Upload
others
View
30
Download
0
Embed Size (px)
Citation preview
2
议程
• 数据中心的存储困境
• 基于英特尔®架构(IA)的横向扩展存储解决方案概述
• 使用英特尔®产品增强横向扩展存储的性能
• 测试和优化实践
– Swift*
– Ceph*
– Gluster FS*
• 总结
3
存储投入分析
0
20,000
40,000
60,000
80,000
100,000
120,000
140,000
160,000
180,000
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
全球企业不同存储类型消耗出货量, 2006–2015 (PB)
来源: IDC, 2011 Worldwide Enterprise Storage Systems 2011–2015 Forecast Update, Doc#231051
线性增长
指数级增长
内容仓库和公有云/大量的非结构化数据
传统的非结构数据
传统的结构数据
公有云-企业托管服务
容量(Petabytes)
年
移动和云业务推动存储消耗指数级增长
4
典型的新存储用户方案
传统的存储解决方案能满足新出现的需求吗?
更好的解决方案:基于英特尔®架构平台的横向扩展的存储解决方案
• 大量的非结构化消息和图片 微博
• 大量监控视频、图片和日志文件 平安城市
• 病历/高质量的医疗影像 (CT) 医疗
• 虚拟机镜像 企业云
• 容量: GB级别到TB/EB级别 • 价格: $ per MB • 吞吐量: 同时支持几百几千甚至更多的主机
• 响应时间: 扩展存储的同时保持响应时间和吞吐量不变
• 灵活性: 业务弹性要求存储可动态配置并易于管理
• 容错: 没有单点失效
• 大批量集中存储阵列 • 主机通过硬件控制器/电缆连接到存储阵列
• 高性能/高吞吐量 • 磁盘级别容错 • 昂贵的解决方案
新的存储需求
传统的纵向扩展存储
5
什么是横向扩展存储?
Metadata
Server
Data Server
控制流 数据流
定义:
• 大规模低成本的硬件基础设施。英特尔®架构平台是最好的选择。
• 可扩展的系统架构,多个数据服务器共同承担存储负载,元数据服务器定位存储信息
• 高性能/高吞吐量
• 高可靠性/高可用性
• 高可扩展性
类别:
• 分布式文件系统
• 分布式对象存储
• 分布式块设备
特点:
• 冷数据,访问频率及实时性要求不高
• 非结构化数据
可扩展的存储设计通常是与业务紧密结合
IA Platform IA Platform
IA Platform
Data Server Metadata
Server
IA Platform
元数据
服务器
IA Platform
Data Server
IA Platform
IA Platform
数据服务器
客户端
†IA Platform
Client
IA Platform
Client
IA Platform
客户端
IA Platform
†IA Platform = Intel® Architecture Platform
6
横向扩展存储方案
基于文件的商用横向扩展NAS
基于对象的商用横向扩展存储
横向扩展存储分类概述 IBM* SONAS*
EMC* lsilon*
Dell* FluidFS* HP* StoreAll* Storage
DDN* EXAScaler*
Hitachi* NAS (HNAS) Quantum StorNext Huawei* OceanStor* N9000 Red Hat* Storage Server 2.0 Oracle* ZFS
…
EMC* Atmos*
DDN* WOS*
Amplidata* AmpliStor* Object
Storage system
…
GlusterFS*
Lustre*
Ceph*
HDFS*
Swift
Ceph
Sheepdog
…
MogileFS
MooseFS
FastDFS
…
廉价存储解决方案=英特尔®至强™ 处理器平台 + 开源软件栈
基于文件的开源横向扩展存
储
基于对象的开源横向扩展存储
7
开源横向扩展存储
7
项目名称 主要特性 存储类型 成熟度
Swift* • 支持多个代理服务器,没有单点失效 • 支持多租户,基于Python* • PB基本存储 • 兼容AWS S3
对象存储 商用部署较少
Ceph* • 包括多个元数据服务器,没有单点失效 • 兼容POSIX,基于C语言 • 支持块存储,对象存储和文件系统
文件存储/对象存储
新兴的解决方案; Inktank*提供了企业级的Ceph商业支持
GlusterFS* • 没有元服务器也没有单点失效 • 兼容POSIX,基于C语言 • 支持NFS,CIFS,HTTP,FTP,Gluster SDK/ API访问 • 设计支持几百个PB级的数据
文件存储 100多个国家或地区正在使用 GlusterFS
Lustre* • 包括元服务器,有单点失效 • 兼容POSIX,基于C语言 • 支持10K+节点 PB+存储,100GB/ s
文件存储 排名前100位HPC项目中超过40%项目采用Lustre*
9
使用业界领先的英特尔®固态硬盘提升横向扩展存储的性能
Intel® SSD DC S3500/S3700 系列
高效而稳定 的性能
SATA III 6 Gbps 接口 75K/36K IOPS 4K 随机读写 50/65us 平均延迟 <500us 最大延迟 500/460 MBps 稳定顺序读
英特尔® HET 技术保证高持久
超过5年的10 DWPD 负荷 JEDEC 持久性标准
压力无关的 数据保护
端到端数据保护 断电保护 256位AES加密 ECC校验保护 2百万小时的MTBF
大容量
2.5英寸: 100/200/400/800 GB 1.8英寸: 200/400GB
10
• 全新技术
– 可首先使用独立网卡支持万兆技术,当需求量大于50%时可选择板载网卡
• 使用10GbE以太网卡建设新数据中心
– 降低成本和能耗, 缩减布线复杂度, 利于长远规划
– 完美支持数据中心虚拟化
– 统一数据中心网络结构(LAN, iSCSI, FCoE)
• 英特尔®全新服务器平台提供多种10GbE选择:
– 独立网卡 – 提升服务器整体级别
– 网卡扩展槽– 低成本易于扩充
– 1GB/10G双重接口 – 为日后升级做好准备
– 10G baseT + 10G SFP+ 板载网卡– 最低成本配置
1 GbE 服务器网线连接
10GbE 服务器网线连接
降低
15% 基础设施投
入
减少
80% 电缆与端
口
降低
45% 机架能耗
2x 服务器带宽提升
使用业界领先的英特尔®万兆网卡提升横向扩展存储的性能
12
测试和调优实践的议程
对于每一个的解决方案(Swift*, Ceph*, GlusterFS*),我们将讨论:
• 解决方案的架构
• 测试环境和工作负荷
• 基准测试的性能
• 分步骤的性能调优
• 总结
14
Swift*: 架构综述
• Swift*
– 易扩展的分布式对象数据库 – 被优化支持多租户和高并发 – Swift 是理想的备份,web和移动内容, 以及其他非结构化数据存储方案,可以无限扩展
• 主要模块 – Proxy服务 – Account服务 – Container服务 – Object服务 – Authentication服务
• 主要功能 – 持久性 (区域, 复制) – 无单点服务 (NWR理论) – 扩展性 – 多租户
15
Swift*: 测试环境 • 硬件列表
• 软件栈
用途 数目 CPU 内存 磁盘 网卡
压力客户端 4 X5670 2.93GHz 2*6
24G SATA 1000Mbit/s
Proxy服务器 1 E5-2680 2.70GHz 2*8
64G SATA 1000Mbit/s*2
Storage服务器 4 E5-2680 2.70GHz 2*8
64G SATA 1000Mbit/s
软件 版本
swauth 1.04
Swift 1.7.4
COSBench 2.1.0
collectd 4.10.1
16
Swift*: 压力测试工作负荷
名称 配置 指标 用途
读小文件 文件大小64KB,运行5分钟 IOPS, RESP TIME 网站
读大文件 文件大小1MB,运行5分钟 IOPS, RESP TIME 音乐
写小文件 文件大小64KB,运行5分钟 IOPS, RESP TIME 在线游戏
写大文件 文件大小1MB,运行5分钟 IOPS, RESP TIME 企业备份
IOPS: 每秒IO数目 RESP TIME: 响应时间
COSBench*
• Intel开发的基准测试工具,用来测量云对象存储服务的性能 • 组件:
– 控制节点 – 被控节点 – 命令行工具/门户
性能敏感指标: CPU使用率, 网卡带宽使用
17
Swift*: 基准测试性能指标
名称 IOPS REPS (ms) 成功率
读小文件 1615.25 313.63 99.8%
读大文件 108.16 4772.13 99.8%
写小文件 493.58 1039.64 100%
写大文件 37.96 6852.46 99.94%
Swift 配置: 1. Proxy服务实例: 64 2. Object服务实例: 16 3. Account服务实例:16 4. Container服务实例: 16 5. XFS inode大小: 1024 6. 其他配置采用默认值
Proxy服务器: CPU使用率~50%, 网卡使用~100% Storage服务器: 网卡使用~50%, CPU使用率 ~40%
使用Intel®10G 网卡替换原来的千兆网卡
Proxy服务器网卡带宽耗尽
18
调优– 使用Intel® 82599EB万兆网卡
名称 IOPS REPS (ms) 成功率 相比基底数据
读小文件 4271.4 159.74 99.9% >150%
读大文件 406.42 2478.9 99.49% >150%
写小文件 560.64 916.97 100% ~13.5%
写大文件 94.76 3980.7 100% ~150%
没有达到预期
Proxy服务器: CPU使用率 ~50%, 网卡使用率 ~30% Storage服务器: 网卡使用率~50%, CPU使用率 ~40%
详细分析Proxy服务器
0
10
20
30
40
50
60
70
80
90
100
总计
cpu
0
cpu
1
cpu2
cpu3
cpu4
cpu5
cpu6
cpu7
cpu8
cpu9
cpu1
0
cpu1
1
cpu1
2
cpu1
3
cpu1
4
cpu1
5
cpu1
6
cpu1
7
cpu1
8
cpu1
9
cpu2
0
cpu2
1
cpu2
2
cpu2
3
cpu2
4
cpu2
5
cpu2
6
cpu2
7
cpu2
8
cpu2
9
cpu3
0
cpu3
1
soft%sys%user%
proxy服务器
CPU0耗尽,90%用来处理软中断
19
• 了解网卡特性
– Intel万兆网卡上拥有多个队列
– 每个队列对应一个中断号
dmesg | grep ixgbe
cat /proc/softirqs | grep NET
软中断不均衡
深入分析: 内核工具 stap & addr2line
最佳使用方案: 每个中断绑定到不同的core上
调优– 使用Intel® 82599EB万兆网卡(续)
20
调优– 使用Intel® 82599EB万兆网卡(续) • 中断数目远远小于CPU的核数
– 最佳使用方案: 将中断绑定到同一socket或同一NUMA节点下的核上
• 了解CPU的架构
按下面顺序轮流绑定中断到核上: cpu0-cpu7, cpu16-cpu23 cpu8-cpu15, cpu24-cpu31
21
调优– 使用Intel® 82599EB万兆网卡(续)
• 重要的组件: memcached
– 用途:
缓存客户端的token
缓存Ring用来查找
– 调优方式:
增大初始化内存大小
增加并发客户连接的数目
• Dmesg错误: ip_conntrack: table is full, dropping packet
– 最佳使用方案: 增加Linux NAT hash表的大小
例如: net.ipv4.netfilter.ip_conntrack_max = 655350
• 其他:
– Linux* ulimit调节
22
名称 IOPS REPS (ms) 成功率 跟调优前相比
读小文件 7571.4 189.74 99.9% >90%
读大文件 736.42 2678.9 99.49% >90%
写小文件 563.34 716.97 100% ~0%
写大文件 121.38 3280.7 100% ~30%
(除了写小文件外)Proxy服务器: CPU 使用率~50%, NIC使用率~40% Storage服务器: 网卡使用率~50%, CPU使用率 ~40%
0
20000
40000
60000
80000
100000
120000
140000
转发 接收
速度KB/S Proxy网卡
0
10
20
30
40
50
60
user% sys% iowait%
CPU % Storage CPU使用
调优– 使用Intel® 82599EB万兆网卡(续)
23
调优—增加磁盘数目
名称
IOPS REPS (ms) 成功率 跟调优前相比
写小文件 723.34 696.17 100% ~28%
增加存储节点磁盘数目: 从2块SATA磁盘增加到4块SATA磁盘
0
50000
100000
150000
200000
250000
转发 接收
速度KB/S Proxy网卡使用
0
10
20
30
40
50
60
70
user% sys% iowait%
CPU % storage CPU使用率
24
调优–使用Intel® SSD 320系列存储账号和容器数据
• Intel® SSD 可以显著提高磁盘的性能,可以用来存热数据或者相关索引数据
• Account和Container数据可以被存储在SSD中,可以提高性能
名称 IOPS REPS (ms) 成功率
专门的测试流程 245.19 303.19 100%
名称 IOPS REPS (ms) 成功率 跟替换前比较
专门的测试流程 298.13 292.23 100% >20%
专门的测试流程: 容器中已经包含大量的数据,这种情况下再写入数据…
25
Swift* 调优总结
• 配置示例 – 硬件 Proxy服务器使用万兆网卡或者负载均衡器使用万兆网卡
Storage服务器挂载更多的磁盘,增加并行写
SSD存储Account和Container数据
– 软件 绑定中断号到不同的CPU核上
增加memcached的内存大小和并发连接数
增加内核NAT hash表的大小
– Swift*配置 Proxy服务实例: 64(两倍于CPU核数)
Object服务实例: 16(CPU核数一半)
Account服务实例:16(CPU核数一半)
Container服务实例: 16(CPU核数一半)
XFS inode大小: 1024
其他配置采用默认值
Memcached做缓存
26
Swift* 调优总结
名称 IOPS REPS (ms) 成功率 相比于基底数据
读小文件 7571.4 189.74 99.9% 350%
读大文件 736.42 2678.9 99.49% 350%
写小文件 723.34 696.17 100% ~50%
写大文件 121.38 3280.7 100% ~220%
大规模部署示例
28
Ceph*: 架构总览
Ceph* 独一无二的提供对象存储、块设备存储与文件系统的统一存储架构. Ceph* 具备高可靠性、基于管理并且免费。
3种接口:
1. CephFS
2. Ceph RADOS Gateway
3. Ceph Block Devices (RBD)
我们会集中讨论 Ceph RBD的性能.
APP APP HOST/VM Client
RADOS
可靠、自管理、分布式的对象存储, 由自恢复、自管理的智能存储节点组成
LIBRADOS
允许应用直接访问RADOS的库, 支持r C, C++, Java*, Python*, Ruby和 PHP
CEPH FS
POSIX兼容的分布式文件系统, 支持内核模块和FUSE接口
RBD
可靠和完全分布式的块设备存储, 具有内核和QEMU/KVM两种接口
RADOSGW
基于数据块的REST网关. 与S3和Swift兼容
29
Ceph*: 架构总览(续)
系统架构: 客户端的IO操作通过直接去OSD通信来实现. 每一个客户端进程或者直接使用系统提供的库, 或者操作挂载的文件系统
• MDS (Metadata Server Cluster) • OSD (Object Storage Cluster) • MON (Cluster Monitors) • Client
30
测试环境
节点 IP 主机名 OS 版本
MON&MDS 192.168.3.22 NEW-MDS Ubuntu* 12.04.2 LTS
OSD0 192.168.3.19 NEW-OSD0 Ubuntu 12.04.2 LTS
OSD1 192.168.3.20 NEW-OSD1 Ubuntu 12.04.2 LTS
OSD2 192.168.3.21 NEW-OSD2 Ubuntu 12.04.2 LTS
Client 192.168.3.7 compute1 Ubuntu 12.04.2 LTS
Client/MON&MDS/OSD0/OSD1/OSD2: CPU: Intel® Xeon® Processor E5-2680 0 @ 2.70GHz MEM: 8x8GB DDR3 1600Mhz HDD: SATA Seagate* 1TB 7200PRM x 3 SSD: Intel® SSD 320 300GB 10GB NIC Chipset: Intel® 82599EB 10 Gigabit Ethernet Controller 1GB NIC Chipset: Intel® Ethernet Controller I350
31
负载与基准测试结果
• 基准测试性能
- 测试工具 : iozone v3.397 - 单客户端读写测试
• 负载
iozone -i 0 -i 1 -r X -s Y -f /mnt/rbd-block/iozone -Rb ./rbd-X-Y.xls –I -+r
X 为块大小, Y 为文件大小 -I 所有操作使用 O_DIRECT -+r 所有操作使用 O_RSYNC|O_SYNC
0
20,000
40,000
60,000
80,000
100,000
120,000
256M 512M 1G 2G
吞吐
量(K
B)
文件大小(字节)
1客户端读写测试
Write 1M
Write 4M
Write 16M
Read 1M
Read 4M
Read 16M
块大小
系统网络IO
32
性能调优实践
步骤1: 利用Intel® SSD 替换部分硬盘
观察:
mkfs.xfs -n size=64k /dev/sde mount /dev/sde /srv/ceph/osd0 ceph.conf: osd journal = /srv/ceph/osd0/journal
调整: 利用Intel® SSD 存储日志文件 Result: 明显的写性能提升
0
20000
40000
60000
80000
100000
1M 4M 16M
Thro
ug
hp
ut(
KB
/S)
HDD(Baseline) SSD
2.69x
2.73x
1.47x
33
性能调优实践
步骤2: 为 OSD集群使用专有网络
原因: Ceph* 可以通过为OSD配置独立的专有网络来分离内部数据传输, 达到为对外数据带宽减负的目的.
ceph.conf: [osd] cluster network = 192.168.3.0/24 public network = 10.0.0.0/24 [osd.0] public addr = 10.0.0.19:6802 cluster addr = 192.168.3.19
调整: 配置 Ceph* 使用独立专有网络
0
20000
40000
60000
80000
100000
1M 4M 16MTh
rou
gh
pu
t(K
B/S
)
SSD SSD-Private
结果: 性能略有提升 1.02x
1.04x
1.06x
34
步骤3: 采用1Gbe 网卡绑定
原因:
调优: 配置客户端采用网卡绑定
我们可以观察到客户端的带宽已经用光了
0
20000
40000
60000
80000
100000
120000
1M 4M 16M
Thro
ug
hp
ut(
KB
/S)
SSD-Private SSD-private-Bonding
结果: 写性能略有提升
1.10x
1.02x
1.02x
性能调优实践
35
0
100000
200000
300000
400000
500000
600000
ReWrite Read
Thro
ug
hp
ut(
KB
/s)
1G Bonding/SSD 10G/SSD
步骤4: 使用10Gbe 网卡替换1Gbe网卡
原因:
调优: 可调整bonding分流算法; 但考虑到满流量也仅200MB/s, 直接使用万兆网卡
模拟块设备await很大, 网卡流量不均衡, Bonding效果不佳
结果: 读性能大幅提升
4.33x
1.02x
性能调优实践
36
1G/HDD 1G/SSD 1G Bonding/SSD 10G/SSDReWrite 28831 91946 113719 119980Read 101846 107920 119314 516217
0
100000
200000
300000
400000
500000
600000吞
吐量
(KB
/s)
)
ReWrite
Read
Ceph* 调优总结
参考值: Local SATA 7200RPM 写性能= 101,403 KB/S
38
GlusterFS*:架构
以一个可堆叠的用户空间设计为基础的横向扩展网络附加存储(NAS)的文件系统
• 服务器
• Brick
• 客户端
• 子卷
• 卷
GlusterFS* 客户端
Volume
bric
k bric
k bric
k bric
k
Volume
bric
k bric
k bric
k
存储网关
Volume
bric
k bric
k bric
k bric
k
RDMA
NFS CIFS(Samba)
Gluster 存储云
服务器端
39
Gluster FS*:测试环境
• 硬件:
– GlusterFS* 客户端: 1-2
– GlusterFS* 服务器: 2
• 软件:
– 操作系统: Ubuntu* 12.04 LTS
– GlusterFS * 版本: 3.2.5
– IOzone 用于大文件测试(读/写)
CPU: Intel® Xeon® Processor E5-2680 2.70GHz MEM: 8x8GB DDR3 1600Mhz HDD: SATA Seagate* 1TB 7200PRM x 3 SSD: Intel® SSD 320 300GB 10GB NIC: Intel® 82599EB 10 Gigabit Ethernet Controller
40
Gluster FS*: 基准测试 • Gluster FS* 卷
– 类型: Distributed
– 卷选项
读文件
io-thread-count: 16
cache-size: 32MB
cache-max-file-size 16384PB
cache-min-file-size 0
写文件
write-behind-window-size: 1MB
write-behind: off
io-thread-count: 16
flush-behind: on
• 负载:
– Record size: 4K~16M; 较大的record size更有利于写操作
– 2个客户端: 每个客户端上一个IOzone.
iozone -a -s 2g -i 0 -i 1 -f /mnt/glusterfs/iozone0 -Rb 2Clt2Svr-Dtbt-2G.xls -+r
41
Gluster FS*: 卷选项的优化
• Gluster FS* 卷
– 类型: Distributed
– 卷选项
读大文件
io-thread-count: 16->64
cache-size: 32MB->2GB
cache-max-file-size 16384PB
cache-min-file-size 0
写大文件
write-behind-window-size:1MB->1GB
write-behind: on
io-thread-count: 16->64
flush-behind: on
95
100
105
110
115
120
Baseline Options
Read(MB/s)
Network(MB/s)
115.698MB/s
103.286MB/s
0
20
40
60
80
100
120
Baseline Options
Write(MB/s)
Network(MB/s)5.2X
119.826MB/s
20.863MB/s
42
0
50
100
150
200
250
300
350
Read(MB/s)
Network(MB/s)
248.94
Gluster FS*: 硬件优化
0
50
100
150
200
250
Write(MB/s)
Network(MB/s)
150.783
2.2X
5.2X
5.7X
1.3X
7.2X
• Gluster FS* 卷
– 类型: Distributed
– 卷选项: 启用卷优化选项
• 硬件优化
– 用 Intel® SSD硬盘替换HDD硬盘
– 用 Intel® 10G网卡替换千兆网卡
Baseline: 禁用卷优化选项; Options: 启用相关卷优化选项; SSD: bricks建在SSD硬盘上; 10G: 客户端和服务器都使用10G网卡
43
Gluster FS*: 压力测试
• Gluster FS* 卷
– 类型: Distributed
– 卷选项: 启用卷优化选项
– 12 Bricks: 6 SSD, 6 HDD
iozone -s 24g -r 16m -i 0 -i 1 -t 12 –F iozone0 iozone1 iozone2 iozone3 iozone4 iozone5 iozone6 iozone7 iozone8 iozone9 iozone10 iozone11 -Rb [email protected] -+r
0100200300400500600700800900
1000
Performance
478.358
802.676
689.156
937.337
WriteReadNetwork(Write)Network(Read)
MB/s
44
0
50
100
150
200
250
300
350
Read(MB/s)
Network(MB/s)
317.006
93.803
322.687
Gluster FS*:Striped卷调优
050
100150200250300350400
Write(MB/s)
Network(MB/s)
158.869
355.532
3.38X
3.25X
3.29X
1.19X 3.9X
• Gluster FS*卷
– 类型: Striped
– 卷选项
• 硬件优化
– 用 Intel® SSD硬盘替换HDD硬盘
– 用 Intel® 10G网卡替换千兆网卡
Baseline: 禁用卷优化选项; Options: 启用相关卷优化选项; SSD: bricks建在SSD硬盘上; 10G: 客户端和服务器都使用10G网卡
45
调优方法总结 • GlusterFS*卷选项优化
– 读大文件
io-thread-count: 64
cache-size: 2GB
cache-max-file-size and cache-min-file-size
– 写大文件
write-behind-window-size: 1GB
write-behind: on
io-thread-count: 64
flush-behind: on
• 硬件优化 – 用Intel® SSD硬盘取代HDD硬盘
– 用Intel® 10G网卡取代千兆网卡
49
关于这一主题的其它信息,请参照:
• 其他会议 – TECS003 - Lustre*: 现在归英特尔所有的 Exascale 文件系统 - 房间 306B 17:00
• 展位演示 – Teamsun* OpenStack* Swift* Scale-Out storage solution based on Intel 10GBE
– Customer Application Case Study: Intel® Xeon Phi™ Platform After Porting and
Tuning
– Resource Scheduler & Performance Monitoring for Intel® Xeon® Processor &
Intel Xeon Phi Hybrid Cluster
• 更多的WEB信息
– http://www.intel.cn/content/www/cn/zh/ethernet-controllers/ethernet-
controllers.html (Chinese)
– http://www.intel.cn/content/www/cn/zh/solid-state-drives/solid-state-drives-
ssd.html (Chinese)
– http://www.intel.cn/content/www/cn/zh/intelligent-systems/embedded-software-
tools-for-developers-to-debug-and-optimize.html (Chinese)
50
Legal Disclaimer INFORMATION IN THIS DOCUMENT IS PROVIDED IN CONNECTION WITH INTEL PRODUCTS. NO LICENSE, EXPRESS OR IMPLIED, BY ESTOPPEL OR OTHERWISE, TO ANY INTELLECTUAL PROPERTY RIGHTS IS GRANTED BY THIS DOCUMENT. EXCEPT AS PROVIDED IN INTEL'S TERMS AND CONDITIONS OF SALE FOR SUCH PRODUCTS, INTEL ASSUMES NO LIABILITY WHATSOEVER AND INTEL DISCLAIMS ANY EXPRESS OR IMPLIED WARRANTY, RELATING TO SALE AND/OR USE OF INTEL PRODUCTS INCLUDING LIABILITY OR WARRANTIES RELATING TO FITNESS FOR A PARTICULAR PURPOSE, MERCHANTABILITY, OR INFRINGEMENT OF ANY PATENT, COPYRIGHT OR OTHER INTELLECTUAL PROPERTY RIGHT. • A "Mission Critical Application" is any application in which failure of the Intel Product could result, directly or indirectly, in
personal injury or death. SHOULD YOU PURCHASE OR USE INTEL'S PRODUCTS FOR ANY SUCH MISSION CRITICAL APPLICATION, YOU SHALL INDEMNIFY AND HOLD INTEL AND ITS SUBSIDIARIES, SUBCONTRACTORS AND AFFILIATES, AND THE DIRECTORS, OFFICERS, AND EMPLOYEES OF EACH, HARMLESS AGAINST ALL CLAIMS COSTS, DAMAGES, AND EXPENSES AND REASONABLE ATTORNEYS' FEES ARISING OUT OF, DIRECTLY OR INDIRECTLY, ANY CLAIM OF PRODUCT LIABILITY, PERSONAL INJURY, OR DEATH ARISING IN ANY WAY OUT OF SUCH MISSION CRITICAL APPLICATION, WHETHER OR NOT INTEL OR ITS SUBCONTRACTOR WAS NEGLIGENT IN THE DESIGN, MANUFACTURE, OR WARNING OF THE INTEL PRODUCT OR ANY OF ITS PARTS.
• Intel may make changes to specifications and product descriptions at any time, without notice. Designers must not rely on the absence or characteristics of any features or instructions marked "reserved" or "undefined". Intel reserves these for future definition and shall have no responsibility whatsoever for conflicts or incompatibilities arising from future changes to them. The information here is subject to change without notice. Do not finalize a design with this information.
• The products described in this document may contain design defects or errors known as errata which may cause the product to deviate from published specifications. Current characterized errata are available on request.
• Intel product plans in this presentation do not constitute Intel plan of record product roadmaps. Please contact your Intel representative to obtain Intel's current plan of record product roadmaps.
• Intel processor numbers are not a measure of performance. Processor numbers differentiate features within each processor family, not across different processor families. Go to: http://www.intel.com/products/processor_number.
• Contact your local Intel sales office or your distributor to obtain the latest specifications and before placing your product order. • Copies of documents which have an order number and are referenced in this document, or other Intel literature, may be
obtained by calling 1-800-548-4725, or go to: http://www.intel.com/design/literature.htm • Intel, Xeon, Xeon Phi, Sponsors of Tomorrow and the Intel logo are trademarks of Intel Corporation in the United States and
other countries.
• *Other names and brands may be claimed as the property of others. • Copyright ©2013 Intel Corporation.
51
• Any software source code reprinted in this document is furnished under a software license and may only be used or copied in accordance with the terms of that license.
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
• Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark* and MobileMark*, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more information go to http://www.intel.com/performance.
Legal Disclaimer
52
Intel's compilers may or may not optimize to the same degree for non-Intel microprocessors for optimizations that are not unique to Intel microprocessors. These optimizations include SSE2, SSE3, and SSE3 instruction sets and other optimizations. Intel does not guarantee the availability, functionality, or effectiveness of any optimization on microprocessors not manufactured by Intel. Microprocessor-dependent optimizations in this product are intended for use with Intel microprocessors. Certain optimizations not specific to Intel microarchitecture are reserved for Intel microprocessors. Please refer to the applicable product User and Reference Guides for more information regarding the specific instruction sets covered by this notice. Notice revision #20110804
53
Risk Factors The above statements and any others in this document that refer to plans and expectations for the first quarter, the year and the future are forward-looking statements that involve a number of risks and uncertainties. Words such as “anticipates,” “expects,” “intends,” “plans,” “believes,” “seeks,” “estimates,” “may,” “will,” “should” and their variations identify forward-looking statements. Statements that refer to or are based on projections, uncertain events or assumptions also identify forward-looking statements. Many factors could affect Intel’s actual results, and variances from Intel’s current expectations regarding such factors could cause actual results to differ materially from those expressed in these forward-looking statements. Intel presently considers the following to be the important factors that could cause actual results to differ materially from the company’s expectations. Demand could be different from Intel's expectations due to factors including changes in business and economic conditions; customer acceptance of Intel’s and competitors’ products; supply constraints and other disruptions affecting customers; changes in customer order patterns including order cancellations; and changes in the level of inventory at customers. Uncertainty in global economic and financial conditions poses a risk that consumers and businesses may defer purchases in response to negative financial events, which could negatively affect product demand and other related matters. Intel operates in intensely competitive industries that are characterized by a high percentage of costs that are fixed or difficult to reduce in the short term and product demand that is highly variable and difficult to forecast. Revenue and the gross margin percentage are affected by the timing of Intel product introductions and the demand for and market acceptance of Intel's products; actions taken by Intel's competitors, including product offerings and introductions, marketing programs and pricing pressures and Intel’s response to such actions; and Intel’s ability to respond quickly to technological developments and to incorporate new features into its products. The gross margin percentage could vary significantly from expectations based on capacity utilization; variations in inventory valuation, including variations related to the timing of qualifying products for sale; changes in revenue levels; segment product mix; the timing and execution of the manufacturing ramp and associated costs; start-up costs; excess or obsolete inventory; changes in unit costs; defects or disruptions in the supply of materials or resources; product manufacturing quality/yields; and impairments of long-lived assets, including manufacturing, assembly/test and intangible assets. Intel's results could be affected by adverse economic, social, political and physical/infrastructure conditions in countries where Intel, its customers or its suppliers operate, including military conflict and other security risks, natural disasters, infrastructure disruptions, health concerns and fluctuations in currency exchange rates. Expenses, particularly certain marketing and compensation expenses, as well as restructuring and asset impairment charges, vary depending on the level of demand for Intel's products and the level of revenue and profits. Intel’s results could be affected by the timing of closing of acquisitions and divestitures. Intel’s current chief executive officer plans to retire in May 2013 and the Board of Directors is working to choose a successor. The succession and transition process may have a direct and/or indirect effect on the business and operations of the company. In connection with the appointment of the new CEO, the company will seek to retain our executive management team (some of whom are being considered for the CEO position), and keep employees focused on achieving the company’s strategic goals and objectives. Intel's results could be affected by adverse effects associated with product defects and errata (deviations from published specifications), and by litigation or regulatory matters involving intellectual property, stockholder, consumer, antitrust, disclosure and other issues, such as the litigation and regulatory matters described in Intel's SEC reports. An unfavorable ruling could include monetary damages or an injunction prohibiting Intel from manufacturing or selling one or more products, precluding particular business practices, impacting Intel’s ability to design its products, or requiring other remedies such as compulsory licensing of intellectual property. A detailed discussion of these and other factors that could affect Intel’s results is included in Intel’s SEC filings, including the company’s most recent Form 10-Q, report on Form 10-K and earnings release. Rev. 1/17/13