Timestamping at Scale: Usable bandwidth capacity without the backlog

Timestamping at Scale: Usable bandwidth capacity without the backlog
When your infrastructure is ingesting market data or executing orders across multiple venues, every packet needs to be precisely timestamped. In addition to ensuring compliance, this process allows firms to reconstruct event sequences, provide evidence of the data transmission, and build algorithms that react in real time. Regulatory rules demand timestamping, but global trading strategies in HFTs are an even higher incentive for firms to accurately include timestamps with transactions. NeoTap X, built on our powerful Neo X platform, is designed for institutions that require fast, accurate, and reliable timestamping and aggregation of high-bandwidth data sources.
For trading strategies to be successful, timestamps have to be reliable and trusted. Your firm may be trying to tie together activities between NASDAQ, CME, and Eurex, for example, and if your time source isn’t consistent across devices, racks, and data centers, you run the risk of ending up with gaps, drift, or misalignment across exchanges, negatively impacting your operation. The current industry standard for performance timestamping systems is for 100 ps (picoseconds) of accuracy. This level of precision is essential as it provides a sufficient resolution to detect jitter, measure hardware latency, and even account for physical differences, such as minute latency variations (roughly 50 picoseconds per centimeter of fiber).
Some firms, especially HFT shops and those trading across multiple venues, require “down to the picosecond” level of detail. They're using timestamp data not only to comply with regulations, but to evaluate transceivers, cables, and the timing differences between two events happening on different exchanges. This is what NeoTap X was built for: inline, real-time timestamping at precision levels the industry already expects, and soon beyond as the next version is set to be even more precise.
Timestamping alone isn’t enough if you can’t move the data. A realistic deployment might involve 48 × 10G ports, which is a theoretical 480 Gbps of incoming traffic. Exchanges are also increasingly moving to support 25 GbE due to higher trading volumes, which results in even higher transmission speeds. If a device stamps packets correctly but cannot push out the data, the rest backs up, gets dropped, or forces firms to add more hardware. NeoTap X avoids this entirely. Today, it can output up to 400 Gbps of timestamped data in real time. That means firms don’t need to split handoffs across multiple timestamping devices when feeds spike or when they add more connections. As data volumes increase, the device keeps up instead of becoming the next bottleneck in your trade execution pipeline.
The team at LDA works with clients to create custom solutions, allowing for additional headroom. The hardware itself can process well above 1 Tb per second per FPGA, with a customized configuration, allowing LDA to push the system closer to full capacity. For firms that want to store directly on the box, NeoTap X can integrate that too, removing the need for separate capture servers. In power-constrained or space-limited data centers, that’s often the difference between installing one solution and redesigning an entire rack.
Regarding client benefits, the impact is straightforward: NeoTap X is a single device that doesn’t become obsolete the moment data packet volumes increase. Instead of stacking timestamping appliances or reworking your infrastructure, firms can rely on NeoTap X.
Benefits for NeoTap X include:
- Enhanced Bandwidth and Buffering Memory: with LDA’s Orion HBM card, NeoTap X offers up to 1.088 Tbps of processing bandwidth and 96GB packet buffer memory per FPGA, making it the most powerful and reliable off-the-shelf tap aggregation and timestamping system available.
- Optimize rack space and infrastructure: An optional integrated storage solution offers up to 60 TB of total capacity with up to 400 Gbps of total bandwidth, reducing the dependency on external storage, simplifying deployment, and optimizing rack space
- Low-Latency processing option: Time-sensitive applications benefit from ultra-low latency in-band timestamping (100-200ns packet-in, packet-out), keeping analytics accurate in real-time
- Future-proof & scalable: Dual-FPGA configurations, 100G expansion, and flexible traffic filtering make NeoTap X adaptable to evolving network demands.
What LDA is seeing now is a shift in market data volumes, changes in market timing expectations, and regulatory pressure. As feeds become heavier, multi-exchange trading becomes more commonplace, and regulations grow stricter, firms need to invest in upgrading tap aggregation and timestamping systems to avoid falling behind. Latency-sensitive strategies depend on both precision and throughput. If the device cannot keep up with timestamping at scale while reliably moving the data, the whole setup could come crashing down.
NeoTap X is purpose-built for your growth, so you don’t hit those limits.