10
1 Copyright © 2012 Oracle and/or its affiliates. All rights reserved. Convergence of HPC, Databases, and Analytics Tirthankar Lahiri Senior Director, Oracle TimesTen In- Memory Database

Convergence of HPC, Databases, and Analytics

  • Upload
    yan

  • View
    36

  • Download
    0

Embed Size (px)

DESCRIPTION

Convergence of HPC, Databases, and Analytics. Tirthankar Lahiri Senior Director, Oracle TimesTen In-Memory Database. Enabling The Real-Time World. Financial Services. Social Media. Real-Time Analytics . Telecommunications. eCommerce. Next Generation DBMSs. High-Speed Networks. - PowerPoint PPT Presentation

Citation preview

Page 1: Convergence of HPC, Databases, and Analytics

1 Copyright © 2012 Oracle and/or its affiliates. All rights reserved.

Convergence of HPC, Databases, and AnalyticsTirthankar Lahiri Senior Director, Oracle TimesTen In-Memory Database

Page 2: Convergence of HPC, Databases, and Analytics

2 Copyright © 2012 Oracle and/or its affiliates. All rights reserved.

Real-Time Analytics

Social Media

Telecommunications

Enabling The Real-Time World

eCommerce

Multicore 64-bitProcessors

Massive DRAM capacity

High-SpeedNetworks

Financial Services

Next Generation DBMSs

SSD/Persistent Memory

Page 3: Convergence of HPC, Databases, and Analytics

3 Copyright © 2012 Oracle and/or its affiliates. All rights reserved.

Analytics Requirements

• Intuitive interfaces • Instantaneous response time• Real-time reporting • Structured / Unstructured data• Extreme concurrency

Page 4: Convergence of HPC, Databases, and Analytics

4 Copyright © 2012 Oracle and/or its affiliates. All rights reserved.

Hardware Trends

• Processor throughput – Cores per socket: 5x in last 5 years– Clock frequency: Flat in last 2 years

• Memory capacity 2x every 2 years• High-speed networking: Infiniband, RapidIO, 10GigE, etc. • Persistent memory technologies

– Flash replacing disks for online storage– PC-RAM, MRAM : extension of RAM, or superfast disks for hot data

Page 5: Convergence of HPC, Databases, and Analytics

5 Copyright © 2012 Oracle and/or its affiliates. All rights reserved.

Database Design Trends

• Emerging Industry Trend Exadata, Exalytics, Big-Data Appliance

• Balanced Compute/Capacity/Power• Co-developed components• Built-in scale up, scale out• Built-in interoperability• Pushdown functionality into hardware

Exadata Smart Storage

Integrated Appliances

Page 6: Convergence of HPC, Databases, and Analytics

6 Copyright © 2012 Oracle and/or its affiliates. All rights reserved.

In-Memory Analytics Software

Memory Optimized Essbase

TimesTen for Exalytics

Adaptive In-Memory Tools

1 TB RAM40 Processing Cores

High Speed Networking

In-Memory Analytics Hardware

Optimized Oracle Business Intelligence Foundation Suite

Example: Oracle Exalytics In-Memory Machine

Page 7: Convergence of HPC, Databases, and Analytics

7 Copyright © 2012 Oracle and/or its affiliates. All rights reserved.

Database Design Trends

• CPU core performance is flat• Parallelize (don’t paralyze)

– Coarse-grained parallelism Exploit workload parallelism Parallelize query execution Parallelize maintenance operations (backup/restore) Exploit high speed communication primitives (e.g. Infiniband RDS)

– Fine-grained parallelismVector execution Multi-threading of low-level primitives

Parallelism Everywhere

Page 8: Convergence of HPC, Databases, and Analytics

8 Copyright © 2012 Oracle and/or its affiliates. All rights reserved.

Database Design Trends

• Use all the available tools:– In-memory storage when applicable– Column storage for sequential accesses– Row storage for random accesses

• Advanced compression techniques– More bang for your storage buck

• !! Beware of NUMA !!– NUMA locality awareness – Lock free or well partitioned data structures– Avoid global updates to shared memory

Storage Management

Page 9: Convergence of HPC, Databases, and Analytics

9 Copyright © 2012 Oracle and/or its affiliates. All rights reserved.

Database Design Trends

• Enhance for analytics Analytic functions, data mining models, graph

models, etc

• Cache-friendly access methods Sequential scans better than random access

• Cost modeling for modern hardware – Disk IOs are no longer the dominant cost– Cache Misses / Memory References– CPU cycles, execution time

Query OptimizationWhere’s that %^@#$ plan??

Page 10: Convergence of HPC, Databases, and Analytics

10 Copyright © 2012 Oracle and/or its affiliates. All rights reserved.