Binary Serializers

Binary formats sacrifice human readability for significant improvements in size (20-50% smaller than JSON) and parsing speed.

When to Use Binary

Choose binary when: - Internal microservices where you control all endpoints - High-throughput data pipelines - Caching layers (Redis with binary values) - Mobile apps where bandwidth affects battery life - Network protocols where every byte matters

Common Characteristics

Aspect	Value
Human-readable	No
Schema	Optional
Size	30-50% smaller than JSON
Speed	Faster than JSON
Debugging	Requires hex dumps or tools

Format Comparison

Feature	MessagePack	CBOR
Standard	De facto standard	IETF RFC 8949
Speed	Faster (C extensions)	Slower (pure Python)
Size	Small	Often smallest
Datetime	Extension type	Native tags
Streaming	Chunks	Indefinite length
Popularity	Very high	Growing

C# Binary Serializers

Serializer	Format	Best For	Speed
GroBuf	Custom binary	Maximum speed	★★★★★
MS Bond Fast	Bond	.NET ecosystem	★★★★★
MS Bond Compact	Bond	Balanced	★★★★☆
MemoryPack	Custom	Zero-copy	★★★★☆
NetSerializer	Custom	General use	★★★★☆
FlatSharp	FlatBuffers	Zero-allocation	★★★★☆
ProtoBuf	Protocol Buffers	Cross-language	★★★★☆
MessagePack-CSharp	MessagePack	Wide support	★★★★☆
Hyperion	Custom	Akka.NET	★★★☆☆
Ceras	Custom	Ease of use	★★★☆☆
Migrant	Custom	Complex graphs	★★☆☆☆
FsPickler	Custom	F# integration	★★★☆☆
MS Binary	BinaryFormatter legacy	Legacy only	★★☆☆☆

Example: MemoryPack

// Serialize to byte array
byte[] bytes = MemoryPackSerializer.Serialize(person);

// Deserialize
Person person = MemoryPackSerializer.Deserialize<Person>(bytes);

Example: MessagePack

// Serialize
byte[] bytes = MessagePackSerializer.Serialize(person);

// Deserialize
Person person = MessagePackSerializer.Deserialize<Person>(bytes);

Python Binary Serializers

Serializer	Format	Best For	Speed
msgpack	MessagePack	Speed + ecosystem	★★★★☆
cbor2	CBOR	Standards compliance	★★★☆☆

Example: msgpack

import msgpack

# Serialize
packed = msgpack.packb(person, use_bin_type=True)

# Deserialize
person = msgpack.unpackb(packed, raw=False)

Example: cbor2

import cbor2

# Serialize (handles datetime natively)
encoded = cbor2.dumps(person, datetime_as_timestamp=True)

# Deserialize
person = cbor2.loads(encoded)

Performance Comparison

C# Binary Leaders

Serializer	Ops/Sec (Person)	Size (Person)
MS Bond Fast	79,563	~400 bytes
NetSerializer	35,785	~600 bytes
MemoryPack	59,444*	~500 bytes*
ProtoBuf	22,735	665 bytes

*MemoryPack: tested on SimpleObject (Person not supported in benchmark)

Python Binary

Serializer	Ops/Sec (Person)	Size (Person)
msgpack	2,658	1,631 bytes
cbor2	2,113	~1,400 bytes

Note: C# binary serializers generally outperform Python due to value types and struct optimizations.

Binary vs JSON Trade-offs

Aspect	Binary (MessagePack)	JSON
Size	30-50% smaller	Larger
Speed	Faster parsing	Slower (text parsing)
Readability	None	Human-readable
Debugging	Harder	Easier
Schema	None	None
Cross-language	Wide support	Universal

Type Handling

MessagePack Type Mapping

Python/C#	MessagePack	Notes
int	Integer	Variable encoding
float	Float	IEEE 754
str/string	String	UTF-8
bytes/byte[]	Binary	Native support
list/List	Array	Variable length
dict/Dictionary	Map	Key-value pairs
datetime	Extension	Needs custom handling
tuple	Array	Becomes list in Python

CBOR Type Mapping

Type	CBOR	Notes
int	Integer	Smaller encoding than MessagePack
float	Float/IEEE 754	Multiple precision options
str	Text string	UTF-8
bytes	Byte string	Native
datetime	Tag 0/1	Native support
array	Array	Definite/indefinite length

When to Choose Which

Scenario	Recommendation
Maximum speed	MessagePack (C#) or GroBuf
Standards compliance	CBOR (RFC 8949)
IoT/embedded	CBOR (smallest)
Existing ecosystem	MessagePack (more libraries)
Streaming large data	CBOR (indefinite length)

Security Considerations

MessagePack: Can deserialize arbitrary objects if object_hook is misused
CBOR: Similar risks with custom decoders
Best practice: Use default safe modes, validate after deserialization