> For the complete documentation index, see [llms.txt](https://docs.vega365.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.vega365.com/redundancy-management-framework/solution-blueprints/streaming-locality.md).

# Streaming Locality

Stream locally, fail over globally

### About

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FIK2MYX0Q9QXEdTKGtzg8%2Fimage.png?alt=media&amp;token=7bc49bb2-542c-401d-b85f-ef6dcd79aa9a" alt=""><figcaption></figcaption></figure>

In a distributed multi-site XProtect deployment, the recorder closest to an operator is almost always the best one to stream from. It's the lowest-latency path, it keeps viewing traffic off the WAN, and it scales naturally as more sites are added. But "closest" only stays best as long as that recorder is healthy.&#x20;

Streaming Locality lets administrators steer each user group to its nearest recorder by default, simply by authoring views with the appropriate copy of each camera.&#x20;

RMF's stream-level failover and failback continue to run in the background, so when the local recorder is lost, due to hardware failure, network disruption, or planned maintenance, every tile transparently picks up the peer recorder's stream and returns home when the local recorder recovers.&#x20;

Local by default. Global when it has to be.

> **Available in RMF 8.0 and later.** This is supported in Single Site and Federated architectures. For Independent mode, use a PSIM that integrates with the RMF API.

### How It Works

* RMF Smart Sync ensures that both the P and S copies of every licensed camera are present in the XProtect configuration. Streams are mirrored.

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2F62CF0AnBxZdh4Ni6GyE2%2Fimage.png?alt=media&amp;token=aee6b672-b66c-4216-b9a8-5fd8ed139c0b" alt="" width="563"><figcaption></figcaption></figure>

* An administrator creates view groups scoped to one or more XProtect roles and drags the desired copy (P or S) of each camera into the corresponding tile in that view group.

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FVKFxOq8XGjLW7e7fMMaq%2Fimage.png?alt=media&amp;token=b58bf03b-4d6f-4af9-8f7d-d5a17a72a0e0" alt="" width="563"><figcaption></figcaption></figure>

* When a member of that role opens the view in the Smart Client, each tile connects to the data center for the authored copy.

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FhATLDv1A8MVwylijkGeF%2Fimage.png?alt=media&amp;token=a659e40c-abda-4413-99d9-e8b9dd768cef" alt="" width="563"><figcaption></figcaption></figure>

* RMF democratically monitors both streams in the background. If the authored source is lost, the plugin switches to the peer recorder within the configured Reaction Time.

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FJTQZLrWUtKidEtaA4sD4%2Fimage.png?alt=media&amp;token=375e1901-de35-45f7-a061-698d7afd1e5a" alt="" width="563"><figcaption></figcaption></figure>

* When the authored source is recovered and validated, the plugin fails back to it after the Failback interval

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FpufOfJ8bubfk62sEyoHK%2Fimage.png?alt=media&amp;token=282de60f-0a64-43e7-b48f-bc8afc598e37" alt="" width="563"><figcaption></figcaption></figure>

No additional configuration UI is involved. The default-source decision is encoded in the view itself.

### Use Cases

Each scenario uses the same pattern: P or S cams are dragged into different view groups so each user group streams from the recorder closest to it, with RMF stream-level failover and failback running in the background.

#### Metro Rail Transit

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2Fq9YtGJ1VHDMmbmqKKmJx%2Fimage.png?alt=media&amp;token=4e0d4aac-7f1f-49e5-aed4-55f7b58c5155" alt="" width="563"><figcaption></figcaption></figure>

A metro rail operator runs cameras across dozens of stations, with each station served by a local recorder for its own cameras and a Central Data Center holding the redundant copies of every camera system-wide. Each station's operations staff gets a view group authored from the local recorder's copies: they monitor their own platforms, concourses, and fare gates from a recorder sitting in the same station, so traffic stays on the station LAN and response is instant. The Central Monitoring Group at headquarters receives a view group authored from the Central Data Center copies, so dispatchers can pivot across the whole network from a single consolidated source during incidents, drills, or special-event coverage. When a station recorder is taken offline for maintenance or fails, station staff stream-level failover to the Central Data Center transparently and keep operating. When the Central Data Center is unavailable, headquarters fails over to the station recorders. Every group keeps a continuous live view and playback regardless of which side is down.

#### Multi-Campus Hospital Network

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FltWNLaSB5Dw9LVIvJl2D%2Fimage.png?alt=media&amp;token=1da2bd17-ad23-4c33-a9ef-408481c3ba34" alt="" width="563"><figcaption></figcaption></figure>

A hospital system runs three campuses (Main, North, South) plus a regional command center. Cameras at each campus dual-record to a campus recorder (local) and to a central DC (remote). Security staff at each campus get view groups authored from their own campus recorder's copies, so they pull video locally. The regional command center receives a view group authored from the central DC copies, so it can monitor all campuses from one place. During a central DC outage, command center operators stream-level failover transparently to whichever campus recorder still holds the camera; during a campus recorder outage, local staff stream-level failover to the central DC.

#### Distributed Retail Chain with Regional Loss-Prevention Teams

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FHPoLTcov2TBB5z2D4Ahj%2Fimage.png?alt=media&amp;token=5dc25003-0270-46b4-8447-973c8297dec8" alt="" width="563"><figcaption></figcaption></figure>

A retailer with 400 stores groups them into 12 regions. Each region has a regional recorder, and everything also replicates to a national DC. Regional loss-prevention analysts get view groups authored with the regional recorder's copies of the cameras for their stores — fast scrubbing across thousands of cameras with no WAN penalty. The national LP team and corporate investigators can view groups authored from the national DC copies, enabling them to pivot across the entire footprint from a single consolidated source. Both groups see the same evidence; the default recorder differs.

#### Federal Courthouse Complex with Multiple Judicial Districts

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FZ8yOZxDwMIcW8BiAtbYQ%2Fimage.png?alt=media&amp;token=37705625-da5f-4705-a6c7-e9c2db92cf2b" alt="" width="563"><figcaption></figcaption></figure>

A courthouse campus hosts three judicial districts that must keep their video evidence local, plus a shared central DC for disaster recovery. Each district's security staff gets a view group authored from the district recorder's copies. Cross-district auditors and the inspector general get a separate view group authored from the central DC copies, so they can review any district without going through district staff. Forensic imaging of a district recorder takes that recorder offline; district staff stream-level failover to the central DC transparently.

#### Global Manufacturing — Plants and Corporate SOC

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FiAg2TckN58hFLJiDZjCq%2Fimage.png?alt=media&amp;token=a8a48a57-84f9-486b-a892-15cd9b678b79" alt="" width="563"><figcaption></figcaption></figure>

A manufacturer operates plants in Mexico, Vietnam, and Poland with a corporate Security Operations Center in Frankfurt. Each plant gets a view group authored from its local plant recorder's copies of its own cameras. The Frankfurt SOC gets a view group authored from the corporate DC copies for cross-plant visibility. Plant security responds in seconds; the SOC sees everything from one place. When a plant recorder is taken down for maintenance, both groups automatically fail over to the corporate DC at the stream level.

#### Higher-Education Campus — Police, Residence Life, and Athletics

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2F6CB1RigY3216dgoAKWsB%2Fimage.png?alt=media&amp;token=9d01fc7e-6441-439c-a899-78f7569f4d89" alt="" width="563"><figcaption></figcaption></figure>

A university has campus police dispatch, residence-life staff in each dorm tower, and an athletics operations center at the stadium. Residence-life staff get a view group authored from the dorm recorder cluster's copies. Athletics ops gets a view group authored from the stadium recorder's copies. Campus police get a view group authored from the central police DC copies, so dispatch can see anything during a campus-wide incident. Stream-level failover between recorder clusters holds during scheduled maintenance.

#### Utility or Power Generation — Plant Control Rooms and Regional Ops

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2F7BdvMte2Eh1TCWnL6Uec%2Fimage.png?alt=media&amp;token=99e4d6ca-6245-418f-8e62-5b1f0023a6b0" alt="" width="563"><figcaption></figcaption></figure>

A utility runs five generating stations plus a regional operations center. Each plant has a control-room recorder; the regional ops center has its own DC. Plant operators get view groups authored from the plant recorder's copies of substation and yard cameras, keeping the plant LAN self-sufficient if the regional WAN is cut. Regional ops staff get a view group authored from the regional DC copies. During cyber-lockdown drills, plants keep operating in isolation; regional ops stream-level fails over to whichever plant recorders remain reachable.

#### Hotel Chain — Property Managers and Brand-Protection Team

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2F1ilNTCw9xWOdEJvaptQ4%2Fimage.png?alt=media&amp;token=0b79ef00-1335-410d-b416-3d9ea0e2dbb0" alt="" width="563"><figcaption></figcaption></figure>

A hospitality group operates 80 hotels under three brands. Property security at each hotel gets a view group authored from that property's recorder copies. The HQ brand-protection team gets a view group authored from the central DC copies, allowing them to investigate incidents across any property. Insurance investigators are given a third role with their own view group, also authored from the central DC copies but with read-only export rights — stream-level failover semantics identical, access policy different.

#### Casino — Floor Operations vs. Surveillance Department

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FEp9Xc20uiTXydmGuSuwT%2Fimage.png?alt=media&amp;token=fac810d1-27ed-4f44-b277-fc317e77624d" alt="" width="563"><figcaption></figcaption></figure>

Gaming regulators in many jurisdictions require the surveillance department's video to be functionally independent from floor operations. A casino deploys two recorder clusters — one operational and one for surveillance and compliance — each receiving dual streams. The floor-operations role gets a view group authored from the operations recorder copies. The surveillance and compliance role gets a view group authored from the surveillance recorder copies. Regulators see clear default separation; RMF preserves stream-level failover between recorders behind the scenes to satisfy uptime obligations.
