Pico Market Data, RedlineFeed and Corvil Analytics

Pico supplies market data and trading infrastructure to financial institutions. Its market data service offers raw venue feeds through PicoNet, normalized delivery through RedlineFeed and historical packet captures for replay. Corvil Analytics addresses a separate operational question: whether the feed that reached a trading application was complete and timely. A broker or trading firm can evaluate these services together, but the best starting point is the data each application consumes and how staff will investigate a gap or stale quote during a live session.

Raw feeds through PicoNet

Pico describes dedicated and shared port options for raw venue data delivered over PicoNet. Raw delivery makes sense when the buyer needs venue-native messages or already operates feed handlers built around a specific protocol. The scope should identify which feed and channel each application needs, where delivery occurs and who handles permissioning and exchange onboarding. A commercial count of available products may include different depths, channels or asset classes from the same venue. Compare the actual feeds in the proposal rather than headline coverage.

For an order book application, test the initial snapshot, incremental updates and recovery from a missing sequence. Check the effect of duplicate packets and out-of-order delivery, including whether a secondary channel can fill gaps before recovery is requested. Capture peak traffic as well as average bandwidth. If a venue sends a burst during a reopening, a receiver can run out of queue space even when the circuit appears underused on a minute-by-minute chart. Use application counters and packet captures together to distinguish the two.

Normalized delivery through RedlineFeed

Pico presents RedlineFeed as a consolidated, normalized service with a common API. This can help a team that needs several markets for monitoring or analytics and wants to limit the number of venue-specific integrations. A common schema still needs to represent venue behavior faithfully. Compare mappings for instrument identifiers, precision, depth levels, trade corrections and auction states. Determine which source events are mapped to common fields, which are supplied through extensions and which are unavailable in the feed under consideration.

A useful comparison runs raw and normalized feeds side by side for a limited set of instruments. Record the time at which the application can use each update, then compare book state and trade events across ordinary and unusual sessions. An application may only need the latest bid and offer. A strategy sensitive to queue changes may need more granular updates. Evaluate each consumer separately so a successful dashboard pilot does not stand in for an execution-system test.

Historical packets and repeatable replay

Pico describes historical tick data captured as PCAP files, with replay at original speed. This can help engineers reproduce feed bursts or compare a new parser against a known session. Ask which venues, channels and dates are covered and how the files document capture gaps, timestamp sources and protocol versions. Historical availability can differ from current live availability. Confirm whether the proposed usage allows storage, internal sharing and the intended replay or research workflow.

Start replay with a session that has a known outcome, such as a book state that can be reconstructed independently. Compare a baseline handler with a revised version and inspect the first event where their outputs diverge. For an incident replay, keep original packet order and timestamps available even if the first pass runs at a different rate. A captured feed can reproduce incoming market events, while application logs and order acknowledgments are needed to reconstruct the desk's decisions and venue responses.

Corvil feed monitoring on the Pico product page

How Corvil helps isolate stale data

Pico's Corvil documentation describes passive packet capture and protocol analysis for detecting gaps and tracking latency between observation points. For example, capture at feed ingress and again near the application. If sequence numbers are intact at ingress but incomplete downstream, investigate the internal distribution path. If both observation points see the same missing sequence, compare the redundant channel or recovery stream before assigning responsibility. The evidence is strongest when the monitored protocol can be decoded and the clocks used by the capture points are aligned.

Where an exchange includes a usable timestamp, the platform can compare it with a local capture timestamp. The difference depends on what the source timestamp represents and how both clocks are synchronized. A percentile chart should specify these details along with the capture points and observation window. For tick-to-order measurements, confirm how an order is linked to the triggering market event and how the analysis treats strategies that combine several feeds or send orders after a timer fires.

What to ask for in a procurement pilot

Scope a pilot around one live venue, one application consumer and a historical session that can be replayed. Define the expected fields and recovery behavior, then inject or reproduce a known gap. Operations staff should be able to identify the affected instrument or channel, find the earliest observation of the problem and tell when the application recovered. Where market data is redistributed internally, confirm permissioning and the handoff between the market data team, vendor and application owner.

Pico fits a shortlist when an institution wants feed delivery, repeatable replay or independent visibility into market data quality. A desk replacing venue adapters can prioritize RedlineFeed and compare its mappings with native feeds. A team that already receives all required feeds can evaluate Corvil at the points where evidence is missing. Compare the proposed services using the same venues, observation points and recovery scenarios, then document the responsibilities and data rights that apply in production.