Product Roadmap Q4 2016

[dianasaur323]

**Note: This is a working copy of our product roadmap featured here: https://github.com/cockroachdb/cockroach/wiki

CockroachDB is currently in beta. We track issues using 3 milestones:

• 1.0 is for work to be completed by the 1.0 release.
• Later is for work to be completed after the 1.0 release.
• Issues without a milestone have either not been scheduled or are cleanup work that can be done whenever someone finds the time or inclination.

2016 Roadmap

Goals:

• Developer adoption
• Ready for production deployment

Themes (near term):

• Core DB
  • Scalability
  • Stability
  • Performance
• Varied cluster deployments & tools
• Fully-featured RDBMS
• Functional UI for cluster management

Q4 2016 Roadmap Details

Core & Stability

• Stability
  • Advanced chaos testing
  • Jepsen
• Test clusters with increasingly large range counts (1M ranges, ~37TB)
• Linear scaling (up to 64 nodes, 3, 10 node clusters with p99 latency with 2.5x p50)
• Cross datacenter deployments (intra-region, inter-region, inter-continental) storage: rule solver post merge issues #10275
• Functionality
  • Infrastructure for data migrations
  • Epoch based range leases Epoch-based range leases implementation #10305
  • Rebalancing of leaseholders storage: add Allocator.TransferLease{Source,Target} #10464
  • Raft pre-vote RPC
  • Parallelized dist sender
  • Quiescence improvements (in presence of down replicas) storage: allow quiescing in the presence of dead nodes #9446
  • Zone config expressiveness config: enhance ZoneConfig expressiveness #4868, storage: add the ability to set/adjust zone configs for system ranges #10692
  • Measure clock skew faster than max clock-offset (high/low max-offset)
  • Queues for high contention
  • Leader lease optimizations

SQL Functionality

• Support top 5 ORMs sql: Hibernate should work with CockroachDB #10079
• Improve performance and isolation for online schema changes sql: schema changes significantly hurt OLTP performance #9149
• Online upgrades WIP: Add cluster upgrade logic for backward-compatible migrations #11658
• Views sql: support non-materialized VIEWs #2971
• Parallelized driver perf: provide batching mechanism for SQL statements #10057
• Collations sql: support collations #2473
• Fuzz testing
• Enterprise backup / restore
• Bulk ingest
• Password-based PGWire cli: missing username in connection URL prevents username from appearing in prompt #10863
• Robustness
  • Minimize resource usage blowups
  • Draining on node shutdown RFC: active draining of nodes #10765
• Distributed SQL
  • Read-only queries (filtering, sorting, aggregations, joins)
  • Push-down filtering to ranges

UI

• UI/UX revamp (explore a new information architecture, reskin UI)
• Improve graphing facility (readability, performance, time scales, UX)
• Snappy performance
• Auto-pruning of time series data

[dianasaur323]

Closing this issue in favor of releasing a new issue with items needed for 1.0.