Enterprise: VDM
Hosted data management platform.
Ediphy VDM is a hosted data management platform: normalisation, enrichment, and analytics delivery across all fixed-income data sources, venue feeds, reference data, tape data, joined into one clean platform. The same layer Ediphy runs internally for its own products.
What it is
One canonical store. Shared across every product you run.
Every institutional desk eventually has the same realisation: the data it relies on arrives across many sources, venue feeds, reference data, tape data, in proprietary formats, with inconsistent timestamps and undocumented quirks, and the in-house team that cleans it is a cost centre that quietly becomes critical.
VDM is the hosted alternative: the same normalisation, deduplication, and distribution layer Ediphy runs internally for its own products, available as a service. It manages every data type, venue, reference, and tape. Raw formats go in; clean, canonical, analytics-ready data comes out, with a single SLA, a single contract, and a single integration point.
REGULATOR VIEW
Users may be paying higher prices for the data they buy than if competition was working more effectively.
– FCA · Wholesale Data Market Study MS23/1.5 · February 2024
Coverage, quality, access.
The four axes that decide whether a data layer is worth building in-house, all handled.
Venues and instruments
All major EU and UK fixed-income venues, APAs, and tape sources. Government bonds, credit (IG and HY), cleared IRS, and inflation across the full institutional curve.
VENUES · INSTRUMENTS
Canonical data model
A canonical instrument model and canonical trade and quote schemas, with consistent deduplication across regimes. One clean shape for every consumer.
CANONICAL · CONSISTENT
Real-time and batch tiers
Real-time stream for trading workflows, near-real-time batch for analytics, T+1 files for reporting and research. Same data, three delivery shapes.
REALTIME · BATCH
Feed shapes
FIX for trading-side consumers, WebSocket for custom UIs, REST for on-demand queries, files for analytics. One data model, every integration pattern.
FIX · WS · REST · FILE
Three dimensions
where hosted wins.
Hosted VDM is structurally more cost-effective than an in-house equivalent on three dimensions.
Fixed-cost amortisation
The cost of running a tier-1 data operation is spread across every Ediphy customer. You pay for access to the result, not for the team and infrastructure that produce it.
AMORTISED · SHARED
Operational accountability
Out-of-hours incidents are managed by Ediphy, not escalated to your team. A named Ediphy operations team holds the SLA.
24/7 · SLA
Quality floor
The same normalisation pipeline runs underneath fairEx and every Ediphy analytics product. The quality standard is set by the most demanding consumer in the Ediphy ecosystem, not by your available budget.
QUALITY · CANONICAL
Point-in-time, by default
Every fact, every version.
Every value in VDM is omnitemporal, recorded with the time it referred to and the time it landed in the store. Any historical question is answered by a query. The same query is available to research, risk, and trading.
A snippet of how that looks in practice, the pattern is consistent across the JupyterLab interface, the REST API, and the file delivery.
One data layer,
six consumer roles.
VDM is the managed data layer every function in a fixed-income business depends on. The shape of the access changes with the role.
Positioning and attribution
Clean price history, consistent curves, and position-level context, the same data the risk team sees.
PRICES · CURVES · POSITIONS
Pre-trade and in-flight
Real-time venue data and normalised composite prints on the same feed that drives Ediphy's own execution products.
REALTIME · COMPOSITE
Full history, point-in-time
Point-in-time queries on every field, a full version history for every value, native JupyterLab access against a single live store.
POINT-IN-TIME · VERSIONED
One source of truth
Risk, PnL, and scenario engines run on the same normalised layer as trading and research, no reconciliation between desk data and risk data.
UNIFIED · RECONCILED
One vendor, one SLA
Consolidates venues, APAs, and tape sources into a single integration, a single contract, and a single accountable operator.
CONSOLIDATION · SLA
Costs, coverage, controls
A fixed-price managed layer with a full audit trail and no concentration risk in a specialised data function.
COST · AUDIT · CONTROL
Architecture
Omnitemporal by default.
Compute goes to the data.
The VDM layer sits on Ediphy Graph, an omnitemporal store that records every version of every fact. Any historical question – "what did we think the curve was at 14:32 last Tuesday? What would it look like with the updated version of our model?" – is answered by a point-in-time query against the same data the live desk uses. Consumers access it through a JupyterLab-native interface where compute runs next to the data, not by shipping TB-scale extracts into a desk-local pipeline. The result: research, risk, and trading operate on the same canonical record, and cross-function reconciliation becomes a non-event.
Cost curves
Hosted or BYOC: three dimensions where Ediphy wins.
The build-versus-buy conversation around venue data has the same pattern at every desk. The matrix below is the short version of the cost case, fixed-cost amortisation, operational load, and quality floor.
VDM is available in both a hosted model and as a BYOC deployment, where the same stack runs in your own cloud with Ediphy providing the operational layer.
| VDM (hosted) | In-house data team | Generic vendor feed | |
|---|---|---|---|
| Same layer Ediphy runs for its own products | – | – | |
| Consistent canonical data model | Tribal | Vendor-defined | |
| Point-in-time queries | Partial | – | |
| Compute next to data | – | ||
| Single SLA, single contract | – | Per source | |
| Quality floor set by Ediphy products | Set by your budget | Vendor-defined |