序列化之protobuf与avro对比(Java)
Posted upupgo
tags:
篇首语：本文由小常识网(cha138.com)小编为大家整理，主要介绍了序列化之protobuf与avro对比(Java)相关的知识，希望对你有一定的参考价值。
最近在做socket通信中用到了关于序列化工具选型的问题，在调研过程中开始趋向于用protobuf，可以省去了编解码的过程。能够实现快速开发，且只需要维护一份协议文件即可。
但是调研过程中发现了protobuf的一些弊端，比如需要生成相应的文件类，和业务绑定太紧密，所以在看了AVRO之后发现它完美解决了这个问题。
下面记录下对这两种序列化工具的入门与测评。
一、protobuf基本操作
protobuf简介：
Protocol Buffers (a.k.a., protobuf) are Google\'s language-neutral, platform-neutral, extensible mechanism for serializing structured data.
protobuf是google提供的一种跨语言、跨平台、可扩展的序列化工具。
1.1定义协议文件(部分字段)(TCPLog.proto)：
syntax = "proto2"; message TCPLog{ optional int32 total_byteps = 1; optional int64 flow_start_time =2; optional int64 date =3; }
1.2生成对应的Java类：
生成过程可以使用ecplise 的插件 或者 直接在控制台中使用命令生成。
命令行中生成规则如下：
protoc.exe -I=proto的输入目录 --java_out=java类输出目录 proto的输入目录包括包括proto文件
生成java类如下：
// Generated by the protocol buffer compiler. DO NOT EDIT! // source: TCPLog.proto public final class TCPLogOuterClass { private TCPLogOuterClass() {} public static void registerAllExtensions( com.google.protobuf.ExtensionRegistryLite registry) { } public static void registerAllExtensions( com.google.protobuf.ExtensionRegistry registry) { registerAllExtensions( (com.google.protobuf.ExtensionRegistryLite) registry); } public interface TCPLogOrBuilder extends // @@protoc_insertion_point(interface_extends:TCPLog) com.google.protobuf.MessageOrBuilder { /** * <code>optional int32 total_byteps = 1;</code> */ boolean hasTotalByteps(); /** * <code>optional int32 total_byteps = 1;</code> */ int getTotalByteps(); /** * <code>optional int64 flow_start_time = 2;</code> */ boolean hasFlowStartTime(); /** * <code>optional int64 flow_start_time = 2;</code> */ long getFlowStartTime(); /** * <code>optional int64 date = 3;</code> */ boolean hasDate(); /** * <code>optional int64 date = 3;</code> */ long getDate(); /** * <code>optional int64 server_total_packet = 4;</code> */ boolean hasServerTotalPacket(); /** * <code>optional int64 server_total_packet = 4;</code> */ long getServerTotalPacket(); /** * <code>optional int64 client_total_byte = 5;</code> */ boolean hasClientTotalByte(); /** * <code>optional int64 client_total_byte = 5;</code> */ long getClientTotalByte(); /** * <code>optional int32 link_id = 6;</code> */ boolean hasLinkId(); /** * <code>optional int32 link_id = 6;</code> */ int getLinkId(); /** * <code>optional int64 total_byte = 7;</code> */ boolean hasTotalByte(); /** * <code>optional int64 total_byte = 7;</code> */ long getTotalByte(); /** * <code>optional int64 flow_end_time = 8;</code> */ boolean hasFlowEndTime(); /** * <code>optional int64 flow_end_time = 8;</code> */ long getFlowEndTime(); /** * <code>optional int32 client_port = 9;</code> */ boolean hasClientPort(); /** * <code>optional int32 client_port = 9;</code> */ int getClientPort(); /** * <code>optional int32 protocol = 10;</code> */ boolean hasProtocol(); /** * <code>optional int32 protocol = 10;</code> */ int getProtocol(); /** * <code>optional int64 total_packet = 11;</code> */ boolean hasTotalPacket(); /** * <code>optional int64 total_packet = 11;</code> */ long getTotalPacket(); /** * <code>optional int64 flow_duration = 12;</code> */ boolean hasFlowDuration(); /** * <code>optional int64 flow_duration = 12;</code> */ long getFlowDuration(); /** * <code>optional string id = 13;</code> */ boolean hasId(); /** * <code>optional string id = 13;</code> */ java.lang.String getId(); /** * <code>optional string id = 13;</code> */ com.google.protobuf.ByteString getIdBytes(); /** * <code>optional string server_ip_addr = 14;</code> */ boolean hasServerIpAddr(); /** * <code>optional string server_ip_addr = 14;</code> */ java.lang.String getServerIpAddr(); /** * <code>optional string server_ip_addr = 14;</code> */ com.google.protobuf.ByteString getServerIpAddrBytes(); /** * <code>optional string direction_mask = 15;</code> */ boolean hasDirectionMask(); /** * <code>optional string direction_mask = 15;</code> */ java.lang.String getDirectionMask(); /** * <code>optional string direction_mask = 15;</code> */ com.google.protobuf.ByteString getDirectionMaskBytes(); /** * <code>optional int32 app = 16;</code> */ boolean hasApp(); /** * <code>optional int32 app = 16;</code> */ int getApp(); /** * <code>optional int32 client_country_id = 17;</code> */ boolean hasClientCountryId(); /** * <code>optional int32 client_country_id = 17;</code> */ int getClientCountryId(); /** * <code>optional int32 client_netsegment_id = 18;</code> */ boolean hasClientNetsegmentId(); /** * <code>optional int32 client_netsegment_id = 18;</code> */ int getClientNetsegmentId(); /** * <code>optional int64 client_total_packet = 19;</code> */ boolean hasClientTotalPacket(); /** * <code>optional int64 client_total_packet = 19;</code> */ long getClientTotalPacket(); /** * <code>optional string client_ip_addr = 20;</code> */ boolean hasClientIpAddr(); /** * <code>optional string client_ip_addr = 20;</code> */ java.lang.String getClientIpAddr(); /** * <code>optional string client_ip_addr = 20;</code> */ com.google.protobuf.ByteString getClientIpAddrBytes(); /** * <code>optional int32 tcp_status = 21;</code> */ boolean hasTcpStatus(); /** * <code>optional int32 tcp_status = 21;</code> */ int getTcpStatus(); /** * <code>optional int32 server_country_id = 22;</code> */ boolean hasServerCountryId(); /** * <code>optional int32 server_country_id = 22;</code> */ int getServerCountryId(); /** * <code>optional int32 server_netsegment_id = 23;</code> */ boolean hasServerNetsegmentId(); /** * <code>optional int32 server_netsegment_id = 23;</code> */ int getServerNetsegmentId(); /** * <code>optional int64 avg_pkt_size = 24;</code> */ boolean hasAvgPktSize(); /** * <code>optional int64 avg_pkt_size = 24;</code> */ long getAvgPktSize(); /** * <code>optional int32 server_port = 25;</code> */ boolean hasServerPort(); /** * <code>optional int32 server_port = 25;</code> */ int getServerPort(); /** * <code>optional int64 server_total_byte = 26;</code> */ boolean hasServerTotalByte(); /** * <code>optional int64 server_total_byte = 26;</code> */ long getServerTotalByte(); /** * <code>optional int32 total_packetps = 27;</code> */ boolean hasTotalPacketps(); /** * <code>optional int32 total_packetps = 27;</code> */ int getTotalPacketps(); } /** * Protobuf type {@code TCPLog} */ public static final class TCPLog extends com.google.protobuf.GeneratedMessageV3 implements // @@protoc_insertion_point(message_implements:TCPLog) TCPLogOrBuilder { // Use TCPLog.newBuilder() to construct. private TCPLog(com.google.protobuf.GeneratedMessageV3.Builder<?> builder) { super(builder); } private TCPLog() { totalByteps_ = 0; flowStartTime_ = 0L; date_ = 0L; serverTotalPacket_ = 0L; clientTotalByte_ = 0L; linkId_ = 0; totalByte_ = 0L; flowEndTime_ = 0L; clientPort_ = 0; protocol_ = 0; totalPacket_ = 0L; flowDuration_ = 0L; id_ = ""; serverIpAddr_ = ""; directionMask_ = ""; app_ = 0; clientCountryId_ = 0; clientNetsegmentId_ = 0; clientTotalPacket_ = 0L; clientIpAddr_ = ""; tcpStatus_ = 0; serverCountryId_ = 0; serverNetsegmentId_ = 0; avgPktSize_ = 0L; serverPort_ = 0; serverTotalByte_ = 0L; totalPacketps_ = 0; } @java.lang.Override public final com.google.protobuf.UnknownFieldSet getUnknownFields() { return this.unknownFields; } private TCPLog( com.google.protobuf.CodedInputStream input, com.google.protobuf.ExtensionRegistryLite extensionRegistry) throws com.google.protobuf.InvalidProtocolBufferException { this(); int mutable_bitField0_ = 0; com.google.protobuf.UnknownFieldSet.Builder unknownFields = com.google.protobuf.UnknownFieldSet.newBuilder(); try { boolean done = false; while (!done) { int tag = input.readTag(); switch (tag) { case 0: done = true; break; default: { if (!parseUnknownField(input, unknownFields, extensionRegistry, tag)) { done = true; } break; } case 8: { bitField0_ |= 0x00000001; totalByteps_ = input.readInt32(); break; } case 16: { bitField0_ |= 0x00000002; flowStartTime_ = input.readInt64(); break; } case 24: { bitField0_ |= 0x00000004; date_ = input.readInt64(); break; } case 32: { bitField0_ |= 0x00000008; serverTotalPacket_ = input.readInt64(); break; } case 40: { bitField0_ |= 0x00000010; clientTotalByte_ = input.readInt64(); break; } case 48: { bitField0_ |= 0x00000020; linkId_ = input.readInt32(); break; } case 56: { bitField0_ |= 0x00000040; totalByte_ = input.readInt64(); break; } case 64: { bitField0_ |= 0x00000080; flowEndTime_ = input.readInt64(); break; } case 72: { bitField0_ |= 0x00000100; clientPort_ = input.readInt32(); break; } case 80: { bitField0_ |= 0x00000200; protocol_ = input.readInt32(); break; } case 88: { bitField0_ |= 0x00000400; totalPacket_ = input.readInt64(); break; } case 96: { bitField0_ |= 0x00000800; flowDuration_ = input.readInt64(); break; } case 106: { com.google.protobuf.ByteString bs = input.readBytes(); bitField0_ |= 0x00001000; id_ = bs; break; } case 114: { com.google.protobuf.ByteString bs = input.readBytes(); bitField0_ |= 0x00002000; serverIpAddr_ = bs; break; } case 122: { com.google.protobuf.ByteString bs = input.readBytes(); bitField0_ |= 0x00004000; directionMask_ = bs; break; } case 128: { bitField0_ |= 0x00008000; app_ = input.readInt32(); break; } case 136: { bitField0_ |= 0x00010000; clientCountryId_ = input.readInt32(); break; } case 144: { bitField0_ |= 0x00020000; clientNetsegmentId_ = input.readInt32(); break; } case 152: { bitField0_ |= 0x00040000; clientTotalPacket_ = input.readInt64(); break; } case 162: { com.google.protobuf.ByteString bs = input.readBytes(); bitField0_ |= 0x00080000; clientIpAddr_ = bs; break; } case 168: { bitField0_ |= 0x00100000; tcpStatus_ = input.readInt32(); break; } case 176: { bitField0_ |= 0x00200000; serverCountryId_ = input.readInt32(); break; } case 184: { bitField0_ |= 0x00400000; serverNetsegmentId_ = input.readInt32(); break; } case 192: { bitField0_ |= 0x00800000; avgPktSize_ = input.readInt64(); break; } case 200: { bitField0_ |= 0x01000000; serverPort_ = input.readInt32(); break; } case 208: { bitField0_ |= 0x02000000; serverTotalByte_ = input.readInt64(); break; } case 216: { bitField0_ |= 0x04000000; totalPacketps_ = input.readInt32(); break; } } } } catch (com.google.protobuf.InvalidProtocolBufferException e) { throw e.setUnfinishedMessage(this); } catch (java.io.IOException e) { throw new com.google.protobuf.InvalidProtocolBufferException( e).setUnfinishedMessage(this); } finally { this.unknownFields = unknownFields.build(); makeExtensionsImmutable(); } } public static final com.google.protobuf.Descriptors.Descriptor getDescriptor() { return TCPLogOuterClass.internal_static_TCPLog_descriptor; } protected com.google.protobuf.GeneratedMessageV3.FieldAccessorTable internalGetFieldAccessorTable() { return TCPLogOuterClass.internal_static_TCPLog_fieldAccessorTable .ensureFieldAccessorsInitialized( TCPLogOuterClass.TCPLog.class, TCPLogOuterClass.TCPLog.Builder.class); } private int bitField0_; public static final int TOTAL_BYTEPS_FIELD_NUMBER = 1; private int totalByteps_; /** * <code>optional int32 total_byteps = 1;</code> */ public boolean hasTotalByteps() { return ((bitField0_ & 0x00000001) == 0x00000001); } /** * <code>optional int32 total_byteps = 1;</code> */ public int getTotalByteps() { return totalByteps_; } public static final int FLOW_START_TIME_FIELD_NUMBER = 2; private long flowStartTime_; /** * <code>optional int64 flow_start_time = 2;</code> */ public boolean hasFlowStartTime() { return ((bitField0_ & 0x00000002) == 0x00000002); } /** * <code>optional int64 flow_start_time = 2;</code> */ public long getFlowStartTime() { return flowStartTime_; } public static final int DATE_FIELD_NUMBER = 3; private long date_; /** * <code>optional int64 date = 3;</code> */ public boolean hasDate() { return ((bitField0_ & 0x00000004) == 0x00000004); } /** * <code>optional int64 date = 3;</code> */ public long getDate() { return date_; } public static final int SERVER_TOTAL_PACKET_FIELD_NUMBER = 4; private long serverTotalPacket_; /** * <code>optional int64 server_total_packet = 4;</code> */ public boolean hasServerTotalPacket() { return ((bitField0_ & 0x00000008) == 0x00000008); } /** * <code>optional int64 server_total_packet = 4;</code> */ public long getServerTotalPacket() { return serverTotalPacket_; } public static final int CLIENT_TOTAL_BYTE_FIELD_NUMBER = 5; private long clientTotalByte_; /** * <code>optional int64 client_total_byte = 5;</code> */ public boolean hasClientTotalByte() { return ((bitField0_ & 0x00000010) == 0x00000010); } /** * <code>optional int64 client_total_byte = 5;</code> */ public long getClientTotalByte() { return clientTotalByte_; } public static final int LINK_ID_FIELD_NUMBER = 6; private int linkId_; /** * <code>optional int32 link_id = 6;</code> */ public boolean hasLinkId() { return ((bitField0_ & 0x00000020) == 0x00000020); } /** * <code>optional int32 link_id = 6;</code> */ public int getLinkId() { return linkId_; } public static final int TOTAL_BYTE_FIELD_NUMBER = 7; private long totalByte_; /** * <code>optional int64 total_byte = 7;</code> */ public boolean hasTotalByte() { return ((bitField0_ & 0x00000040) == 0x00000040); } /** * <code>optional int64 total_byte = 7;</code> */ public long getTotalByte() { return totalByte_; } public static final int FLOW_END_TIME_FIELD_NUMBER = 8; private long flowEndTime_; /** * <code>optional int64 flow_end_time = 8;</code> */ public boolean hasFlowEndTime() { return ((bitField0_ & 0x00000080) == 0x00000080); } /** * <code>optional int64 flow_end_time = 8;</code> */ public long getFlowEndTime() { return flowEndTime_; } public static final int CLIENT_PORT_FIELD_NUMBER = 9; private int clientPort_; /** * <code>optional int32 client_port = 9;</code> */ public boolean hasClientPort() { return ((bitField0_ & 0x00000100) == 0x00000100); } /** * <code>optional int32 client_port = 9;</code> */ public int getClientPort() { return clientPort_; } public static final int PROTOCOL_FIELD_NUMBER = 10; private int protocol_; /** * <code>optional int32 protocol = 10;</code> */ public boolean hasProtocol() { return ((bitField0_ & 0x00000200) == 0x00000200); } /** * <code>optional int32 protocol = 10;</code> */ public int getProtocol() { return protocol_; } public static final int TOTAL_PACKET_FIELD_NUMBER = 11; private long totalPacket_; /** * <code>optional int64 total_packet = 11;</code> */ public boolean hasTotalPacket() { return ((bitField0_ & 0x00000400) == 0x00000400); } /** * <code>optional int64 total_packet = 11;</code> */ public long getTotalPacket() { return totalPacket_; } public static final int FLOW_DURATION_FIELD_NUMBER = 12; private long flowDuration_; /** * <code>optional int64 flow_duration = 12;</code> */ public boolean hasFlowDuration() { return ((bitField0_ & 0x00000800) == 0x00000800); } /** * <code>optional int64 flow_duration = 12;</code> */ public long getFlowDuration() { return flowDuration_; } public static final int ID_FIELD_NUMBER = 13; private volatile java.lang.Object id_; /** * <code>optional string id = 13;</code> */ public boolean hasId() { return ((bitField0_ & 0x00001000) == 0x00001000); } /** * <code>optional string id = 13;</code> */ public java.lang.String getId() { java.lang.Object ref = id_; if (ref instanceof java.lang.String) { return (java.lang.String) ref; } else { com.google.protobuf.ByteString bs = (com.google.protobuf.ByteString) ref; java.lang.String s = bs.toStringUtf8(); if (bs.isValidUtf8()) { id_ = s; } return s; } } /** * <code>optional string id = 13;</code> */以上是关于序列化之protobuf与avro对比(Java)的主要内容，如果未能解决你的问题，请参考以下文章
Netty_05_六种序列化方式(JavaIO序列化 XML序列化 Hessian序列化 JSON序列化 Protobuf序列化 AVRO序列化)(实践类)
Netty_05_六种序列化方式(JavaIO序列化 XML序列化 Hessian序列化 JSON序列化 Protobuf序列化 AVRO序列化)(实践类)