Skip to main content

Fury Java Serialization Format

Spec overview

Fury Java Serialization is an automatic object serialization framework that supports reference and polymorphism. Fury will convert an object from/to fury java serialization binary format. Fury has two core concepts for java serialization:

  • Fury Java Binary format
  • Framework to convert object to/from Fury Java Binary format

The serialization format is a dynamic binary format. The dynamics and reference/polymorphism support make Fury flexible, much more easy to use, but also introduce more complexities compared to static serialization frameworks. So the format will be more complex.

Here is the overall format:

| fury header | object ref meta | object class meta | object value data |

The data are serialized using little endian byte order overall. If bytes swap is costly for some object, Fury will write the byte order for that object into the data instead of converting it to little endian.

Fury header

Fury header consists starts one byte:

|     4 bits    | 1 bit | 1 bit | 1 bit  | 1 bit |          optional 4 bytes          |
+---------------+-------+-------+--------+-------+------------------------------------+
| reserved bits | oob | xlang | endian | null | unsigned int for meta start offset |
  • null flag: 1 when object is null, 0 otherwise. If an object is null, other bits won't be set.
  • endian flag: 1 when data is encoded by little endian, 0 for big endian.
  • xlang flag: 1 when serialization uses xlang format, 0 when serialization uses Fury java format.
  • oob flag: 1 when passed BufferCallback is not null, 0 otherwise.

If meta share mode is enabled, an uncompressed unsigned int is appended to indicate the start offset of metadata.

Reference Meta

Reference tracking handles whether the object is null, and whether to track reference for the object by writing corresponding flags and maintaining internal state.

Reference flags:

FlagByte ValueDescription
NULL FLAG-3This flag indicates the object is a null value. We don't use another byte to indicate REF, so that we can save one byte.
REF FLAG-2This flag indicates the object is already serialized previously, and fury will write a ref id with unsigned varint format instead of serialize it again
NOT_NULL VALUE FLAG-1This flag indicates the object is a non-null value and fury doesn't track ref for this type of object.
REF VALUE FLAG0This flag indicates the object is referencable and the first time to serialize.

When reference tracking is disabled globally or for specific types, or for certain types within a particular context(e.g., a field of a class), only the NULL and NOT_NULL VALUE flags will be used for reference meta.

Class Meta

Fury supports to register class by an optional id, the registration can be used for security check and class identification. If a class is registered, it will have a user-provided or an auto-growing unsigned int i.e. class_id.

Depending on whether meta share mode and registration is enabled for current class, Fury will write class meta differently.

Schema consistent

If schema consistent mode is enabled globally or enabled for current class, class meta will be written as follows:

  • If class is registered, it will be written as a fury unsigned varint: class_id << 1.
  • If class is not registered:
    • If class is not an array, fury will write one byte 0bxxxxxxx1 first, then write class name.
      • The first little bit is 1, which is different from first bit 0 of encoded class id. Fury can use this information to determine whether to read class by class id for deserialization.
    • If class is not registered and class is an array, fury will write one byte dimensions << 1 | 1 first, then write component class subsequently. This can reduce array class name cost if component class is or will be serialized.
    • Class will be written as two enumerated fury unsigned by default: package name and class name. If meta share mode is enabled, class will be written as an unsigned varint which points to index in MetaContext.

Schema evolution

If schema evolution mode is enabled globally or enabled for current class, class meta will be written as follows:

  • If meta share mode is not enabled, class meta will be written as schema consistent mode. Additionally, field meta such as field type and name will be written with the field value using a key-value like layout.
  • If meta share mode is enabled, class meta will be written as a meta-share encoded binary if class hasn't been written before, otherwise an unsigned varint id which references to previous written class meta will be written.

Meta share

This mode will forbid streaming writing since it needs to look back for update the start offset after the whole object graph writing and meta collecting is finished. Only in this way we can ensure deserialization failure doesn't lost shared meta. Meta streamline will be supported in the future for enclosed meta sharing which doesn't cross multiple serializations of different objects.

For Schema consistent mode, class will be encoded as an enumerated string by full class name. Here we mainly describe the meta layout for schema evolution mode:

|      8 bytes meta header      | meta size |   variable bytes   |  variable bytes   | variable bytes |
+-------------------------------+-----------|--------------------+-------------------+----------------+
| 7 bytes hash + 1 bytes header | 1~2 bytes | current class meta | parent class meta | ... |

Class meta are encoded from parent class to leaf class, only class with serializable fields will be encoded.

Meta header

Meta header is a 64 bits number value encoded in little endian order.

  • Lowest 4 digits 0b0000~0b1110 are used to record num classes. 0b1111 is preserved to indicate that Fury need to read more bytes for length using Fury unsigned int encoding. If current class doesn't has parent class, or parent class doesn't have fields to serialize, or we're in a context which serialize fields of current class only( ObjectStreamSerializer#SlotInfo is an example), num classes will be 1.
  • 5rd bit is used to indicate whether this class needs schema evolution.
  • 6rd bit is used to indicate whether the size sum of all layers meta is less than 256.
  • Other 56 bits is used to store the unique hash of flags + all layers class meta.

Meta size

  • If the size sum of all layers meta is less than 256, then one byte is written next to indicate the length of meta.
  • Otherwise, write size as two bytes in little endian.

Single layer class meta

|      unsigned varint       |      meta string      |     meta string     |  field info: variable bytes   | variable bytes  | ... |
+----------------------------+-----------------------+---------------------+-------------------------------+-----------------+-----+
| num fields + register flag | header + package name | header + class name | header + type id + field name | next field info | ... |
  • num fields: encode num fields << 1 | register flag(1 when class registered) as unsigned varint.
    • If class is registered, then an unsigned varint class id will be written next, package and class name will be omitted.
    • If current class is schema consistent, then num field will be 0 to flag it.
    • If current class isn't schema consistent, then num field will be the number of compatible fields. For example, users can use tag id to mark some field as compatible field in schema consistent context. In such cases, schema consistent fields will be serialized first, then compatible fields will be serialized next. At deserialization, Fury will use fields info of those fields which aren't annotated by tag id for deserializing schema consistent fields, then use fields info in meta for deserializing compatible fields.
  • Package name encoding(omitted when class is registered):
    • encoding algorithm: UTF8/ALL_TO_LOWER_SPECIAL/LOWER_UPPER_DIGIT_SPECIAL
    • Header: 6 bits size | 2 bits encoding flags. The 6 bits size: 0~63 will be used to indicate size 0~63, the value 63 the size need more byte to read, the encoding will encode size - 63 as a varint next.
  • Class name encoding(omitted when class is registered):
    • encoding algorithm: UTF8/LOWER_UPPER_DIGIT_SPECIAL/FIRST_TO_LOWER_SPECIAL/ALL_TO_LOWER_SPECIAL
    • header: 6 bits size | 2 bits encoding flags. The 6 bits size: 0~63 will be used to indicate size 0~63, the value 63 the size need more byte to read, the encoding will encode size - 63 as a varint next.
  • Field info:
    • header(8 bits): 3 bits size + 2 bits field name encoding + polymorphism flag + nullability flag + ref tracking flag. Users can use annotation to provide those info.
      • 2 bits field name encoding:
        • encoding: UTF8/ALL_TO_LOWER_SPECIAL/LOWER_UPPER_DIGIT_SPECIAL/TAG_ID
        • If tag id is used, i.e. field name is written by an unsigned varint tag id. 2 bits encoding will be 11.
      • size of field name:
        • The 3 bits size: 0~7 will be used to indicate length 1~7, the value 6 the size read more bytes, the encoding will encode size - 7 as a varint next.
        • If encoding is TAG_ID, then num_bytes of field name will be used to store tag id.
      • ref tracking: when set to 1, ref tracking will be enabled for this field.
      • nullability: when set to 1, this field can be null.
      • polymorphism: when set to 1, the actual type of field will be the declared field type even the type if not final.
    • type id:
      • For registered type-consistent classes, it will be the registered class id.
      • Otherwise it will be encoded as OBJECT_ID if it isn't final and FINAL_OBJECT_ID if it's final. The meta for such types is written separately instead of inlining here is to reduce meta space cost if object of this type is serialized in current object graph multiple times, and the field value may be null too.
    • Field name: If type id is set, type id will be used instead. Otherwise meta string encoding length and data will be written instead.

Field order are left as implementation details, which is not exposed to specification, the deserialization need to resort fields based on Fury field comparator. In this way, fury can compute statistics for field names or types and using a more compact encoding.

Other layers class meta

Same encoding algorithm as the previous layer except:

  • header + package name:
    • Header:
      • If package name has been written before: varint index + sharing flag(set) will be written
      • If package name hasn't been written before:
        • If meta string encoding is LOWER_SPECIAL and the length of encoded string <= 64, then header will be 6 bits size + encoding flag(set) + sharing flag(unset).
        • Otherwise, header will be 3 bits unset + 3 bits encoding flags + encoding flag(unset) + sharing flag(unset)

Meta String

Meta string is mainly used to encode meta strings such as class name and field names.

Encoding Algorithms

String binary encoding algorithm:

AlgorithmPatternDescription
LOWER_SPECIALa-z._$|every char is written using 5 bits, a-z: 0b00000~0b11001, ._$|: 0b11010~0b11101, prepend one bit at the start to indicate whether strip last char since last byte may have 7 redundant bits(1 indicates strip last char)
LOWER_UPPER_DIGIT_SPECIALa-zA-Z0~9._every char is written using 6 bits, a-z: 0b00000~0b11001, A-Z: 0b11010~0b110011, 0~9: 0b110100~0b111101, ._: 0b111110~0b111111, prepend one bit at the start to indicate whether strip last char since last byte may have 7 redundant bits(1 indicates strip last char)
UTF-8any charsUTF-8 encoding

Encoding flags:

Encoding FlagPatternEncoding Algorithm
LOWER_SPECIALevery char is in a-z._$|LOWER_SPECIAL
FIRST_TO_LOWER_SPECIALevery char is in a-z[c1,c2] except first char is upper casereplace first upper case char to lower case, then use LOWER_SPECIAL
ALL_TO_LOWER_SPECIALevery char is in a-zA-Z[c1,c2]replace every upper case char by | + lower case, then use LOWER_SPECIAL, use this encoding if it's smaller than Encoding LOWER_UPPER_DIGIT_SPECIAL
LOWER_UPPER_DIGIT_SPECIALevery char is in a-zA-Z[c1,c2]use LOWER_UPPER_DIGIT_SPECIAL encoding if it's smaller than Encoding FIRST_TO_LOWER_SPECIAL
UTF8any utf-8 charuse UTF-8 encoding
Compressionany utf-8 charlossless compression

Notes:

  • For package name encoding, c1,c2 should be ._; For field/type name encoding, c1,c2 should be _$;
  • Depending on cases, one can choose encoding flags + data jointly, uses 3 bits of first byte for flags and other bytes for data.

Shared meta string

The shared meta string format consists of header and encoded string binary. Header of encoded string binary will be inlined in shared meta header.

Header is written using little endian order, Fury can read this flag first to determine how to deserialize the data.

Write by data

If string hasn't been written before, the data will be written as follows:

| unsigned varint: string binary size + 1 bit: not written before | 56 bits: unique hash | 3 bits encoding flags + string binary |

If string binary size is less than 16 bytes, the hash will be omitted to save spaces. Unique hash can be omitted too if caller pass a flag to disable it. In such cases, the format will be:

| unsigned varint: string binary size + 1 bit: not written before  | 3 bits encoding flags + string binary |

Write by ref

If string has been written before, the data will be written as follows:

| unsigned varint: written string id + 1 bit: written before |

Value Format

Basic types

Bool

  • size: 1 byte
  • format: 0 for false, 1 for true

Byte

  • size: 1 byte
  • format: write as pure byte.

Short

  • size: 2 byte
  • byte order: little endian order

Char

  • size: 2 byte
  • byte order: little endian order

Unsigned int

  • size: 1~5 byte
  • Format: The most significant bit (MSB) in every byte indicates whether to have the next byte. If first bit is set i.e. b & 0x80 == 0x80, then the next byte should be read until the first bit of the next byte is unset.

Signed int

  • size: 1~5 byte
  • Format: First convert the number into positive unsigned int by (v << 1) ^ (v >> 31) ZigZag algorithm, then encoding it as an unsigned int.

Unsigned long

  • size: 1~9 byte
  • Fury PVL(Progressive Variable-length Long) Encoding:
    • positive long format: first bit in every byte indicates whether to have the next byte. If first bit is set i.e. b & 0x80 == 0x80, then the next byte should be read until the first bit is unset.

Signed long

  • size: 1~9 byte
  • Fury SLI(Small long as int) Encoding:
    • If long is in [-1073741824, 1073741823], encode as 4 bytes int: | little-endian: ((int) value) << 1 |
    • Otherwise write as 9 bytes: | 0b1 | little-endian 8 bytes long |
  • Fury PVL(Progressive Variable-length Long) Encoding:
    • First convert the number into positive unsigned long by (v << 1) ^ (v >> 63) ZigZag algorithm to reduce cost of small negative numbers, then encoding it as an unsigned long.

Float

  • size: 4 byte
  • format: convert float to 4 bytes int by Float.floatToRawIntBits, then write as binary by little endian order.

Double

  • size: 8 byte
  • format: convert double to 8 bytes int by Double.doubleToRawLongBits, then write as binary by little endian order.

String

Format:

| header: size << 2 | 2 bits encoding flags | binary data |
  • size + encoding will be concat as a long and encoded as an unsigned var long. The little 2 bits is used for encoding: 0 for latin, 1 for utf-16, 2 for utf-8.
  • encoded string binary data based on encoding: latin/utf-16/utf-8.

Which encoding to choose:

  • For JDK8: fury detect latin at runtime, if string is latin string, then use latin encoding, otherwise use utf-16.
  • For JDK9+: fury use coder in String object for encoding, latin/utf-16 will be used for encoding.
  • If the string is encoded by utf-8, then fury will use utf-8 to decode the data. But currently fury doesn't enable utf-8 encoding by default for java. Cross-language string serialization of fury uses utf-8 by default.

Collection

All collection serializers must extend AbstractCollectionSerializer.

Format:

length(unsigned varint) | collection header | elements header | elements data

Collection header

  • For ArrayList/LinkedArrayList/HashSet/LinkedHashSet, this will be empty.
  • For TreeSet, this will be Comparator
  • For subclass of ArrayList, this may be extra object field info.

Elements header

In most cases, all collection elements are same type and not null, elements header will encode those homogeneous information to avoid the cost of writing it for every element. Specifically, there are four kinds of information which will be encoded by elements header, each use one bit:

  • If track elements ref, use the first bit 0b1 of the header to flag it.
  • If the collection has null, use the second bit 0b10 of the header to flag it. If ref tracking is enabled for this element type, this flag is invalid.
  • If the collection element types are not declared type, use the 3rd bit 0b100 of the header to flag it.
  • If the collection element types are different, use the 4rd bit 0b1000 header to flag it.

By default, all bits are unset, which means all elements won't track ref, all elements are same type, not null and the actual element is the declared type in the custom class field.

The implementation can generate different deserialization code based read header, and look up the generated code from a linear map/list.

Elements data

Based on the elements header, the serialization of elements data may skip ref flag/null flag/element class info.

CollectionSerializer#write/read can be taken as an example.

Array

Primitive array

Primitive array are taken as a binary buffer, serialization will just write the length of array size as an unsigned int, then copy the whole buffer into the stream.

Such serialization won't compress the array. If users want to compress primitive array, users need to register custom serializers for such types.

Object array

Object array is serialized using the collection format. Object component type will be taken as collection element generic type.

Map

All Map serializers must extend AbstractMapSerializer.

Format:

| length(unsigned varint) | map header | key value pairs data |

Map header

  • For HashMap/LinkedHashMap, this will be empty.
  • For TreeMap, this will be Comparator
  • For other Map, this may be extra object field info.

Map Key-Value data

Map iteration is too expensive, Fury won't compute the header like for collection before since it introduce considerable overhead. Users can use MapFieldInfo annotation to provide header in advance. Otherwise Fury will use first key-value pair to predict header optimistically, and update the chunk header if the prediction failed at some pair.

Fury will serialize map chunk by chunk, every chunk has 127 pairs at most.

|    1 byte      |     1 byte     | variable bytes  |
+----------------+----------------+-----------------+
| chunk size: N | KV header | N*2 objects |

KV header:

  • If track key ref, use the first bit 0b1 of the header to flag it.
  • If the key has null, use the second bit 0b10 of the header to flag it. If ref tracking is enabled for this key type, this flag is invalid.
  • If the key types of map are different, use the 3rd bit 0b100 of the header to flag it.
  • If the actual key type of map is not the declared key type, use the 4rd bit 0b1000 of the header to flag it.
  • If track value ref, use the 5th bit 0b10000 of the header to flag it.
  • If the value has null, use the 6th bit 0b100000 of the header to flag it. If ref tracking is enabled for this value type, this flag is invalid.
  • If the value types of map are different, use the 7rd bit 0b1000000 header to flag it.
  • If the value type of map is not the declared value type, use the 8rd bit 0b10000000 of the header to flag it.

If streaming write is enabled, which means Fury can't update written chunk size. In such cases, map key-value data format will be:

|    1 byte      | variable bytes  |
+----------------+-----------------+
| KV header | N*2 objects |

KV header will be a header marked by MapFieldInfo in java. The implementation can generate different deserialization code based read header, and look up the generated code from a linear map/list.

Enum

Enums are serialized as an unsigned var int. If the order of enum values change, the deserialized enum value may not be the value users expect. In such cases, users must register enum serializer by make it write enum value as an enumerated string with unique hash disabled.

Object

Object means object of pojo/struct/bean/record type. Object will be serialized by writing its fields data in fury order.

Depending on schema compatibility, objects will have different formats.

Field order

Field will be ordered as following, every group of fields will have its own order:

  • primitive fields: larger size type first, smaller later, variable size type last.
  • boxed primitive fields: same order as primitive fields
  • final fields: same type together, then sorted by field name lexicographically.
  • collection fields: same order as final fields
  • map fields: same order as final fields
  • other fields: same order as final fields

Schema consistent

Object fields will be serialized one by one using following format:

Primitive field value:
| var bytes |
+----------------+
| value data |
+----------------+
Boxed field value:
| one byte | var bytes |
+-----------+---------------+
| null flag | field value |
+-----------+---------------+
field value of final type with ref tracking:
| var bytes | var objects |
+-----------+-------------+
| ref meta | value data |
+-----------+-------------+
field value of final type without ref tracking:
| one byte | var objects |
+-----------+-------------+
| null flag | field value |
+-----------+-------------+
field value of non-final type with ref tracking:
| one byte | var bytes | var objects |
+-----------+-------------+-------------+
| ref meta | class meta | value data |
+-----------+-------------+-------------+
field value of non-final type without ref tracking:
| one byte | var bytes | var objects |
+-----------+------------+------------+
| null flag | class meta | value data |
+-----------+------------+------------+

Schema evolution

Schema evolution have similar format as schema consistent mode for object except:

  • For this object type itself, schema consistent mode will write class by id/name, but schema evolution mode will write class field names, types and other meta too, see Class meta.
  • Class meta of final custom type needs to be written too, because peers may not have this class defined.

Class

Class will be serialized using class meta format.

Implementation guidelines

  • Try to merge multiple bytes into an int/long write before writing to reduce memory IO and bound check cost.
  • Read multiple bytes as an int/long, then split into multiple bytes to reduce memory IO and bound check cost.
  • Try to use one varint/long to write flags and length together to save one byte cost and reduce memory io.
  • Condition branches are less expensive compared to memory IO cost unless there are too many branches.