lock(&scheduler.server_lock);
+        worker->status = 0;
+        pthread_mutex_unlock(&scheduler.server_lock);
+        // fetch the next in queue
+        it = next_task();
+        if(!it) continue;
+        //LOG("worker processing \n");
+        pthread_mutex_lock(&scheduler.server_lock);
+        worker->status = 1;
+        pthread_mutex_unlock(&scheduler.server_lock);
+        // execute the task
+        antd_execute_task(it);
+    }
+}
+
+static antd_callback_t* callback_of( void* (*callback)(void*) )
+{
+    antd_callback_t* cb = NULL;
+    if(callback)
+    {
+        cb = (antd_callback_t*)malloc(sizeof *cb);
+        cb->handle = callback;
+        cb->next = NULL;
+    }
+    return cb;
+}
+
+static void free_callback(antd_callback_t* cb)
+{
+    antd_callback_t* it = cb;
+    antd_callback_t* curr;
+    while(it)
+    {
+        curr = it;
+        it = it->next;
+        free(curr);
+    }
+}
+
+static void enqueue_callback(antd_callback_t* cb, antd_callback_t* el)
+{
+    antd_callback_t* it = cb;
+    while(it && it->next != NULL)
+        it = it->next;
+    if(!it) return; // this should not happend
+    it->next = el;
+}
+
+static void execute_callback(antd_task_t* task)
+{
+    antd_callback_t* cb = task->callback;
+    if(cb)
+    {
+        // call the first come call back
+        task->handle = cb->handle;
+        task->callback = task->callback->next;
+        free(cb);
+        antd_add_task(task);
+    }
+    else
+    {
+        free(task);
+    }
+}
+
+
+/*
+    Main API methods
+    init the main scheduler
+*/
+void antd_scheduler_init(int n)
+{
+    time_t t;
+    srand((unsigned) time(&t));
+    scheduler.n_workers = n;
+    scheduler.status = 1;
+    scheduler.workers = (antd_worker_t*)malloc(n*(sizeof(antd_worker_t)));
+    if(!scheduler.workers)
+    {
+        LOG("Cannot allocate memory for worker\n");
+        exit(-1);
+    }
+    for(int i = 0; i < N_PRIORITY; i++) scheduler.task_queue[i] = NULL;
+    // create scheduler.workers
+    for(int i = 0; i < scheduler.n_workers;i++)
+    {
+        scheduler.workers[i].status = 0;
+        if (pthread_create(&scheduler.workers[i].pid , NULL,(void *(*)(void *))work, (void*)&scheduler.workers[i]) != 0)
+        {
+            scheduler.workers[i].status = -1;
+            perror("pthread_create: cannot create worker\n");
+        }
+    }
+    LOG("Antd scheduler initialized with %d worker\n", scheduler.n_workers);
+}
+void antd_task_lock()
+{
+    pthread_mutex_lock(&scheduler.task_lock);
+}
+void antd_task_unlock()
+{
+    pthread_mutex_unlock(&scheduler.task_lock);
+}
+/* 
+    destroy all pending task
+*/
+void antd_scheduler_destroy()
+{
+    // free all the chains
+    antd_task_item_t it, curr;
+    for(int i=0; i < N_PRIORITY; i++)
+    {
+        it = scheduler.task_queue[i];
+        while(it)
+        {
+            // first free the task
+            if(it->task && it->task->callback) free_callback(it->task->callback);
+            if(it->task) free(it->task);
+            // then free the placeholder
+            curr = it;
+            it = it->next;
+            free(curr);
+        }
+    }
+    stop();
+}
+
+/*
+    create a task
+*/
+antd_task_t* antd_create_task(void* (*handle)(void*), void *data, void* (*callback)(void*))
+{
+    antd_task_t* task = (antd_task_t*)malloc(sizeof *task);
+    task->stamp = (unsigned long)time(NULL);
+    task->id = rand();
+    task->data = data;
+    task->handle = handle;
+    task->callback = callback_of(callback);
+    task->priority = NORMAL_PRIORITY;
+    return task;
+}
+
+/*
+    scheduling a task
+*/
+void antd_add_task(antd_task_t* task)
+{
+    pthread_mutex_lock(&scheduler.server_lock);
+    enqueue(task);
+    pthread_mutex_unlock(&scheduler.server_lock);
+}
+
+
+void antd_execute_task(antd_task_item_t taski)
+{
+    // execute the task
+    void *ret = (*(taski->task->handle))(taski->task->data);
+    // check the return data if it is a new task
+    if(!ret)
+    {
+        // call the first callback
+        execute_callback(taski->task);
+        free(taski);
+    }
+    else
+    {
+        antd_task_t* rtask = (antd_task_t*) ret;
+        if(taski->task->callback)
+        {   
+            if(rtask->callback)
+            {
+                enqueue_callback(rtask->callback, taski->task->callback);
+            }
+            else
+            {
+                rtask->callback = taski->task->callback;
+            }
+        }
+        if(!rtask->handle)
+        {
+            // call the first callback
+            execute_callback(rtask);
+            free(taski->task);
+            free(taski);
+        }
+        else
+        {
+            antd_add_task(rtask);
+            free(taski->task);
+            free(taski);
+        }
+    }
+}
+int antd_scheduler_busy()
+{
+    int ret = 0;
+    if(available_workers() != scheduler.n_workers) return 1;
+
+    pthread_mutex_lock(&scheduler.server_lock);
+	for(int i = 0; i < N_PRIORITY; i++)
+        if(scheduler.task_queue[i] != NULL)
+        {
+            ret = 1;
+            break;
+        }
+	pthread_mutex_unlock(&scheduler.server_lock);
+    return ret;
+}
+int antd_scheduler_status()
+{
+    return scheduler.status;
+}
\ No newline at end of file
diff --git a/libs/scheduler.h b/libs/scheduler.h
new file mode 100644
index 0000000..9d1ff65
--- /dev/null
+++ b/libs/scheduler.h
@@ -0,0 +1,97 @@
+#ifndef ANT_SCHEDULER
+#define ANT_SCHEDULER
+
+#include "utils.h"
+#include <pthread.h>
+// thread pool of workers
+#define N_PRIORITY 10
+#define NORMAL_PRIORITY ((int)((N_PRIORITY - 1) / 2))
+#define LOW_PRIORITY (N_PRIORITY - 1)
+#define HIGH_PRIORITY 0
+
+// callback definition
+typedef struct __callback_t{
+    void* (*handle)(void*);
+    struct __callback_t * next;
+} antd_callback_t;
+// task definition
+typedef struct {
+    /*
+        creation time of a task
+    */
+    unsigned long stamp;
+    /*
+        unique id
+    */
+    int id;
+    /*
+        priority from 0 to 9
+        higher value is lower priority
+    */
+    uint8_t priority;
+    /*
+        the callback
+    */
+   void* (*handle)(void*);
+   antd_callback_t* callback;
+   /*
+        user data if any
+   */
+  void * data;
+
+} antd_task_t;
+
+typedef struct {
+    pthread_t pid;
+    uint8_t status; // -1 quit, 0 available, 1 busy
+} antd_worker_t;
+
+
+typedef struct __task_item_t{
+    antd_task_t* task;
+    struct __task_item_t* next;
+}* antd_task_item_t;
+
+typedef antd_task_item_t antd_task_queue_t;
+
+typedef struct {
+    pthread_mutex_t server_lock;
+    pthread_mutex_t task_lock;
+    antd_task_queue_t task_queue[N_PRIORITY];
+    uint8_t status; // 0 stop, 1 working
+    antd_worker_t* workers;
+    int n_workers;
+} antd_scheduler_t;
+
+/*
+    init the main scheduler
+*/
+void antd_scheduler_init();
+/*
+    destroy all pending task
+*/
+void antd_scheduler_destroy();
+
+/*
+    create a task
+*/
+antd_task_t* antd_create_task(void* (*handle)(void*), void *data, void* (*callback)(void*));
+
+/*
+    scheduling a task
+*/
+void antd_add_task(antd_task_t*);
+
+void antd_task_lock();
+void antd_task_unlock();
+/*
+    Execute a task
+*/
+int antd_scheduler_status();
+/*
+    execute and free a task a task
+*/
+void antd_execute_task(antd_task_item_t);
+
+int antd_scheduler_busy();
+#endif
\ No newline at end of file
diff --git a/libs/utils.h b/libs/utils.h
index 139df74..c9b8a3d 100644
--- a/libs/utils.h
+++ b/libs/utils.h
@@ -58,6 +58,9 @@ THE SOFTWARE.
 #else
     #define LOG(a,...) do{}while(0)
 #endif
+// add this to the utils
+#define UNUSED(x) (void)(x)
+
 #define	BUFFLEN 	1024
 #define HASHSIZE 	1024
 #define DHASHSIZE 	50
diff --git a/relay.c b/relay.c
new file mode 100644
index 0000000..455df91
--- /dev/null
+++ b/relay.c
@@ -0,0 +1,123 @@
+#include "http_server.h"
+#include "libs/scheduler.h"
+/*
+this node is a relay from the http
+to https
+*/
+#define MATCH(s, n) strcmp(section, s) == 0 && strcmp(name, n) == 0
+int server_sock = -1;
+void stop_serve(int dummy) {
+    UNUSED(dummy);
+    antd_scheduler_destroy();
+	close(server_sock);
+}
+/*
+HTTP/1.1 301 Moved Permanently
+Location: http://www.example.org/
+Content-Type: text/html
+Content-Length: 174
+*/
+void* antd_redirect(void* user_data)
+{
+    void** data = (void**)user_data;
+    void* client = data[0];
+    char* host = (char*)data[1];
+    __t(client,"%s", "HTTP/1.1 301 Moved Permanently");
+    __t(client, "Location: https://%s", host);
+    __t(client, "%s", "Content-Type: text/html");
+    __t(client, "");
+    __t(client, "This page has moved to https://%s", host);
+    free(host);
+    free(user_data);
+    return antd_create_task(NULL,client, NULL);
+}
+
+void* antd_free_client(void* client)
+{
+    antd_client_t * source = (antd_client_t *) client;
+    if(source->ip) free(source->ip);
+    close(source->sock);
+    free(client);
+    return NULL;
+}
+
+void* antd_get_host(void * client)
+{
+    char buff[1024];
+    char* line, *token;
+    char* host = NULL;
+    while((read_buf(client,buff,sizeof(buff))) && strcmp("\r\n",buff))
+	{
+        line = buff;
+        trim(line, '\n');
+        trim(line, '\r');
+        token = strsep(&line, ":");
+        trim(token,' ');
+		trim(line,' ');
+        if(token && strcasecmp(token,"Host")==0)
+            if(line)
+            {
+                host = strdup(line);
+                break;
+            }
+    }
+    if(!host) host = strdup("lxsang.me");
+    void** data = (void**)malloc(2*(sizeof *data));
+    data[0] = client;
+    data[1] = (void*)host;
+    LOG("Host is %s\n", host);
+    return antd_create_task(antd_redirect,data, NULL);
+}
+
+int main(int argc, char* argv[])
+{
+    UNUSED(argc);
+    UNUSED(argv);
+// load the config first
+	unsigned port = 80;
+	int client_sock = -1;
+	struct sockaddr_in client_name;
+	socklen_t client_name_len = sizeof(client_name);
+	// ignore the broken PIPE error when writing 
+	//or reading to/from a closed socked connection
+	signal(SIGPIPE, SIG_IGN);
+	signal(SIGABRT, SIG_IGN);
+	signal(SIGINT, stop_serve);
+	server_sock = startup(&port);
+    // 1 worker is
+    antd_scheduler_init(1);
+	LOG("httpd running on port %d\n", port);
+
+	while (antd_scheduler_status())
+	{
+		client_sock = accept(server_sock,(struct sockaddr *)&client_name,&client_name_len);
+		if (client_sock == -1)
+		{
+			perror("Cannot accept client request\n");
+			continue;
+		}
+        antd_client_t* client = (antd_client_t*)malloc(sizeof(antd_client_t));
+		// set timeout to socket
+		struct timeval timeout;      
+		timeout.tv_sec = 0;
+		timeout.tv_usec = 500;
+
+		if (setsockopt (client_sock, SOL_SOCKET, SO_RCVTIMEO, (char *)&timeout,sizeof(timeout)) < 0)
+			perror("setsockopt failed\n");
+
+		if (setsockopt (client_sock, SOL_SOCKET, SO_SNDTIMEO, (char *)&timeout,sizeof(timeout)) < 0)
+			perror("setsockopt failed\n");
+
+        /*
+			get the remote IP
+		*/
+		client->ip = NULL;
+		if (client_name.sin_family == AF_INET)
+			client->ip = strdup(inet_ntoa(client_name.sin_addr));
+        client->sock = client_sock;
+		//accept_request(&client);
+        antd_add_task(antd_create_task(antd_get_host,(void*)client, antd_free_client ));
+	}
+
+	return(0);
+}
\ No newline at end of file
diff --git a/relayd b/relayd
new file mode 100644
index 0000000..dd2f1cc
--- /dev/null
+++ b/relayd
@@ -0,0 +1,8 @@
+#!/bin/sh
+UNAME=`uname -s`
+
+if [ "$UNAME" = "Darwin" ]; then
+	DYLD_LIBRARY_PATH=$(dirname "$0")/plugins/ $(dirname "$0")/relay
+else
+	LD_LIBRARY_PATH=$(dirname "$0")/plugins/ $(dirname "$0")/relay
+fi