AccuRev workspace and stream counting

AccuRev organises source control around streams and workspaces, and an audit can mistake those structures for a user count, charging a license for every workspace or stream as though each represented a separate person. A single developer routinely holds several workspaces and works across many streams, so counting structures rather than users inflates the AccuRev finding by a wide margin. Establishing what the metric actually charges for is the first move in any AccuRev defense.

AccuRev, the stream based version control product that came to OpenText through the Micro Focus acquisition, models development as a hierarchy of streams through which changes promote, with each developer working in one or more private workspaces. The architecture is the reason the product is useful, and it is also the reason an audit can overcount: workspaces and streams are plentiful by design, while licensable users are far fewer. When a finding counts the plentiful thing instead of the chargeable thing, the number climbs past anything the entitlement supports.

What AccuRev workspaces and streams actually are

A workspace in AccuRev is a developer's private working area, a place to make and test changes before promoting them up the stream hierarchy. A single developer typically maintains several workspaces, one per task, branch, or environment, so the workspace count is a multiple of the headcount, not a proxy for it. A stream is a configuration line through which changes flow, and an organisation may run dozens or hundreds of streams across its projects, each shared by many developers. Neither structure corresponds one to one with a user, which is why counting them as licenses is a category error.

The licensable unit in most of these products is the user, governed by the Additional License Authorizations, and whether that user is counted on a named or concurrent basis is itself a question that decides the finding, the subject of Dimensions CM concurrent versus named licensing in the related Dimensions family. AccuRev raises the same first question, set out in AccuRev user and stream license questions: what is the licensed metric, and does it count people, streams, or something else. Until that is answered, no count is reliable.

Where the count goes wrong

The decisive question is what the entitlement meters. If AccuRev is licensed by user, the count is the population of developers with access, cleaned of leavers, duplicates, and service accounts, not the number of workspaces those developers happen to keep. If the entitlement is concurrent, the count is the peak of simultaneous sessions, a far smaller figure than either the workspace count or the headcount, and proving it follows the same method as any concurrent challenge, described in how to challenge an ALM concurrent user headcount.

Streams add a particular hazard because automation creates them freely. Build pipelines, integration jobs, and release processes can each spawn streams that no human uses interactively, and counting those as user licenses charges for machine activity. The same reasoning that keeps automated and integration activity out of a user count applies here, the principle examined in how ALM API and integration users are counted. A stream created by a pipeline is infrastructure, not a person.

How the structures map to the entitlement

The reason this matters is that the audit's easiest number to produce is the structure count, because workspaces and streams are listed in the AccuRev server records and can be exported in bulk. The defensible number, the cleaned user count or the peak concurrency, takes work to derive, so a finding left on the structure count rests on the path of least resistance rather than the licensed metric. Shifting the basis from structures to users is the core of the rebuttal, and it depends on reading the server records for what they represent rather than tallying rows.

Reconciling the entitlement against the deployment is the discipline that establishes the correct basis, the work set out in reconciling ALM entitlements before an audit. Once the metric is confirmed, the workspace and stream counts fall into their proper place as evidence of how developers work, not as a tally of licenses owed.

How we defend an AccuRev counting finding under the four Rs

Respond. OpenText gives seven days notice before an audit and the right to copy relevant records, and the seven day notice clock starts immediately. Within that window we take over the single controlled channel and preserve the AccuRev server records, the user inventory, and the session data together, so the relationship between structures and people can be demonstrated rather than assumed.

Reconstruct. We build the effective license position against entitlements and the Additional License Authorizations independently, establishing whether AccuRev is licensed by named user, concurrent user, or another metric, and deriving the count the confirmed metric requires.

Rebut. We challenge every line that counts workspaces or streams as licenses, presenting the user inventory cleaned of non human and departed accounts and, where the model is concurrent, the peak simultaneous session figure. The finding falls by the gap between structures and licensable users.

Resolve. We settle on the count that reflects the licensed metric and, where it serves you, convert forward into an OpenPass agreement that records how AccuRev is measured, so the next review cannot rebuild the finding from a structure count.

An anonymised outcome

The counting basis matters because the remedy is severe. On noncompliance the licensee is deemed to have acquired licenses at then current list price, owes back maintenance and support, owes first year maintenance on the new licenses, and reimburses the cost OpenText incurs performing the audit, so counting structures rather than users multiplies the error across every charge. Our anonymised case files show what correcting the basis achieves: an insurance ECM seat count finding fell from $7.2M to $1.6M, a 78 percent reduction built on disqualifying what the raw count had treated as licensable. An AccuRev finding answers to the same correction, because a workspace is not a person.

Confirm the metric before you accept the count

The durable point is that AccuRev is defended by confirming what the metric charges for, because workspaces and streams are plentiful by design while licensable users are few. A buyer who establishes the metric and produces the cleaned user count or the peak concurrency holds the finding to the basis that was purchased. To build the position, read AccuRev user and stream license questions, Dimensions CM concurrent versus named licensing, how to challenge an ALM concurrent user headcount, and how ALM API and integration users are counted. For the full method see our ALM and LoadRunner audit defense track and our complete OpenText audit defense playbook for 2026. If an AccuRev finding has counted your structures as licenses, open a case.