Download as PDF, PPTX
![Blasting out I/O operations
future<>
make_data_requests(digest_resolver_ptr resolver, targets_iterator begin,
targets_iterator end) {
return parallel_for_each(begin, end, [this, resolver =
std::move(resolver)] (gms::inet_address ep) {
return make_data_request(ep).then_wrapped([resolver, ep]
(future<foreign_ptr<lw_shared_ptr<query::result>>> f) {
try {
resolver->add_data(ep, f.get0());
} catch (...) {
resolver->error(ep, std::current_exception());
}
});
});
}](https://image.slidesharecdn.com/scyllacassandrasummit2015noanimation-150928185155-lva1-app6891/75/Scylla-1-Million-CQL-operations-per-second-per-server-15-2048.jpg)
Uploaded byAvi Kivity
Scylla: 1 Million CQL operations per second per server
ScyllaDB is a high-performance NoSQL database designed to handle up to 1 million operations per second per node with low latency, fully compatible with Apache Cassandra. It utilizes a unique architecture that eliminates locks and threads, improving efficiency by exploiting hardware resources and enabling linear scaling. The technology supports existing Cassandra drivers and queries, promoting data modeling and eliminating the need for caching layers.
More Related Content
Similar to Scylla: 1 Million CQL operations per second per server
Scylla: 1 Million CQL operations per second per server
- 1.
- 2. 1,000,000 CQL operations perserver Avi Kivity, CTO @ ScyllaDB Cassandra Summit 2015
- 3. Capable of 1,000,000operations per second PER NODE With predictable, low latencies Compatible with Apache Cassandra Scylla: A new NoSQL Database
- 4. THE POWER OF CASSANDRAAT THE SPEED OF REDIS SCYLLA DB
- 5.
- 6.
- 7. FULLY COMPATIBLE ❏ UsesCassandra SSTables ❏ Use your existing drivers ❏ Use your existing CQL queries ❏ Use your existing cassandra.yaml ❏ Manage with nodetool or other JMX console ❏ Use your existing code with no change ❏ Copy over a complete Cassandra database ❏ Works with the Cassandra ecosystem (Spark etc.)
- 8.
- 9. WHAT WOULD YOUDO WITH 1 MILLION TPS? Shrink your cluster by a factor of 10X Handle 10X traffic spikes on Black Friday Model your data instead of using Cassandra as K/V Get the most out of your data - Run more queries Maintain your clusters while serving Stop using caches in front of the database
- 10.
- 11. SCYLLA IS QUITEDIFFERENT Shard-per-core, no locks, no threads, zero-copy Based on the Seastar C++ application framework Efficient, unified DB cache (not using Linux page cache) CQL-oriented storage engine Exploit all hardware resources - NUMA, multiqueue NICs, etc
- 12. SCYLLA DB: ARCHITECTURECOMPARISON Kernel Cassandra TCP/IPScheduler queuequeuequeuequeuequeue threads NIC Queues Kernel Traditional stack Scylla sharded stack Memory Lock contention Cache contention NUMA unfriendly Application TCP/IP Task Scheduler queuequeuequeuequeuequeuesmp queue NIC Queue DPDK Kernel (isn’t involved) Userspace Application TCP/IP Task Scheduler queuequeuequeuequeuequeuesmp queue NIC Queue DPDK Kernel (isn’t involved) Userspace Application TCP/IP Task Scheduler queuequeuequeuequeuequeuesmp queue NIC Queue DPDK Kernel (isn’t involved) Userspace Core Database TCP/IP Task Scheduler queuequeuequeuequeuequeuesmp queue NIC Queue DPDK Kernel (isn’t involved) Userspace No contention Linear scaling NUMA friendly
- 13. Scylla has itsown task scheduler Traditional stack Scylla stack Promise Task Promise Task Promise Task Promise Task CPU Promise Task Promise Task Promise Task Promise Task CPU Promise Task Promise Task Promise Task Promise Task CPU Promise Task Promise Task Promise Task Promise Task CPU Promise Task Promise Task Promise Task Promise Task CPU Promise is a pointer to eventually computed value Task is a pointer to a lambda function Scheduler CPU Scheduler CPU Scheduler CPU Scheduler CPU Scheduler CPU Thread Stack Thread Stack Thread Stack Thread Stack Thread Stack Thread Stack Thread Stack Thread Stack Thread is a function pointer Stack is a byte array from 64k to megabytes Context switch cost is high. Large stacks pollutes the caches No sharing, millions of parallel events
- 14.
- 15. Blasting out I/Ooperations future<> make_data_requests(digest_resolver_ptr resolver, targets_iterator begin, targets_iterator end) { return parallel_for_each(begin, end, [this, resolver = std::move(resolver)] (gms::inet_address ep) { return make_data_request(ep).then_wrapped([resolver, ep] (future<foreign_ptr<lw_shared_ptr<query::result>>> f) { try { resolver->add_data(ep, f.get0()); } catch (...) { resolver->error(ep, std::current_exception()); } }); }); }
- 16. Unified cache Cassandra Scylla Keycache Row cache On-heap / Off-heap Linux page cache SSTables Unified cache SSTables
- 17. Unified cache Cassandra Scylla Keycache Row cache On-heap / Off-heap Linux page cache SSTables Unified cache SSTables Tuning Caching unparsed data Parasitic rows Page faults
- 18. ❏ Implement missingCQL features ❏ Stabilize clustering ❏ Complete nodetool support ❏ Spark integration, other connectors ❏ Thrift support ❏ Authentication and encryption Work in progress
- 19. ❏ Open source@ github ❏ In Beta ❏ Try it out! RPM, Docker images available ❏ Live Demo @ ScyllaDB booth! ❏ → http://scylladb.com OUTLOOK
- 20. ❏ Multi-tenancy ❏ Newprotocols (redis, …) ❏ Vertical integration (solr, …) FUTURE
- 21. SCYLLA, NoSQL GOESNATIVE Thank you. http://scylladb.com