Using ScyllaDB for Real-Time Write-Heavy Workloads
•
0 likes•76 views
Learn about major architectural shifts that enable new levels of elasticity and operational simplicity ScyllaDB just launched the first release featuring our new “tablets” architecture. Tablets builds upon a multiyear project to re-architect our legacy ring architecture. Our metadata is now fully consistent, thanks to the assistance of the Raft consensus protocol. Together, these changes enable new levels of elasticity, speed, simplicity, and efficiency. Data is dynamically redistributed as the workload and topology evolve. New nodes can be spun up in parallel and start adapting to the load in near real-time. Join ScyllaDB Co-founder Dor Laor to learn what this new approach means for: - Rapidly responding to traffic spikes without overprovisioning - Performing safe, concurrent, fast bootstrapping - Simplifying cluster administration (e.g., cleanup, repair, tombstone garbage collection) - Increasing efficiency and eliminating overprovisioning
Recommended
More Related Content
Similar to Using ScyllaDB for Real-Time Write-Heavy Workloads
Similar to Using ScyllaDB for Real-Time Write-Heavy Workloads (20)
More from ScyllaDB
More from ScyllaDB (20)
Recently uploaded
Recently uploaded (20)
Using ScyllaDB for Real-Time Write-Heavy Workloads
- 1. ScyllaDB Fast Forward: True Elastic Scale Dor Laor, CEO and Co-Founder, ScyllaDB Felipe Cardeneti Mendes, Solutions Architect++, ScyllaDB
- 2. Poll How often do you scale your database?
- 3. Presenters Felipe Cardeneti Mendes, Solutions Architect + Puppy Lover + Open Source Enthusiast + ScyllaDB passionate! Dor Laor, CEO and Co-Founder + Puppy Lover + Open Source, snowboard and MTB Enthusiast + ScyllaDB passionate!
- 4. Why ScyllaDB? Best High Availability in the industry Best Disaster Recovery in the industry Best scalability in the industry Best Price/Performance in the industry Auto-tune - out of the box performance Compatible with Cassandra & DynamoDB The power of Cassandra at the speed of Redis with the usability of DynamoDB No Lock-in Open Source Software
- 5. +400 Gamechangers Leverage ScyllaDB
- 6. Agenda + Wasn’t ScyllaDB elastic before? + The value proposition + Demo time + Deeper dive + What's next
- 7. < 6.0 ScyllaDB Releases Were Good 2X 2X 2X 2X 2X 3 x 4-vcpu VMs 1B keys 3 x 8-vcpu VMs 2B keys 3 x 16-vcpu VMs 4B keys 3 x 32-vcpu VMs 8B keys 3 x 64-vcpu VMs 16B keys 3 x 128-vcpu VMs 32B keys
- 8. < 6.0 Latency Under Repair/Add/Remove Nodes
- 9. ScyllaDB < 6.0 Problems + Eventual Consistency of cluster metadata + Great architecture for scalable, partition-tolerable, widely distributed DB + But + Topology changes are allowed one-at-a-time + Rely on 30+ second timeouts for poor’s man linearizability + Node failed/down block scale + Streaming time is a function of the schema - parsing partitions + Additional complex operations: Cleanup, rebuild, etc
- 10. Our journey from eventual to immediate consistency & tablets
- 11. ScyllaDB 6.0 Value Elasticity + Faster bootstrap/operations + Incremental bootstrap + Concurrent node operations Simplicity + Decouple topology operations + Transparent cleanups + Semi transparent repairs + Parallel maintenance operations Speed + Streaming sstables - 30x faster + Load balancing of tablets Consistency + Small tables fit in few (single) nodes TCO + Shrink free space + Reduce static, over provisioned deployments
- 12. Demo Time!
- 14. Raft- Consistent Metadata Protocol for state machine replication Total order broadcast of state change commands X = 0 X += 1 CAS(X, 0, 1) X = 0
- 15. node A bootstrap bootstrap Linearizable Token Metadata node B node C system.token_metadata Read barrier Read barrier
- 16. + Fencing - each write is signed with topology version + If there is a version mismatch, the write doesn’t go through Changes in the Data Plane Replica Coordinator Topology coordinator R e q ( V 1 ) Req(V 1 ) F e n c e ( V 1 ) D r a i n ( V 1 ) Req(V 2 )
- 17. Linearizable Schema Version No re-hash of the entire schema on change 10x less CPU with large schemas. TimeUUID-based Schema version Hash-based schema version 5.x: 6.x:
- 18. Tablet History: ClayDB > Bigtable > ScyllaDB
- 19. Consistent Metadata & Tablet Journey RAFT Safe schema changes Safe topology changes Dynamic partitioning Consistent tables Tablets 5.0 5.2 5.2+ 6.0
- 20. + Introduce a new layer of indirection - the tablets table + Each table has its own token range to node mapping + Mapping can change independently of node addition and removal + Different tables can have different tablet counts + Managed by Raft Implementation - Metadata System, tablets Query Replica Set Token
- 21. + Each tablet replica is isolated into its own memtable+ SSTables + Forms its own little Log-Structured Merge Tree + With compaction and stuff + Can be migrated as a unit + Migration: copy the unit + Cleanup: delete the unit + Split/merge as the table grows/shrinks Implementation - Data Path
- 22. Standard Tables {R1, R2, R3} R1 R2 key1 key2 Sharding function generates good load distribution between CPUs R3
- 23. RAFT group No. 299238 RAFT group No. 299236 RAFT group No. 299237 Raft Tables key1 key2 tablet tablet replica tablet replica
- 24. + Source of truth for the cluster + How many tablets for each table + Token boundaries for each tablets + On which nodes and shards do we have replicas + What kind of transition the tablet is undergoing + Which nodes and shards will host tablet replicas after the transition + Managed using the Raft protocol + Replicated on every node The Tablets Table CREATE TABLE system.tablets ( table_id uuid, last_token bigint, keyspace_name text static, replicas list<tuple<uuid, int>>, new_replicas list<tuple<uuid, int>>, session uuid, stage text, transition text, table_name text static, tablet_count int static, PRIMARY KEY (table_id, last_token) );
- 25. Tablets - Balancing Depends on fault-tolerant, reliable, and fast topology changes.
- 26. + Hosted on one node + But can be migrated freely if the node is down + Synchronized via Raft + Collects statistics on tables and tablets + Migrates to balance space + Evacuates nodes to decommission + Migrates to balance CPU load + Rebuilds and repairs Implementation - Load Balancer
- 27. Tablet Scheduler Scheduler globally controls movement, maintenance operation on a per tablet basis repair migration tablet 0 tablet 1 schedule schedule repair Backup
- 28. Tablets Resharding is cheap. SStables split at tablet boundary. Reassign tablets to shards (logical operation).
- 29. Tablets Cleanup after topology change is cheap. Just delete SSTables. Runs automatically, under the hood
- 30. Tablets - Streaming (Enterprise) Send files over RPC. No per schema, per row processing. Up to 30x faster, saturate links. Sstables files Sstables files Sstables files
- 31. + Drivers are tablet-aware + But without reading the tablets table + Driver contacts a random node/shard + On miss, gets updated routing information for that tablet + Ensures fast start-up even with 100,000 entries in the tablets table Driver Support
- 32. Authentication and Service Levels on Raft ScyllaDB 5.x was Manual ScyllaDB 6.x: + Automatically replicated on every node + Linearizable with CREATE/DROP + No denial of service if a node is down
- 33. What’s Coming
- 34. Tablets - Repair - Fire & Forget + Continuous, transparent controlled by the load balancer + Auto GC grace period
- 35. Serverless (VM Based..) Typeless Sizeless Limitless
- 37. Poll How long does it take for you to scale your existing database?
- 38. Keep Learning scylladb.com/category/engineering Register now at p99conf.io Visit our blog for more on ScyllaDB engineering
- 39. Thank you for joining us today. @scylladb scylladb/ slack.scylladb.com @scylladb company/scylladb/ scylladb/
- 40. 40 New metrics in keyspace dashboard ● scylla-monitoring provides an updated dashboard that shows tablets over time per instance and tablet count per instance at the moment
- 42. Tablet Tablet Tablet Tablet Growing a cluster - Tablets Tablet Tablet Tablet Tablet Serving reads/writes
- 43. Tablet Tablet Tablet Tablet Growing a cluster - Tablets Tablet Tablet Tablet Tablet
- 44. Tablet Tablet Tablet Tablet Growing a cluster - Tablets Tablet Tablet Tablet Tablet Start Serving reads/writes