Enhanced Market Data Interface (EMDI)

The Enhanced Market Data Interface (EMDI) is Deutsche Boerse’s medium-latency market data feed that combines standardized FIX protocol encoding with aggregated order book data. EMDI provides a balanced solution for firms requiring comprehensive market data without the complexity of ultra-low-latency implementations, offering price-level aggregated order books with depths up to 15 levels.

Protocol Architecture

EMDI is built on industry-standard protocols, making it accessible to firms with existing market data infrastructure:

Protocol Stack:

FIX 5.0 Service Pack 2 (FIX 5.0 SP2)
FAST 1.1 and 1.2 (FIX Adapted for Streaming) encoding
UDP multicast delivery for efficient distribution
Standardized FIXML schema files for message validation

The use of standardized FIX/FAST protocols distinguishes EMDI from the proprietary EOBI format, allowing firms to leverage existing commercial and open-source libraries. Deutsche Boerse provides XML FAST template files supporting both FAST 1.1 and 1.2 versions, along with FIXML schema files for message structure validation.

Architectural Integration:

Since Release 12.0, EMDI has been fully consolidated into the Matching Engine (ME) process. This architectural change eliminated the separate Market Data Distributor (MDD) process, reducing the data path and improving latency characteristics. The consolidation means EMDI and EOBI now share source IP addresses for consolidated partitions, streamlining network configuration and reducing the latency gap between the two feeds.

Data Content and Structure

EMDI delivers un-netted, price-level aggregated order book data with configurable depth:

Order Book Depth:

Eurex markets: up to 15 price levels per side
Xetra markets: up to 10 price levels per side
Each level contains aggregated quantity at that price
Both bid and ask sides transmitted

Data Elements:

Aggregated quantities at each price level (not individual orders)
Best bid and ask prices with depth
Trade reports with price, quantity, and trade conditions
Market statistics and trading session information
Auction indicators and reference prices
Instrument status updates

The aggregation model significantly reduces data volume compared to EOBI’s order-by-order approach. While EOBI transmits every individual order add, modify, and delete, EMDI consolidates all orders at each price level into a single quantity figure. This makes EMDI more bandwidth-efficient but less granular for market microstructure analysis.

Snapshot and Incremental Update Model

EMDI employs a dual-channel architecture separating complete state snapshots from real-time incremental updates:

Snapshot Channel:

Provides complete order book state at regular intervals
Transmitted on dedicated multicast group separate from incrementals
Contains full depth for all subscribed instruments
Snapshot cycle frequency configurable per product group
Used for initial state acquisition and gap recovery

Incremental Channel:

Real-time updates reflecting order book changes
Minimal message size for efficiency
Only transmits changed price levels, not entire book
Sequential message numbering for gap detection

Synchronization Mechanism:

The synchronization between snapshots and incrementals relies on two critical sequence number fields:

MsgSeqNum: Sequence number of the current message
LastMsgSeqNumProcessed: In snapshots, indicates the last incremental update already reflected in the snapshot state

Recovery Process:

When a gap is detected in the incremental stream:

Request or wait for next snapshot message
Apply snapshot to establish known good state
Note the LastMsgSeqNumProcessed value from snapshot
Buffer any incrementals received during recovery
Apply buffered incrementals with MsgSeqNum > LastMsgSeqNumProcessed
Resume normal incremental processing

This model ensures data consistency while allowing efficient real-time updates during normal operation and reliable recovery during network disruptions or initial connection.

Latency Characteristics

EMDI occupies the middle ground in Deutsche Boerse’s market data latency spectrum:

Relative Latency Positioning:

Faster than MDI (Request/Response interface)
Slower than EOBI (ultra-low-latency feed)
Suitable for algorithmic trading, not optimal for market making

Release 12.0 Improvements:

The consolidation of EMDI into the Matching Engine process in Release 12.0 delivered measurable latency improvements:

Eliminated MDD process hop, reducing data path
EMDI and EOBI now generated from same source process
Shared source IP addresses for consolidated partitions
Gap between EOBI and EMDI latency narrowed

While exact latency figures vary by product and network conditions, typical EMDI latency ranges from several hundred microseconds to low milliseconds from matching event to client receipt. This positions EMDI as appropriate for strategies operating on millisecond to sub-second timescales, but not for microsecond-sensitive high-frequency trading.

Latency Contributors:

Aggregation processing within ME
FIX/FAST encoding overhead
Network transmission time
Client-side FAST decoding

Redundancy and Multicast Delivery

EMDI provides live-live redundancy through dual independent distribution paths:

Side A and Side B:

Two complete, independent multicast streams
Identical data transmitted simultaneously on both sides
Separate multicast IP addresses and port numbers
IGMPv2 protocol for multicast group management

Client Implementation Pattern:

Clients typically subscribe to both Side A and Side B simultaneously:

Receive messages from both sides in parallel
Process first message received for each sequence number
Discard duplicate when received from second side
Monitor both sides for health and failover capability

Benefits:

Zero failover time (already receiving both sides)
Protection against single network path failure
Protection against single source server failure
Ability to detect and diagnose network asymmetries

The live-live model eliminates the detection and switchover delay inherent in active-passive failover schemes, critical for maintaining continuous market data flow.

Pricing and Commercial Terms

EMDI pricing reflects its position as a comprehensive, standards-based market data solution:

Monthly Subscription Fees (CoLo 2.0 pricing):

Xetra EMDI: EUR 5,200 per month
Eurex EMDI: EUR 6,000 per month
Combined Xetra EMDI + EOBI: EUR 7,280 per month (bundled discount)

Connection Rebates:

Eurex dual connection (Side A + Side B): EUR 750 rebate per connection
Incentivizes proper redundant connection implementation

Cost Comparison:

EMDI less expensive than EOBI as standalone product
Combined EMDI + EOBI bundle offers cost efficiency
MDI lower cost but request/response model limits use cases

The bundled pricing for EMDI + EOBI reflects common deployment patterns where firms use EMDI for algorithmic strategies and EOBI for latency-critical market making, allowing cost-effective access to both feeds.

EMDI vs EOBI Comparison

Understanding the differences between EMDI and EOBI guides appropriate feed selection:

Protocol and Implementation:

EMDI: Standardized FIX 5.0 SP2 with FAST encoding; commercial libraries available
EOBI: Proprietary binary format; custom parser implementation required

Data Granularity:

EMDI: Aggregated price-level quantities; total volume at each price
EOBI: Individual order-by-order updates; every add, modify, delete visible

Order Book Depth:

EMDI: Limited to 10 (Xetra) or 15 (Eurex) price levels
EOBI: Full unlimited depth; entire order book visible

Latency:

EMDI: Medium latency; hundreds of microseconds to low milliseconds
EOBI: Ultra-low latency; tens of microseconds

Data Volume:

EMDI: Lower message rate due to aggregation
EOBI: Higher message rate; every order event transmitted

Use Case Fit:

EMDI: Algorithmic trading, DMA systems, visualization, strategies not requiring microsecond latency
EOBI: High-frequency trading, market making, strategies requiring full depth and minimal latency

Implementation Effort:

EMDI: Lower; leverage existing FIX/FAST infrastructure and libraries
EOBI: Higher; custom parser development and optimization required

Use Cases and Target Applications

EMDI serves specific market participant profiles and strategy types:

Algorithmic Trading Platforms:

Execution algorithms requiring market context
Smart order routing based on consolidated depth
VWAP, TWAP, and implementation shortfall strategies
Strategies operating on second to sub-second timescales

Direct Market Access (DMA) Systems:

Retail and institutional order routing platforms
Real-time order book visualization for traders
Pre-trade analytics and market impact estimation
Integration with existing FIX-based OMS/EMS systems

Firms with Existing FIX Infrastructure:

Organizations already using FIX for other markets
Ability to reuse parsing libraries and frameworks
Reduced development time and testing burden
Standardized operational procedures across venues

Cost-Sensitive Deployments:

Firms not requiring ultra-low latency
Strategies where EMDI latency is sufficient
Lower total cost of ownership than EOBI
Reduced complexity and maintenance burden

Multi-Asset Class Strategies:

Cross-market arbitrage with millisecond horizons
Portfolio rebalancing and risk management
Statistical arbitrage with longer holding periods

Implementation Considerations

Implementing EMDI connectivity requires attention to several technical areas:

Parser Development:

Use commercial FAST libraries (e.g., CoralFAST, QuickFAST) or open-source alternatives
Load Deutsche Boerse-provided FAST templates (1.1 or 1.2)
Validate message parsing against FIXML schema files
Implement robust error handling for malformed messages

Snapshot Synchronization Logic: