• PKI systems: Often used Base64-encoded certificates embedded in XML or JSON, but the encoding was not composable
• Blockchain protocols: Typically used pure binary serialization, making debugging difficult
• Web APIs: Used JSON for data structures but had no standard way to embed cryptographic primitives efficiently

The problem became more acute with the rise of verifiable credentials and decentralized identity systems, which needed to:

• Serialize complex data structures (claims, schemas, metadata)
• Include cryptographic commitments (signatures, digests, proofs)
• Support both human inspection and efficient transmission
• Enable streaming protocols without buffering entire messages

Earlier attempts at mixed serialization typically used wrapper envelopes - outer structures that contained both data and cryptographic material as separate sections. However, this approach:

• Required parsing the entire envelope before accessing content
• Created disposable wrapper overhead
• Made stream processing inefficient
• Prevented parallel processing of attachments

KERI's Approach

KERI's solution through CESR interleaved serialization represents a fundamental architectural shift. Rather than wrapping different formats in containers, CESR enables direct concatenation of different serialization types within a single stream, using count codes as intelligent boundaries.

Count Code Architecture

CESR defines special group framing codes (count codes) that serve multiple purposes:

1. Format Identification: The count code indicates what serialization format follows (CESR primitives, JSON, CBOR, or MGPK)
2. Length Specification: The code encodes how many bytes or characters to extract for that section
3. Boundary Marking: Provides unambiguous transition points between formats
4. Self-Framing: Enables extraction without parsing the content itself

These count codes maintain CESR's critical 24-bit alignment constraint - all primitives and groups align on boundaries that are integer multiples of both 4 Base64 characters (text domain) and 3 bytes (binary domain). This alignment ensures composability is preserved even when interleaving different formats.

Practical Implementation Pattern

A typical interleaved CESR stream might contain:

[CESR Count Code for JSON][JSON field map with data][CESR Count Code for primitives][CESR-encoded signatures][CESR Count Code for CBOR][CBOR-encoded metadata]

Each section is independently parseable, yet the entire stream maintains CESR's composability property - it can be converted between text and binary domains en masse without losing the ability to separate individual sections.

Field Map Integration

The interleaving capability is particularly powerful for field maps (dictionaries/hash tables) that contain both:

• Data attributes: Names, dates, claims, schemas - naturally suited to JSON/CBOR/MGPK
• Cryptographic commitments: SAIDs, signatures, digests - naturally suited to CESR encoding

Rather than forcing all content into CESR encoding (which would be inefficient for structured data) or forcing cryptographic material into JSON strings (which loses type information and compactness), interleaved serialization allows each to use its optimal format.

Cold Start Recovery

A critical advantage of interleaved serialization is support for cold start stream parsing. When a parser encounters errors or needs to resynchronize:

• Count codes serve as recovery points
• The parser can skip to the next count code boundary
• No need to flush buffers or restart from the beginning
• Malformed sections can be isolated without corrupting the entire stream

This robustness is essential for distributed systems where network interruptions, partial transmissions, or corrupted data may occur.

ACDC Integration

In ACDC (Authentic Chained Data Container) credentials, interleaved serialization enables:

• Attribute sections encoded in JSON for human readability and schema validation
• SAID commitments encoded in CESR for cryptographic efficiency
• Signature attachments encoded in CESR for compact transmission
• Edge references to other ACDCs using CESR-encoded identifiers

The entire ACDC can be transmitted as a single interleaved stream, with each component using its most appropriate encoding, yet the whole structure remains verifiable and composable.

Practical Implications

Use Cases

Verifiable Credentials: ACDCs use interleaved serialization to combine human-readable claims (JSON) with cryptographic proofs (CESR). A credential might contain:

• Issuer and subject identifiers (CESR-encoded AIDs)
• Claim attributes (JSON field map)
• Schema references (CESR-encoded SAIDs)
• Digital signatures (CESR-encoded signature attachments)

Key Event Logs: KELs can include:

• Event data (JSON/CBOR for flexibility)
• Cryptographic commitments (CESR for efficiency)
• Witness receipts (CESR-encoded signatures)
• Seals to external data (CESR-encoded digests)

Streaming Protocols: Network protocols benefit from:

• Efficient binary transmission of cryptographic material
• Human-readable logging and debugging of data structures
• Parallel processing of attachments without parsing message bodies
• Graceful degradation when encountering unknown formats

Benefits

Format Optimization: Each type of content uses its most appropriate encoding:

• Cryptographic primitives: CESR (compact, type-safe, self-framing)
• Structured data: JSON/CBOR/MGPK (flexible, schema-compatible, tooling support)
• No forced conversion to a single format that compromises both

Tooling Compatibility: Applications can:

• Use standard JSON/CBOR/MGPK libraries for data structures
• Use CESR libraries for cryptographic primitives
• Avoid implementing custom serialization for everything

Stream Processing Efficiency: Parsers can:

• Extract sections without full parsing
• Process cryptographic attachments in parallel
• Skip unknown sections without breaking
• Recover from errors at count code boundaries

Human Readability: When using text domain encoding:

• Data structures remain readable in JSON
• Cryptographic material is still compact and type-identified
• Logs and audit trails are comprehensible
• Debugging is practical

Trade-offs

Parser Complexity: Implementations must:

• Support multiple serialization formats
• Handle count code interpretation
• Maintain state across format transitions
• Implement recovery mechanisms

Format Proliferation: Systems must decide:

• Which formats to support (JSON, CBOR, MGPK, or all three)
• How to handle unknown formats in count codes
• Whether to enforce format restrictions for specific use cases

Alignment Overhead: Count codes add:

• Small overhead for format boundaries
• Padding requirements to maintain 24-bit alignment
• Complexity in calculating section lengths

However, these trade-offs are generally favorable compared to alternatives:

• Wrapper envelopes add more overhead and prevent stream processing
• Single-format encoding forces suboptimal representations
• Non-composable encodings prevent efficient domain conversion

Security Considerations

Interleaved serialization maintains CESR's security properties:

• Type safety: Count codes prevent type confusion attacks
• Boundary integrity: Self-framing prevents buffer overflow exploits
• Cryptographic binding: SAIDs and signatures cover entire structures including format boundaries
• Denial of service resistance: Length limits in count codes prevent resource exhaustion

The ability to skip malformed sections also provides resilience against malicious inputs - a corrupted JSON section doesn't prevent processing of valid CESR primitives that follow.

Conclusion

Interleaved serialization represents a sophisticated solution to the multi-format encoding challenge in cryptographic protocols. By enabling different serialization types to coexist within a single composable stream, CESR provides the flexibility needed for modern verifiable data structures while maintaining the efficiency and security properties required for production systems. This capability is fundamental to KERI's architecture and enables the practical implementation of ACDCs, KELs, and other verifiable data structures that combine human-readable content with cryptographic commitments.