The two committed RPC parity reports of xOneRS show 0 differing answers against geth-XDC: 820 equal answers on Apothem and 1,038 on net2040. Most rows are "not backed": the harness had no data to answer them from, so they prove nothing either way. This post gives the numbers from conformance/rpc-parity/report/apothem.json and net2040.json, both last changed on 2026-11-14.
For how the harness works and which methods xOne serves, see the xOne RPC surface. For the live 50-call fleet samples, see the fleet snapshot.
What do the reports say?
Each report is made by xone-rpc-parity against recorded answers of a geth-XDC archive at the pinned commit 6fc8e8c2088b7f18264f5663036669d04fa7becf.
| Network | Replayed tip | Rows | Equal | Differ | Not backed | Allowlisted | Methods |
|---|---|---|---|---|---|---|---|
| Apothem | 1,800 | 2,610 | 820 | 0 | 1,774 | 16 | 60 |
| net2040 | 747,006 | 2,619 | 1,038 | 0 | 1,577 | 4 | 74 |
The row counts are the sum of the four statuses in each file. The method counts are the entries of each file's methods list.
What does "not backed" mean?
A row is not backed when the harness has no data to answer the request from, so it does not compare it. The reason is written in each row's note, and it differs by network:
- Apothem: the harness replayed blocks up to 1,800. A request for a later block (for example 1,801 or the V2 switch at 56,828,700) has "no executed block, receipts or state", so it is not backed.
- net2040: every not-backed row has the same note: "net2040 is backed by its recorded raw blocks only: no state, receipts or consensus snapshots". Block and header reads compare; state, receipt and XDPoS snapshot reads do not.
A not-backed row is not a pass and not a fail.
What does "allowlisted" mean?
An allowlisted row asks for a value that belongs to the live oracle node, not to the chain, so it cannot match a replay. The notes in the Apothem report name the reasons: the oracle's identity (admin_nodeInfo), its peer set (admin_peers), its live head (eth_blockNumber), and fee history "relative to the oracle's live head" (eth_feeHistory). On net2040 the 4 allowlisted rows are latest calls of eth_getBlockByNumber and eth_getHeaderByNumber; their block-keyed calls are compared.
Which methods carry the equal answers?
| Network | Methods with at least one equal answer | Methods with only not-backed rows |
|---|---|---|
| Apothem | 41 | 9 |
| net2040 | 23 | 51 |
The largest equal counts per method:
- Apothem:
eth_getMasternodes58,eth_getBalance57,eth_getBlockByHash38,eth_getBlockByNumber38,eth_getTransactionCount27, and 19 each for mostXDPoS_*reads. - net2040:
eth_createAccessList104,eth_getBlockByHash100,eth_getBlockByNumber100, then 50 each fordebug_getRawBlock,debug_getRawHeaderand the transaction-count methods.
On Apothem the trace methods (debug_traceTransaction, debug_traceBlockByNumber, debug_traceBlockByHash) and eth_call are all not backed in this report. Tracer parity was measured separately on archive calls in the tracer PRs; see tracing a transaction.
What do the reports not show?
- They do not cover a node at the chain tip: the replayed tips are 1,800 (Apothem) and 747,006 (net2040).
- They do not cover state reads on net2040.
- They are a snapshot of 2026-11-14. Methods added later are not in them.
How do other clients describe RPC checks?
The collected competitor articles describe RPC limits and new methods, not a parity report against a reference client, so no comparison table is given. The RPC-limit claims are compared in the xOne RPC surface.
Frequently asked questions
Did any xOne RPC answer differ from geth-XDC?
Not in the two committed reports: 0 differ rows on Apothem and 0 on net2040.
Why are most rows not backed?
On Apothem the harness replayed only up to block 1,800, so later blocks have no data. On net2040 the harness has the raw blocks but no state, receipts or consensus snapshots.
Is a not-backed row a failure?
No. It is not compared at all, so it is neither a pass nor a fail.
Where can I see a single row?
Each report has a rows list with the method, the block, the status, the differences and a note. The files are in conformance/rpc-parity/report/ in xOneRS.
Sources
This article describes the source and evidence at its review date. Check the linked repository for current implementation and qualification status.
Explore the engineering evidenceEdit this page on GitHub opens the Markdown source (content/blog/rpc-parity-in-numbers/index.md). GitHub may ask you to sign in. Report a correction opens a prefilled GitHub issue for this page.