27
ScyllaDB, Present and Future Avi Kivity (@AviKivity) CTO @ScyllaDB

Scylla Summit 2016: ScyllaDB, Present and Future

Embed Size (px)

Citation preview

Page 1: Scylla Summit 2016: ScyllaDB, Present and Future

ScyllaDB, Present and FutureAvi Kivity (@AviKivity)

CTO @ScyllaDB

Page 2: Scylla Summit 2016: ScyllaDB, Present and Future

Agenda

• The Three ScyllaDB Commitments

• ScyllaDB Road Map

Page 3: Scylla Summit 2016: ScyllaDB, Present and Future

Agenda

▸The Three ScyllaDB Commitments

• ScyllaDB Road Map

Page 4: Scylla Summit 2016: ScyllaDB, Present and Future

The Three ScyllaDB Commitments

• High Throughput and Low Latency

•Compatibility with the Apache Cassandra Ecosystem

• Workload Conditioning

Page 5: Scylla Summit 2016: ScyllaDB, Present and Future

Agenda

• The Three ScyllaDB Commitments▸Throughput and latency

• ScyllaDB Road Map

Page 6: Scylla Summit 2016: ScyllaDB, Present and Future
Page 7: Scylla Summit 2016: ScyllaDB, Present and Future

Throughput and Latency

Page 8: Scylla Summit 2016: ScyllaDB, Present and Future

Throughput and Latency

• Reduces Capital and/or Cloud Expenses

• Reduces the need to manage large clusters

• Reduces support costs

• Reduces failure rate

Page 9: Scylla Summit 2016: ScyllaDB, Present and Future

Not Just Equipment Cost!

• Fewer nodes = fewer emergencies

• Reduce risk of double failure

Page 10: Scylla Summit 2016: ScyllaDB, Present and Future

Not Just Equipment Cost!

• Lose fewer customers due to page-load time

• Win more real-time bids

Page 11: Scylla Summit 2016: ScyllaDB, Present and Future

Agenda

• The Three ScyllaDB Commitments▸Cassandra Compatibility

• ScyllaDB Road Map

Page 12: Scylla Summit 2016: ScyllaDB, Present and Future

Cassandra Ecosystem Compatibility

Page 13: Scylla Summit 2016: ScyllaDB, Present and Future

Cassandra Ecosystem Compatibility

• Reuse existing investments and knowledge

• Leverage existing software

• Reduce dev effort, time to market

Page 14: Scylla Summit 2016: ScyllaDB, Present and Future

Agenda

• The Three ScyllaDB Commitments▸Workload Conditioning

• ScyllaDB Road Map

Page 15: Scylla Summit 2016: ScyllaDB, Present and Future

Workload Conditioning

• Internal feedback loops to balance competing loads

Memtable

Seastar SchedulerCompaction

Query

Repair

Commitlog

SSD

Compaction Backlog Monitor

Memory Monitor

Adjust priorityAdjust priority

WAN

CPU

Page 16: Scylla Summit 2016: ScyllaDB, Present and Future

Workload Conditioning Examples

• Prevent compaction from falling behind

• Ensure repair makes forward progress

• Prevent memtable memory from filling up

• Isolate read loads from write loads

• Ramp up load to a newly started node until its cache

is warm

Page 17: Scylla Summit 2016: ScyllaDB, Present and Future

ScyllaDB Commitments RecapSELECT * FROM ScyllaDB.Commitments;

Commitment | Value------------------------+-------------------------------------

Performance & Latency | Reduced CapEx/CloudExEcosystem Compatibility | Reduced time-to-market and dev cost

Workload Conditioning | Simplified operations

Page 18: Scylla Summit 2016: ScyllaDB, Present and Future

RoadMap

Page 19: Scylla Summit 2016: ScyllaDB, Present and Future

Near Term (4Q16)

• Materialized Views / Secondary Indexes

• Counters

• Lightweight Transactions

• Management stack phase 1

• Formal support for REST API

• Container Orchestration Integration▪ For less critical throughput/latency

Page 20: Scylla Summit 2016: ScyllaDB, Present and Future

Why REST API?

• JMX slow, somewhat clumsy▪ Hard to operate from non-Java applications

• Need standard, documented, simple approach to

automating ScyllaDB cluster operations

Page 21: Scylla Summit 2016: ScyllaDB, Present and Future

Medium-long Term (1 / 2)

• New storage format

• Multitenancy

• Analytics

• Search

• Additional protocol support

Page 22: Scylla Summit 2016: ScyllaDB, Present and Future

Medium-Long Term (2 / 2)

• Filesystem bypass

• NVDIMM / 3DXpoint

Page 23: Scylla Summit 2016: ScyllaDB, Present and Future

New Filesystem Format

• C* 2.x format metadata intensive

• C* 3.x format improves, but large partition support

remains slower

• Scylla will provide first-class large partition performance

Page 24: Scylla Summit 2016: ScyllaDB, Present and Future

Multitenancy

• Many orgs run multiple small-ish clusters▪ Wish to isolate performance considerations

• Problems▪ Underutilized hardware▪ Duplication of ops effort

Page 25: Scylla Summit 2016: ScyllaDB, Present and Future

Multitenancy

• Run several “virtual ScyllaDB clusters” on top of one

physical ScyllaDB cluster

▪ Share resources, ops efforts

▪ Workload Conditioning isolates distinct workloads

▪ Each virtual cluster receives an SLA

Page 26: Scylla Summit 2016: ScyllaDB, Present and Future

Roadmap Recap

• Bridging the gap with Cassandra

• ScyllaDB features

• ScyllaDB meta-features

Page 27: Scylla Summit 2016: ScyllaDB, Present and Future

Thank You!

Contact: [email protected], @AviKivity