DuraSpace customer service — expert guide

Historical context and where support lives today

DuraSpace was formed in 2009 to steward community infrastructure projects such as DSpace and Fedora; the DSpace project itself dates to 2002. In October 2019 DuraSpace formally merged into LYRASIS, and commercial, hosted, and institutional support activities now operate under the LYRASIS umbrella. For current program information and membership/contract channels consult https://www.lyrasis.org and the legacy DuraSpace pages at https://duraspace.org (these redirect or provide archival details).

Understanding that lineage is important for customer service because many contracts, institutional relationships, and historical SLAs were written under the DuraSpace name but are now administered through LYRASIS or partner companies. When you engage for support, clarify whether the offering is “community” (forum/Slack + documentation), “paid support” (ticketing + SLA), or “managed hosting” (24×7 ops, backups, and monitoring) so contract terms reference the correct organization and effective dates (post-2019).

Service models and support tiers

Professional support that grew out of the DuraSpace community typically falls into three tiers: community support (no-charge, peer-driven), commercial support (time-and-materials or retainer-based ticket resolution), and managed hosting/operations (fully operated instances, monitoring, and backups). Typical response and coverage models used by former DuraSpace/Lyrasis engagements are: 9×5 business-hour ticketing for standard support and 24×7 on-call or incident coverage for managed services.

Pricing and contract duration vary by tier and scale. As practical guidance: small, institution-only commercial support retainers often start at US$5,000–$10,000 per year; multi-institution consortial agreements or full managed-hosting contracts commonly range from US$20,000 to US$150,000+ per year depending on number of nodes, storage (TB), and agreed SLAs. Always confirm whether storage, bandwidth, and third-party cloud provider fees are included or invoiced separately.

SLA components and recommended targets

An effective SLA contains at minimum: severity definitions, initial response time, target resolution or workaround time, availability/uptime commitment, backup and restore RTO/RPO, scheduled maintenance windows, and performance credits or remedies. For platform software derived from DuraSpace projects, standard SLA targets used in practice are: initial response — 1 hour for Severity 1 (production-down), 4 hours for Severity 2 (major degradation), 1 business day for Severity 3, and 3–5 business days for low-priority issues.

Availability and recovery targets you should negotiate explicitly: 99.9% uptime (≈8.76 hours annual downtime) is common baseline for hosted repositories; higher 99.95% or 99.99% SLAs are possible for additional cost. Recovery objectives typically quoted are RTO (recover service) = 4 hours for critical incidents and RPO (data loss tolerance) = 1 hour for high-value content; lower-cost tiers may accept daily backups with RTO 24+ hours.

Incident handling, escalation, and metrics

Incident management paperwork should define triage workflow: ticket creation (include incident metadata), initial triage within SLA window, assignment to on-call engineer, progress updates cadence (every 30–60 minutes for Severity 1), and formal escalation points (SRE lead, product manager, executive sponsor). Use incident post-mortems with timelines and corrective action within 72 hours for major outages; require Root Cause Analysis (RCA) within 7 business days.

Track and publish KPIs to measure service quality: average initial response time (<4 hours), mean time to resolution (MTTR) goals (Severity 1 <8–24 hours, Severity 2 <72 hours), first-touch resolution rate (>60–75%), backlog count (<10 active >30-day tickets), and customer satisfaction target (CSAT >4.0 out of 5). These measurable targets make vendor performance contractible and audit-ready.

Onboarding, migrations, and practical checklists

Onboarding for a repository platform usually spans 4–12 weeks depending on complexity. Typical phases: discovery and requirements (1–2 weeks), environment provisioning (1–2 weeks), data migration and validation (2–6 weeks), cutover and acceptance testing (1–2 weeks), and knowledge transfer/training (1 week). Expect dedicated engineering effort: 40–160 staff-hours for small migrations and 400+ staff-hours for very large archives with custom metadata or integrations.

For practical success, require a migration plan with clear acceptance criteria (record counts, checksums, metadata mapping correctness, searchability) and an agreed rollback plan. Ask vendors for sample runbooks and proof-of-concept snapshots before committing to final data transfer windows to limit risk of extended downtime or metadata loss.

• Ticket intake checklist (fields to require): affected service/component, full URL or PID, timestamps (UTC), user impact (how many users affected), recent changes/deployments, reproducible steps, logs or screenshots, and preferred contact for updates. Including accurate metadata reduces time-to-resolution by 20–50% in typical operations.
• Onboarding checklist (must-have deliverables): signed SOW with SLA table, designated technical and executive contacts, deployment architecture diagram (including IP ranges and firewall rules), backup and restore validation report, monitoring/alert subscriptions, and one-run failover test schedule within first 90 days.

Documentation, training, and community versus paid support

Documentation and training are key differentiators. Commercial arrangements usually include a knowledge-transfer package: admin guides, runbooks, and at least 8–24 hours of instructor-led training for core staff. Self-service community resources (mailing lists, developer docs, GitHub issues) are useful for lower-severity problems but cannot substitute contractual SLAs for production-critical systems.

When comparing offers, verify what constitutes “knowledge transfer” and whether updates to documentation are included during the contract term. For long-term resilience, budget 6–12 months of institutional cross-training so that no single person is required to sustain the repository operations.

Where to find current contacts, contracts, and further resources

Because DuraSpace’s programs now operate within LYRASIS, start with https://www.lyrasis.org for current support plans, membership details, and contract intake. For community-specific resources on DSpace or Fedora, use the projects’ official sites and communities: https://duraspace.org (archival redirects), https://dspace.org, and https://project-fedorahosted.org (or the Fedora Project site). Always confirm contract signatory and billing entity (LYRASIS or third-party partner) up front so you know where legal and invoice requests should be directed.

Finally, insist on a trial or pilot, a clear SLA table, defined escalation contacts, and periodic service reviews (quarterly business reviews) in the contract. These practical contract terms — combined with measurable KPIs and a documented onboarding/migration plan — are what make DuraSpace-era customer service effective and auditable for institutional repository programs.