近期研究了一下NetMQ，设想把他用在分布式爬虫上面，NetMQ是一个封装了Socket队列的开源库，他是ZeroMQ的.net移植版，而ZeroMQ是用C写成的,有人测试过他的性能，几乎可以秒杀其他所有的MQ（MSMQ，RabitMQ等等，都不是他的对手），不过他也有一个弱点，消息不支持持久化！当然，这个功能可以自己实现，我这里只讲性能，不需要持久化

下面的例子是我基于NetMQ官网的例子修改的，下面有三个对象Ventilator 消息分发者，Worker 消息处理者，Sink 接受Worker处理消息后返回的结果，耗时的计算处理工作是交给Worker的，如果开多个Worker.exe，可以提升处理速度，Worker的最终目的是分布式计算，部署到多台PC上面，把计算工作交给他们去做（在分布式爬虫上面，每个Worker相当于一个爬虫）。

不废话，上代码（本来打算用作为序列化格式，在多线程环境下老是报一个错，暂时不知道是什么原因，所以这段注释掉了）

首先是定义要发送到消息里的对象

using System;
using ProtoBuf;

namespace Model
{
    [Serializable]
    [ProtoContract]
    public class Person
    {
        [ProtoMember(1)]
        public int Id { get; set; }
        [ProtoMember(2)]
        public string Name { get; set; }
        [ProtoMember(3)]
        public DateTime BirthDay { set; get; }
        [ProtoMember(4)]
        public Address Address { get; set; }
    }
}

using System;
using ProtoBuf;

namespace Model
{
    [Serializable]
    [ProtoContract]
    public class Address
    {
        [ProtoMember(1)]
        public string Line1 { get; set; }
        [ProtoMember(2)]
        public string Line2 { get; set; }
    }
}

然后是消息的发送者

using System;
using System.IO;
using System.Runtime.Remoting.Channels;
using System.Runtime.Serialization.Formatters.Binary;
using System.Threading;
using System.Threading.Tasks;
using Model;
using NetMQ;
using ProtoBuf;
using ProtoBuf.Meta;

namespace Ventilator
{
    sealed class Ventilator
    {
        public void Run()
        {
            Task.Run(() =>
            {
                using (var ctx = NetMQContext.Create())
                using (var sender = ctx.CreatePushSocket())
                using (var sink = ctx.CreatePushSocket())
                {
                    sender.Bind("tcp://*:5557");
                    sink.Connect("tcp://localhost:5558");
                    sink.Send("0");

                    Console.WriteLine("Sending tasks to workers");
                    RuntimeTypeModel.Default.MetadataTimeoutMilliseconds = 300000;

                    //send 100 tasks (workload for tasks, is just some random sleep time that
                    //the workers can perform, in real life each work would do more than sleep
                    for (int taskNumber = 0; taskNumber < 10000; taskNumber++)
                    {
                        Console.WriteLine("Workload : {0}", taskNumber);
                        var person = new Person
                        {
                            Id = taskNumber,
                            Name = "First",
                            BirthDay = DateTime.Parse("1981-11-15"),
                            Address = new Address { Line1 = "Line1", Line2 = "Line2" }
                        };
                        using (var sm = new MemoryStream())
                        {
                            //Serializer.PrepareSerializer<Person>();
                            //Serializer.Serialize(sm, person);
                            //sender.Send(sm.ToArray());

                            var binaryFormatter = new BinaryFormatter();
                            binaryFormatter.Serialize(sm, person);
                            sender.Send(sm.ToArray());
                        }
                    }
                }
            });
        }
    }
}

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading;
using System.Threading.Tasks;
using NetMQ;

namespace Ventilator
{
    public class Program
    {

        public static void Main(string[] args)
        {
            // Task Ventilator
            // Binds PUSH socket to tcp://localhost:5557
            // Sends batch of tasks to workers via that socket
            Console.WriteLine("====== VENTILATOR ======");

            Console.WriteLine("Press enter when worker are ready");
            Console.ReadLine();

            //the first message it "0" and signals start of batch
            //see the Sink.csproj Program.cs file for where this is used
            Console.WriteLine("Sending start of batch to Sink");

            var ventilator = new Ventilator();
            ventilator.Run();

            Console.WriteLine("Press Enter to quit");
            Console.ReadLine();
        }
    }
}

消息的处理者

using System;
using System.IO;
using System.Runtime.Serialization.Formatters.Binary;
using System.Threading;
using System.Threading.Tasks;
using Model;
using NetMQ;
using ProtoBuf;

namespace Worker
{
    sealed class Worker
    {
        public void Run()
        {
            Task.Run(() =>
            {
                using (NetMQContext ctx = NetMQContext.Create())
                {
                    //socket to receive messages on
                    using (var receiver = ctx.CreatePullSocket())
                    {
                        receiver.Connect("tcp://localhost:5557");

                        //socket to send messages on
                        using (var sender = ctx.CreatePushSocket())
                        {
                            sender.Connect("tcp://localhost:5558");

                            //process tasks forever
                            while (true)
                            {
                                //workload from the vetilator is a simple delay
                                //to simulate some work being done, see
                                //Ventilator.csproj Proram.cs for the workload sent
                                //In real life some more meaningful work would be done

                                //string workload = receiver.ReceiveString();

                                var receivedBytes = receiver.Receive();
                                using (var sm = new MemoryStream(receivedBytes))
                                {
                                    // 序列化在多线程方式下报错:
                                    /*
                                      Timeout while inspecting metadata; this may indicate a deadlock. 
                                      This can often be avoided by preparing necessary serializers during application initialization, 
                                      rather than allowing multiple threads to perform the initial metadata inspection; 
                                      please also see the LockContended event
                                     */
                                    //var person = Serializer.Deserialize<Person>(sm);

                                    //采用二进制方式
                                    var binaryFormatter = new BinaryFormatter();
                                    var person = binaryFormatter.Deserialize(sm) as Person;
                                    Console.WriteLine("Person {Id:" + person.Id + ",Name:" + person.Name + ",BirthDay:" +
                                                      person.BirthDay + ",Address:{Line1:" + person.Address.Line1 +
                                                      ",Line2:" + person.Address.Line2 + "}}");
                                    Console.WriteLine("Sending to Sink:" + person.Id);
                                    sender.Send(person.Id + "");
                                }

                                //simulate some work being done
                                //Thread.Sleep(int.Parse(workload));
                            }
                        }
                    }
                }
            });
        }
    }
}

using System;
using System.Collections.Generic;
using System.Linq;
using System.Threading.Tasks;

namespace Worker
{
    public class Program
    {
        public static void Main(string[] args)
        {
            // Task Worker
            // Connects PULL socket to tcp://localhost:5557
            // collects workload for socket from Ventilator via that socket
            // Connects PUSH socket to tcp://localhost:5558
            // Sends results to Sink via that socket
            Console.WriteLine("====== WORKER ======");

            //Task 方式多线程
            //foreach (Worker client in Enumerable.Range(0, 1000).Select(
            //    x => new Worker()))
            //{
            //    client.Run();
            //}

            //多核计算方式多线程
            var actList =
                Enumerable.Range(0, 50).Select(x => new Worker()).Select(client => (Action)(client.Run)).ToList();
            var paraOption = new ParallelOptions { MaxDegreeOfParallelism = Environment.ProcessorCount };
            Parallel.Invoke(paraOption, actList.ToArray());

            Console.ReadLine();
        }
    }
}

接受消息处理的结果

using System;
using System.Collections.Generic;
using System.Diagnostics;
using System.Linq;
using System.Text;
using System.Threading.Tasks;
using NetMQ;

namespace Sink
{
    public class Program
    {
        public static void Main(string[] args)
        {
            // Task Sink
            // Bindd PULL socket to tcp://localhost:5558
            // Collects results from workers via that socket
            Console.WriteLine("====== SINK ======");

            using (NetMQContext ctx = NetMQContext.Create())
            {
                //socket to receive messages on
                using (var receiver = ctx.CreatePullSocket())
                {
                    receiver.Bind("tcp://localhost:5558");

                    //wait for start of batch (see Ventilator.csproj Program.cs)
                    var startOfBatchTrigger = receiver.ReceiveString();
                    Console.WriteLine("Seen start of batch");

                    //Start our clock now
                    Stopwatch watch = new Stopwatch();
                    watch.Start();

                    for (int taskNumber = 0; taskNumber < 10000; taskNumber++)
                    {
                    //while (true)
                    //{
                        var workerDoneTrigger = receiver.ReceiveString();
                        Console.WriteLine(workerDoneTrigger);
                    //}
                    }
                    watch.Stop();
                    //Calculate and report duration of batch
                    Console.WriteLine();
                    Console.WriteLine("Total elapsed time {0} msec", watch.ElapsedMilliseconds);
                    Console.ReadLine();
                }
            }
        }
    }
}
再次提醒，Worker.exe 可以开多个，以提高效率