javaone-2013: save scarce resources by managing terabytes of objects off-heap or even on disk

60
Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 1

Upload: harvraja

Post on 17-Jun-2015

163 views

Category:

Technology


1 download

DESCRIPTION

Presented at JavaOne 2013. A description of Coherence's Elastic Data feature and the reasons to want such a solution that allows storing data in other resource other than heap-based memory.

TRANSCRIPT

Page 1: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

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

Page 2: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Use your current wisely Harvey Raja Oracle

Chris Neal Pegasus

Page 3: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

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

§  Senior Engineer

§  Oracle Coherence

Introduction

Harvey Raja

Page 4: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

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

§  Systems Architect

§  Pegasus Solutions

Introduction

Chris Neal

Page 5: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 5 5 Copyright © 2011, Oracle and/or its affiliates. All rights reserved.

The following is intended to outline our general product direction. It is intended for information purposes only, and may not be incorporated into any contract. It is not a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decisions. The development, release, and timing of any features or functionality described for Oracle’s products remains at the sole discretion of Oracle.

Page 6: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

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

What’s all this about?

§ Big Data / Big Memory on a transistor diet

§ Applications and conceived concerns

§ Object Profiling

§ Elastic Data

§  Improvements

Page 7: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

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

Heap me up

§  JVM manages our objects

§ Understands Live Data –  References –  Free Lists

new Object()

Page 8: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

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

Heap me up

§ Two distinct regions of data locality –  Young Generation –  Old Generation

§ Allows conscious distinction between: –  Long living objects –  Short lived objects

Page 9: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

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

Heap me up

§ All memory allocations are against the same resource

§ Why would it be any different?

§ Provides means to access –  off-heap memory –  File Descriptors a.k.a. any resource

Page 10: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

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

Heap me up

JVM

100% 0%

0%

Page 11: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 11

Applications & Their Objects

§ Every application has very different uses of objects –  Size –  Scope

§ Structures / Containers –  Structures {Containers} –  Containers {Structures}

Page 12: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 12

Application Object Profiles

§ Ye olde faithful… Pet Store

§ Short-lived objects –  Search Results

§ Long-living objects –  Popular items

JEE Pet Store

JEE Pet Store

JEE Pet Store

JEE Pet Store

Pet Store

Page 13: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 13

Application Object Profiles

§ Foreign Exchange Position Keeping

§ Aggregate Trades Values per currency pair

§ Some currencies are a ‘busier’

§ Currencies may have varying SLAs

FX Position Keeping

T

T

T

T

USD-GBP

USD-EUR

T

T T

T

T

T

T

T

T

T

T

T

Page 14: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 14

What we know about applications?

§ The application understands more about each object than the JVM –  Frequency –  Size

§ Keeping everything in RAM is possible but is it efficient?

§ Huge Leap between Object on heap & stored in DB

-  Recency -  Custom characteristics of the object

Page 15: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 15

Can the JVM just make the right choice?

§ The JVM would have to span multiple devices

§ Non-Heap must be serialized

§ Applications are diverse therefore to make generic decisions on object usage would likely lead to false-positives

§ Down to the application or a layer above the JVM allowing users to define resource assignment policies

Page 16: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 16

Municipality

§ The application deems its own usage of each resource

§  JVM provides primitives to load & store to these devices

§ May be useful to have an API that performs this storage appropriate to the device:

–  Routing stores to appropriate device –  Handle concerns of multiple applications on the same JVM

Page 17: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 17

CPU Architectures

§ As latency increases so does capacity

§ Data fetched as required by instructions

§ Data is demoted as well as writes to shared data rippling through the caches

Page 18: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 18

CPU Architectures Speed Capacity

Registry 1ns 0.00

L1 Cache 2-5ns 2x32KB

L2 Cache 5ns 256KB

L3 Cache 20ns 8MB

Main Memory 20-60ns 16GB

Neh

alem

2G

HZ

pr

oces

sor

Mayfair

Kensington

Camden

Wimbledon

Manchester

§ Perhaps we should charge our objects rental premiums?

Page 19: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 19

Blank Canvas

£600pcm

£1800pcm

£2400pcm

£100000pcm £100

Per TimeUnit £500

Per TimeUnit

£250 Per TimeUnit

£1000 Per TimeUnit

Page 20: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 20

How do you choose the right property?

§ How do you select the right property: –  How often are you in the office? –  How long does it take you to get into the

office? –  How much space do you occupy in the

office? –  How do you get to the office?

–  MFU –  Device latency

–  Object size

–  Every device has its own

quirks

Page 21: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 21

Translated to memory

USD -> GBP

USD-> JMD

Heap RAM (off heap)

Disk

GBP -> MUR

Usage

Pet store example

Pedigree Chum

Orijen

Dog Ugg boots

Page 22: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 22

Translated to memory

§ Objects held on heap are: –  Structures (Containers) –  Containers (Structures)

§ Similar to a file descriptor, each object has its own metadata: –  Access Time –  Modified Time

§ With ((Map) Containers) we already have a location to store metadata

-  Size -  Touch Count

Page 23: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 23

Translated to memory

CPU

Heap

NIO

Flash

Mechanical Disk

§ Each object has metadata

§ Some policy can manage these objects

§ Demoted & promoted to the various media types

§ Big jump from heap to NIO

Page 24: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 24

Brainstorm Summary

§ Cherish your high commodity investments

§ Reduce the regularity of going to a highly contended foreign resource

§ Would be ideal to have objects float between high latency resources

using a telepathic API –  Having some metadata could drive our decision for data locality

§ Map provides a nice abstraction for objects that should ripple

Page 25: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 25

Challenges

§ Serialization cost

§ Generally interactions are performed against the object form

§ False Positives

§ Device type peculiarities

§ A handle to the object (key) and metadata must be held on heap

Page 26: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 26

Elastic Data

§ A Feature in Coherence

§ Store binary key and value objects in RAM or Disk

§ Overflow from RAM to Disk

§ RAM can be configured as NIO

store(byte[] key, byte[] val)load(byte[] key)erase(byte[] key)

RAM

Flash

Page 27: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 27

Elastic Data

§  Its simply writing a number of bytes to some stream?

§ How do you maintain handles?

§ Need a pointer to the written data?

§ How about updates, seek & replace?

Easy Peasy

Page 28: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 28

Elastic Data

§ Require a compact structure to hold handles (keys) to device pointers

§  Journal writes to the file system

§ Consistent API regardless of write to RAM, NIO, Flash or Mechanical Disk

§ Buffer writes with thread dedicated to writing

Implementation Details

Page 29: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 29

Elastic Data

FlashJournal

RAMJournal

CollectorOverflow

Preparer

Writer

Binary Key Pointer

0011010100111001001011

...

...Serialize

110101

LFU

Collector

Store Index

Object deemed unworthy of Heap

store(key, value)Pointer returned

Create

Page 30: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 30

Elastic Data

FlashJournal

RAMJournal

CollectorOverflow

Preparer

Writer

Binary Key Pointer

0011010100111001001011

...

...

Collector

Store Index

Deserialize110101

Deserialized object returned

load(pointer)binary value

The binary key provides the physical location of the stored

item

Read

Page 31: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 31

Elastic Data

Elastic Data

100% 100%

100%

Utilization

Page 32: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 32

Elastic Data

§ RAM Journal can be used with NIO ByteBuffers –  Memory managed using same mechanisms between RAM & Flash

§ Consider device specifics prior to use and design components / interactions accordingly

§ Multi-threaded clients

§ Flash vs Mechanical

More Features

Page 33: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 33

Elastic Data

§ Several platters & reading / writing heads

§ Faraday & Lorentz

§ Seek time + Rotational Latency = L

§ Can not get to the same speeds as the disk controllers

Mechanical Media

Page 34: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 34

Elastic Data

§ NAND gates

§ MLC

§ Write in pages erase in blocks

§ No Seek Time or Rotational Latency

Flash Media

Page 35: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 35

Elastic Data

JVM handle *handle *handle *handle *handle *

01110111001101101010111010110100110101

01110111001101101010111010110100110101

01110111001101101010111010110100110101

01110111001101101010111010110100110101

01110111001101101010111010110100110101

01110111001101101010111010110100110101

01110111001101101010111010110100110101

01110111001101101010111010110100110101

01110111001101101010111010110100110101

01110111001101101010111010110100110101

01110111001101101010111010110100110101

01110111001101101010111010110100110101

01110111001101101010111010110100110101

01110111001101101010111010110100110101

01110111001101101010111010110100110101

handle *(disk pointer)

HARRIET *HARVEY *HILARY *HILTON *

Handles are stored in process

Number of handles constrains amount of

data that can be stored

Key Management

Page 36: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 36

Elastic Data H

AR

RIET VEY

IL

ARY TON

HARRIETHARVEYHILARYHILTON

1 2 4 8Tickets:

§ Data structure to hold handles is a Binary Radix Tree

–  allows sharing of common denominators

§ Handles (keys) are stored in serialized form

§ Benefit increases as common bytes increase and less heap memory is used

Page 37: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 37

Elastic Data

§ Writes are journalled

§ Erase is a logical removal

§ Update = erase + write § Avoids seek time or cascading pointer

changes

write write

write

APPEND

APPEND

APPEND

erase

100111010100111100010111001101111010

100111010100111100010111001101111010

100111010100111100010111001101111010

100111010100111100010111001101111010

100111010100111100

100111010100111100010111001101111010

100111010100111100010111001101111010

100111010100111100010111001101111010

100111010100111100

100111010100111100010111001101111010

100111010100111100010111001101111010

100111010100111100010111001101111010

100111010100111100010111001101111010

100111010100111100

File 1

File 2

File 3

Page 38: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 38

Elastic Data

§ Require a Journal Garbage Collector –  Reclaim unreferenced memory

§ Evacuation process for each file

§ Eviction logically removes from Journal File

§ May enter an exhaustive mode –  Synonymous to Full GC

1001110100111100010111001101111010100111101110011000

1001110100111100010111001101111010100111101110011000

1001110100111100010111001101111010100111101110011000

1001110100111100010111001101111010100111101110011000

1001110100111100010111001101111010100111101110011000

1001110100111100010111001101111010100111101110011000

1001110100111100010111001101111010100111101110011000

JournalCollector

sort

evacuate(exhaustive-

mode)

Page 39: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 39

Elastic Data

§ Dedicated threads to unblock writes –  Tuned to device type

§ Client write appears to be as fast as heap write

§ Overwhelming number of writes will result in push back to the client

Preparer Buffer

Writer

Page 40: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 40

Who am I?

§ Chris Neal, Systems Architect § Started with Pegasus in 1994 § Worked with Coherence since 3.3.1 in 2007 § Participate in Coherence CAB

Page 41: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 41

Who is Pegasus Solutions?

§ Founded in 1988 § Provide technology and services to hotels and travel distributors § Three main service areas:

–  Representation Services –  Distribution Services –  Central Reservation Services (CRS)

Page 42: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 42

Distribution Services

–  Connects hotel systems with distribution partners. –  100,000 hotels connected to all major distributors (Expedia, Orbitz, Hotwire, Travelocity, etc) –  Cheaper and easier than a direct connect –  If you book a hotel online, chances are your transaction goes through Pegasus. –  Pegasus processes roughly 8 billion transactions per month… sustained ~3000TPS @200ms

latency or less.

Page 43: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 43

Why do we need a cache?

§  In travel agent days, “Look to book” ratios were 3:1 §  At internet scale, they are roughly 4200:1

–  Travel aggregators like Sidestep, Kayak, Mobissimo, etc burst this to >100,000:1

§  Looks are the most expensive transaction from systems processing perspective. §  We make no money on “looks”, so saving resources by not processing these

transactions is important to both Pegasus and the downstream hotel systems. §  …Hello Coherence….

Page 44: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 44

Distribution Services and Coherence Physical deployment: –  Production cluster consists of 6 servers, 144GB RAM each. –  Each server runs 3 storage enabled JVMs –  Servers run Solaris x64 –  Each Hotspot JVM is 32GB –  Using CMS collector, and having no GC pauses –  The vast majority of the storage space is for AvailibilityCache (22GB) to service the “looks” Client Applications –  ~120 storage disabled clients either in containers or stand alone –  Backing store is so large that NearCache is disabled

Page 45: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 45

The problem…

§ The cache is too small. Empty to full in ~3 hours. § Evicting valuable, usable data § Cache hit rate is too low

Page 46: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 46

The Challenge….

§  Increase the cache size on the current servers from 20M to 200M § Spend as little money as possible § Do it by EOY

Page 47: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 47

The Process…

§  “spend as little as possible” means adding servers to reach 200M is not possible, which means no more RAM.

§ Enter ElasticData §  In terms of $, RAM > SSD > SATA, but is SATA fast enough?

Page 48: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 48

iozone

§ Reads and writes a file to a filesystem as fast as it can § Compare SSD to SATA with regards to throughput § We know SSD is faster, but will SATA do?

www.iozone.org - Filesystem benchmark tool

Page 49: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 49

SSD benchmark results

Page 50: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 50

SATA benchmark results, part 1

Page 51: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 51

SATA benchmark results, part 2

Page 52: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 52

Elastic Data Hardware configuration

Page 53: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 53

What we observe

§ At production volumes in RAM: –  Avg get/put times ~2ms

§ At production volumes on SATA: –  Avg put times ~3ms –  Avg get times ~10ms

Benchmark data through the application

Page 54: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 54

Hurdles We Have Overcome §  Configuring the heap size:

–  Enough room to store keys for 1.6TB of objects, so that Full GCs do not occur

–  Enough room to store puts() while partitions are being evicted –  64GB (up to 72GB with G1)

§  Eviction killing throughput. Eviction process was reading the values from disk at it evicted (for indexes and listeners). Behavior was changed with BlindCompactSerializationCache.

§  Stopping a JVM: Transferring full partitions. Instead, drop the data, then transfer (DropContentPartitionListener)

§  Starting a JVM: Rebalancing partitions to the new JVM. (DropContentPartitionListener)

Page 55: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 55

Configuring Coherence for Elastic Data § Operational overrides: <journaling-config> <ramjournal-manager> <minimum-load-factor>.4</minimum-load-factor> <maximum-size>8GB</maximum-size> </ramjournal-manager> <flashjournal-manager> <minimum-load-factor>.7</minimum-load-factor> <!-- 3.6TB filesystem size total / 2 JVMs is 1843GB each VM --> <!-- That gives 511 files @ 3690MB each per JVM --> <maximum-file-size>3690MB</maximum-file-size> <collector-timeout>30m</collector-timeout> <!-- 1600GB to force a more aggressive prune (same as high-units) --> <high-journal-size>1600GB</high-journal-size> </flashjournal-manager> </journaling-config>

Page 56: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 56

Configuring Coherence for Elastic Data § Cache-config.xml <distributed-scheme> <backup-count>0</backup-count> <partition-listener> <class-name>com.tangosol.net.partition.DropContentPartitionListener</class-name> </partition-listener> <backing-map-scheme> <ramjournal-scheme> <class-name>com.tangosol.net.cache.BlindCompactSerializationCache</class-name> <high-units>1600KB</high-units> <low-units>1400KB</low-units> <unit-calculator>Binary</unit-calculator> <unit-factor>1048576</unit-factor> </ramjournal-scheme>

</backing-map-scheme> </distributed-scheme>

Page 57: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 57

Garbage Collection settings argv[12]: -XX:+UseG1GC argv[13]: -XX:MaxGCPauseMillis=800 argv[14]: -XX:ConcGCThreads=10 argv[15]: -XX:ParallelGCThreads=10 argv[16]: -XX:InitiatingHeapOccupancyPercent=25 argv[17]: -XX:NewRatio=16

Page 58: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 58

Did we meet our goals

§ Goal: 500M cached objects –  Actual: 1.6B cached objects

§ Goal: Spend as little as possible –  Actual: Spent 1400/machine (8640 total), 84x more objects

§ Goal: Do it by EOY –  Actual: On track for production release before EOY

So far, so good…

Page 59: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 59

Graphic Section Divider

Page 60: JavaOne-2013: Save Scarce Resources by Managing Terabytes of Objects off-heap or Even on Disk

Copyright © 2012, Oracle and/or its affiliates. All rights reserved. 60