it infrastructure database tuning – fall 2015. slotnik’s law of effort #1: heterogeneous systems...

Post on 14-Dec-2015

219 Views

Category:

Documents

0 Downloads

Preview:

Click to see full reader

TRANSCRIPT

IT Infrastructure

Database Tuning – Fall 2015

Slotnik’s Law of Effort #1: Heterogeneous Systems

@ Dennis Shasha and Philippe Bonnet, 2013

LOOK UP: Slotnik vs. Amdahl (AFIPS’67), Michael Flynn’s talk on dataflow machines, Ryan Johnson’s paper on bionic databases.

Source: http://www.anandtech.com/show/2933

RAIDcontroller PCI Southbridge

Chipset[z68]

IO Architecture

@ Dennis Shasha and Philippe Bonnet, 2013

SSD

SSD SSD

HDD

HDD

PCI Express

SSD

SSD

SATA ports

Processor[core i7]

Memory bus RAM

LOOK UP: Smart Response Technology (SSD caching managed by z68)

16 GB/sec

2x21GB/sec

5 GB/sec

3 GB/sec

3 GB/sec

Byte addressable

Block addressable

@ Dennis Shasha and Philippe Bonnet, 2013

RAID Controller

CPU RAM

PCI bridge

Host Bus Adapter

• Caching• Write-back / write-through

• Logical disk organization• JBOD• RAID

Batt

erie

s

ReadWrite

Logi

cal a

ddre

ss s

pace

Scheduling& Mapping

Wear Leveling

Garbage collection

ReadProgram

EraseChipChip

Chip…

ChipChip

Chip…

ChipChip

Chip…

ChipChip

Chip…

Flash memory array

Channels

Phys

ical

add

ress

spa

ce

@ Dennis Shasha and Philippe Bonnet, 2013

Slotnik’s Law of Effort #2: The emergence of SSDs

Throughput for 4K read IOs from product specifications

Latency of 5000 random writes on an Intel710 SSD (10 successive passes over 250 KB with 512Brandom writes on a random formatted device).

LOOK UP: The necessary death of the block device interface

@ Dennis Shasha and Philippe Bonnet, 2013

Case: TPC-C Top Performer (01/13)

Redo Log Configuration

Source: http://www.tpc.org/tpcc/results/tpcc_result_detail.asp?id=110120201

Total system cost 30,528,863 USD

Performance 30,249,688 tpmC

Total #processors 108

Total #cores 1728

Total storage 1,76 PB

Total #users 24,300,000

LOOK UP: TPC-C OLTP Benchmark

Warehouse-Scale Computer

@ Dennis Shasha and Philippe Bonnet, 2013Source: http://www.morganclaypool.com/doi/abs/10.2200/S00193ED1V01Y200905CAC006

LOOK UP: Werner Voegels on virtualization.

Cloud Services

• IaSS: compute and storage abstractions (VM, load balancer)– ex: AWS, Azure

• PaSS: execution environment– ex: Google App Engine, AWS, Azure

• SaSS: managed software– ex: SQL Azure, Snowflake

• Private, public, hybrid coud

source: Microsoft

Key metric: number of servers per administrator – Amazon: 10K

Database Appliances

@ Dennis Shasha and Philippe Bonnet, 2013Source: http://www.oracle.com/us/products/database/exadata/overview/index.html

Mainframe

• Fast processors, lots of hardware accelerators, lots of RAM and storage

• Hardware resources are virtualized: e.g., single-level store

• High-performance, general purpose applicance; or alternative to a private cloud

© Philippe Bonnet, 2014 13

Storage Architectures

source: Virtual Geek’s take on storage tree of life – A MUST READ!!

StorageRAMInterconnect

© Philippe Bonnet, 2014 14

Data-Intensive Applications: Server-side Architectures

source: Virtual Geek’s take on storage tree of life – A MUST READ!!

Look up Fabric Computing on Wikipedia.

15

Source: Chad Sakac,EMC, software-defined storage - what's next, vmworld 2014

top related