00 · Header

T5 · Dedicated Instance

DI Admin & Operations

Dedicated Instance operations is a RACI, not a shared login. Cisco builds and restores the cloud UC applications. The customer or partner owns the dial plan, the users, premise DNS, and the service restart — after they tell Cisco.

Expand all sections

01 · Executive summary

Five lines, then why it matters

• R means the person doing the work. A signs off. C is two-way. I is one-way. [cite:1]
• Cisco is accountable and responsible for the core application build, core monitoring, and operational readiness testing before handover. [cite:1]
• A widespread issue is 100 percent of a UC cluster, or more than one cluster. Cisco resolves non-configuration issues there and restores clusters. A smaller issue is the customer or partner’s configuration to fix. [cite:1]
• Service requests live at Calling → Dedicated Instance → Service Request. Day-0 and Day-1 types include DNS, firewall, firmware upload, cloud connectivity, audit log, UC application management, DI–MT inter-op trunks, UC application lifecycle, and IaaS. [cite:3]
• Open a TAC case on Collaboration and Conferencing, sub-technology UCM Cloud / Dedicated Instance. Do not choose the Webex option. Port 21 FTP is not allowed. [cite:4]

02 · Architecture map

Who holds the admin plane

flowchart LR
  CP[Customer or partner] --> CH[Control Hub service request]
  CH --> OPS[Dedicated Instance operations]
  CISCO[Cisco TAC and cloud] --> UC[UC applications]

03 · Core concepts

The split the matrix actually prints

Letters

Customer onboarding rows you will be asked

Operations rows that get mis-assigned

04 · Ecosystem & sister technologies

Glossary, map, sisters, edges

Glossary

RACI

Responsible, Accountable, Consulted, Informed, as the Dedicated Instance matrix defines those four letters. [cite:1]

Service request

The Control Hub object at Calling → Dedicated Instance → Service Request for Day-0 and Day-1 assistance. [cite:3]

Widespread issue

100 percent of a UC cluster, or an issue across multiple clusters. Cisco troubleshoots non-configuration issues in that band. [cite:1]

Alerts Center rule

Calling - Maintenance and Outage. The customer or partner configures email, Webex App, or webhooks. [cite:1]

DRS

Disaster Recovery System on Unified CM, IM and Presence, or Unity Connection. Cisco runs it. Partners do not get access. [cite:2]

Custom SAN

Not supported on Dedicated Instance certificates. [cite:1]

n and n-1

The partner operations guide says customers can run the current Collaboration Systems release or the previous one, then states the printed n-1 has entered end of support. Read the live table. [cite:2] re-check

Tier 1

Initial technical support included in the Webex Calling subscription at no additional cost for the subscription term. Basic support is 24/7 in English. [cite:4]

UC application lifecycle request

The service-request type for maintenance schedule questions covering certificate management, software upgrades, or PSIRT fixes. [cite:3]

DI–MT inter-op request

Changes the location in the selected region on multi-tenant for the default trunks Cisco created. [cite:3]

Mind map

mindmap
  root((DI operations))
    RACI
    Service request
    TAC
    DRS
    Certificates

Sister-technology matrix

Integration map

Cross-platform bridges

Future-integration hooks

• Canon plan: T5-R05 owns the DI and multi-tenant dial plan. This room only records that a DI–MT service request changes the multi-tenant location on the default trunks. [cite:3]
• Canon plan: T5-R06 owns the migration gates. Day-2 documents named there point back at this RACI, the partner operations guide, and the service-request list.
• The partner operations version table and the October 09, 2026 RACI sentence about v15SU4a are not the same snapshot. re-check both before a maintenance window. [cite:1] [cite:2]

Ecosystem troubleshooting

05 · Flow

From a fault to the right owner

06 · Config

Paths you can cite

1. Open Calling → Dedicated Instance → Service Request to create and track a request. [cite:3]
2. For a restart, use Service Request → UC Application Management → Server/Service Restart → Notify Cisco regarding the restart. Do this before the restart, including configuration-related restarts. [cite:1]
3. To change the multi-tenant location on the default DI–MT trunks, raise DI–MT Inter-op SIP trunks and choose the region. [cite:3]
4. For certificate, upgrade, or PSIRT schedule questions, raise UC application lifecycle. [cite:3]
5. Subscribe to Calling - Maintenance and Outage and keep the partner contact current. [cite:1] [cite:2]
6. Open TAC at Support Case Manager. Product and Services, not the Webex option. Entitlement is the Dedicated Instance subscription or service contract. Technology is Collaboration and Conferencing. Sub-technology is UCM Cloud / Dedicated Instance (DI). [cite:4]
7. Leave base UC configuration unchanged. Put customer requirements in the non-base dial plan, integrations, and PSTN. [cite:1]

Operations card (nlh9lde / ntqlhe1 / nuk5o63 / nvfve88). No customer secrets.
  route toward DI:     0.0.0.0/0 from the customer side
  more specific:       not accepted
  restart:             notify Cisco first; TAC cannot restart
  disruptive change:   written notice at least 10 calendar days
  AMER window:         9 p.m. ET through 6 a.m. ET, Mon–Fri
  weekend:             Cisco infrastructure maintenance only
  TAC product pick:    not Webex
  TAC sub-technology:  UCM Cloud / Dedicated Instance (DI)
  FTP:                 port 21 not allowed
  custom SAN:          not supported
  DRS:                 Cisco only; not a change backout

07 · Limits

What is quantified

08 · Troubleshooting

Ten ownership failures

09 · Self-question ladder

Twelve rungs

10 · Lab

A paper lab. Do not open a production service request to learn this.

• Goal: assign ten operations events to the RACI column the articles name.
• Prereq: the four articles in the citation list. No customer org you do not administer.
• Step 1. A site advertises 10.0.0.0/8 only. Expected: the matrix says more specific routes are not accepted.
• Step 2. A dial-plan mistake is one subscriber. Expected: customer or partner owns the configuration; Cisco still owns cluster restore.
• Step 3. Someone wants TAC to restart TFTP. Expected: notify first; TAC cannot restart.
• Step 4. Write the AMER window and the 10-calendar-day notice. Expected: 9 p.m. to 6 a.m. ET on weekdays, and the window cannot be changed.
• Evidence: the service-request type you would use for a certificate date, and the TAC sub-technology string.
• Break it: plan a DRS restore as the backout for a route-pattern change. Recovery: remove that backout. The article does not allow it.
• Reflection: which n-1 sentence you still have to re-check before a real upgrade.

11 · Quiz

Twelve questions

12 · Audio

Slots for Studio

13 · Interview

Oral answers

30 seconds. Dedicated Instance splits work with a RACI. Cisco builds and monitors the core applications and restores clusters. The customer or partner owns dial plan, users, premise DNS, and acceptance testing. No partner means the customer takes that column.

2 minutes. Connectivity advertises 0.0.0.0/0. More specific routes are refused. Base configuration stays as Cisco deployed it. Widespread means the whole cluster or several clusters, and Cisco takes non-configuration faults there. A single mis-set route pattern is still yours. You can restart a service only after the Control Hub request notifies Cisco. TAC cannot restart it. DRS is Cisco’s backup. It is not your backout, and you do not get the backups.

30 seconds, certificates. Cisco manages Dedicated Instance UC certificates and will not add a custom SAN. After renewal, SSO has to be disabled, the IdP metadata reimported, and SSO turned back on. Converted UCM Cloud customers keep their own certificates until the domain and CA move to Cisco.

2 minutes, the window. Disruptive planned maintenance gets at least 10 calendar days of written notice if the partner is enrolled for Maintenance and outages. The AMER weekday window is 9 p.m. to 6 a.m. Eastern and cannot be slid. You can ask for another Cisco date with a lifecycle service request, and you should answer in two or three days. A severe vulnerability or a certificate about to expire does not move. Do not vulnerability-scan the cloud yourself.

What you will not say. You will not merge the October 09, 2026 n-1 sentence with the partner-operations v14.0 table. You will not open the TAC case under the Webex product option. You will not claim port 21 FTP is allowed.

14 · Citations

1. Responsibility assignment matrix — help.webex — article nlh9lde — help.webex.com article nlh9lde — header October 09, 2026, retrieved 2026-10-14. RACI letters. Customer assumes the partner column if no partner is active. 0.0.0.0/0 and refusal of more specific routes. Base configuration is not for the customer to modify. Widespread versus non-widespread. Certificate, firmware, DRS, and restart rows.
2. Partner operations guide — help.webex — article ntqlhe1 — help.webex.com article ntqlhe1 — retrieved 2026-10-14. The retrieved markdown did not print a header date. Ten calendar days. Regional windows that cannot be changed. SSO metadata after certificate renewal. DRS access and the one-business-day application-failure goal. n and n-1 table. Vulnerability scanning is not a partner or customer activity. re-check
3. How to raise a service request — help.webex — article nvfve88 — help.webex.com article nvfve88 — retrieved 2026-10-14. Path Calling → Dedicated Instance → Service Request. Day-0 and Day-1 types. DI–MT inter-op changes the multi-tenant location on the default trunks. UC application lifecycle covers certificates, upgrades, and PSIRT.
4. Dedicated Instance Tier 1 support — help.webex — article nuk5o63 — help.webex.com article nuk5o63 — header November 17, 2025, retrieved 2026-10-14. Basic support 24/7 English, 60 minutes for high severity, next business day for lower severity. Do not choose the Webex option. Sub-technology UCM Cloud / Dedicated Instance (DI). Port 21 FTP is not allowed. Peering debug starts with the customer or partner.

15 · Re-engineer later

• Open questions: the partner operations header date, and whether v14.0 is still the printed n-1 beside the October 09, 2026 RACI sentence that names v15SU4a.
• Facts still marked re-check: the partner operations guide rows in the limits table, because that retrieval did not show a header date.
• Assets missing: service-request screenshot, RACI SVG export, audio.
• Ideas for v1.0: a one-page RACI card generated from nlh9lde at review time, with the header date in the caption.
• Next review: re-check maintenance windows and the n / n-1 table before teaching an upgrade.