Date: 13th November 2018
Location: Fast Data Theatre
Time: 10:30 - 11:00
Speaker: Manish Gupta
Organisation: Redis Labs
About: We live in a world of instant expectations and the technologies underpinning our applications must meet this new demand. This session reviews the mega trends influencing modern application architecture. The talk will discuss use of elements within Redis, the fastest in-memory multi-model database, to support use cases like recommendations engine and personalization that rely on a combination of high-speed analytics and transactions occurring at the same time. Data structures for simultaneous transaction and analytics, probabilistic counting mechanisms, and adaptable machine learning models, will be explored. Real world examples and customer implementations will also be shared. Cost-effectiveness and efficiency of database operations will also be covered in the context of practical enterprise requirements to compete in today's big data world.
9. Cloud Native and Microservices
9
K
Graph Document Columnar RDBMS Cache
App
V
Key/Value
Micro-
service N
Micro-
service 1
Micro-
service 1
Micro-
service 3
Micro-
service 2
Micro-
service 1
App
10. Agnostic to platforms
Ability to talk to old and new systems
No cloud lock-in
Must-have for Enterprises
10
Multi-Cloud and Hybrid
16. Most Popular Database Container
Number of containers (in Billions) launched at Docker Hub (as of OCT 1st, 2018)
16
17. Most Loved Databases 2017 & 2018
Stack Overflow survey, among >100K developers
% of developers who expressed interest in continuing to develop with a database
17
19. An in-memory multi-model database built on top of
open source Redis, delivering a powerful primary
database for enterprises requiring instant experience.
19
20. True Multi-Model Functionality at Any Scale
• Dedicated engine for each
data model (vs. API only)
• Models engines can be
selectively loaded,
according to use case
• All model engines access
the same data, eliminating
the need for transferring
data between them
JSON
GraphSearch
ML
Key-Value
StreamsTime-Series
ENTERPRISE
20
21. Multiple Deployment Options
Cloud
Essentials
Cloud
Pro
Downloadable Native
Kubernetes
Essentials
PCF over Bosh
or PKS
Essentials
Openshift over
Kubernetes
DATABASE-AS-A-SERVICE SOFTWARE
Fully-managed,
server-less and
hosted
Fully-managed,
hosted or in
customer’s
VPC
21
22. .
Average Response Time <1 msec
Sub-millisecond Response Time Under Any Transactions Load
22
23. Highest Performance for Any Use Case
23
ML Serving – x2,000
Search – x5 Graph – x400Serversusedtoachieve1Mwrites/sec
CouchbaseCassandra Redise
$14,832$371,040$2,226,216
25X150X
ANNUAL COST
COST COMPARED
TO REDIS
e
350
300
250
200
150
100
50
0
NoSQL – x150
Streams – x500
24. • Based on years of academic
research on CRDTs (Conflict-Free
Replicated Data Types)
• Global deployment with local
latencies
• Built-in conflict resolution
Active - Active Geo Distribution
24
App
App
App
25. • High Performance
Read and write with low local sub-millisecond latency
• Guaranteed data consistency
CRDT based: The datatypes are conflict-free by design.
All databases eventually converge automatically to the
same state with strong eventual consistency
Supports causal consistency executing read and write
operations in an order that reflects causality
• Simplifies the app design
Develop as if it’s a single app in a single geo, we take
care of all the rest
Tunable Consistency
25
Strong Eventual Consistency and Causal Consistency
Redis App
Redis App
Redis App
26. • Built to run entirely in RAM
• Runs on Flash (NVMe or SATA-based SSD) at sub-millisecond latencies and 80%
lower infrastructure cost
• Runs on persistent memory (NVDIMM, such as Intel 3DXPoint) at near-RAM
performance
Designed for the new generation memory architecture
26
27. Versatile Use Cases
27
ü Authorization
ü Authentication
ü Price Management
ü Advertising Bids
ü Messaging
ü Location-based Processing
ü User Session Management
ü Counting
ü Leaderboards
ü Page Ranking
ü Recommendation Engine
ü Time-series Analysis
ü Session Analysis
ü Secondary Index
ü Accelerated Reporting
ü Real-time Attribution
ü Search
ü Order History
ü Inventory Tracking
TRANSACTIONAL ANALYTICS OPERATIONAL
TRANSLYTICAL / HTAP
27