惠普“动成长企业”
及资源虚拟化技术
纪越
技术咨询部
中国惠普有限公司
2005年7月22日
中国银行业信息化系统建设研讨会
议程
HP“动成长企业”及IT架构
HP资源虚拟化技术
HP与中国同步成长
当前企业在变化环境中面临的主要业务挑战
必须适应不断变化的企业生态环境
内部
外部
信息
流程
过长的/ 复杂的 业务流程
业务环境 快速变化
缺乏交流 / 信息共享
对客户缺乏 信息传播
员工缺乏知识 和培训
竞争环境 快速变化
客户需求 快速变化
需要供应链运行效率
大量的文本工作
在企业内有限的 信息共享
缺乏与客户沟通, 如产品好处和 变化
缺乏产品知识
不能有效利用 IT / 销售工具
更佳的产品出售物
缺少竞争信息
监管和遵纪要求变化
要求追赶技术进步
客户期望递增
缺少产品信息反馈
内部面对陈旧的 IT 系统和复杂的流程
外部需要应对快速,不断变化的环境
基础设施
陈旧的遗留 IT 基础设施/ 系统
缺乏 IT 系统 整合
过时
不灵活 / 无效率
不可靠 (如安全性差)
部门间不兼容
自动化 不够
希望更自动化, 如库存管理和客户管理
挑战:
企业数据中心有未用的服务器容量但仍不能满足需求
Most reports put average utilization at approx 30%
Tremendous amount of unutilized capacity
Yet these systems are unable to handle the load
0
10
20
30
40
50
60
70
80
90
100
Server Average Utilization
Servers
Utilization at an actual HP customer
企业必须适应不断变化的市场生态环境
在21st世纪, 变化是持久的
企业必须管理三个关键的必需因素:
赢利能力 (管理成本)
客户满意 (服务质量)
应对竞争压力 (降低风险)
“灵活性” – 是为不确定的未来需求提供准备和迅速反映
IT经济= 成本 + 质量 + 风险 + 灵活性
“最终生存下来的物种，不是最强大的，也不是最聪明的，而是那些最能适应变化的物种。” – 达尔文
信息技术演变历程与适应性IT基础体系架构特征
“Big Iron” Age
(钢铁时代)
Distributed and Internet Revolution
(分布式与互联网革命时代)
Age of Maturation and Platform Shakeout
(成熟与平台淘汰的时代)
Renaissance
(信息复兴时代)
1980s
1990s
2000-2005
2006-2010
标准硬件
标准应用
高品质的服务
政策性的管理
平台选择
没有标准
各式各样的供应商
凌乱的管理，缺少工具
低品质的服务
平台选择少
应用趋向标准一致
工具改善
扩展服务等级协议
运作成熟
网络系统管理供应商淘汰
主动式管理
有效率的动机分析
自我治疗应用管理
策略管理
自动资源分配
自我发现服务
Resource: Gartner Group 01/04/2004
动成长企业的定义
业务
信息技术
最终的适应状态：业务和IT保持同步，从而充分利用变化
在不断变化的环境中,能自如应对变化和持续增长的企业
HP动成长企业架构
HP动成长达尔文参考体系结构(DRA)
应用
构建与集成 标准化, 模块化，简易化, 集成化
适应性IT基础架构
供应商
雇员
客户
扩展与链接
管理与控制
评估，建议，行动
测量与评估 时间、范围、难易度
业务流程
测量，评估 和维持业务和IT的动态链接
构建和集成 企业的异构IT环境
管理和控制 企业的业务过程，应用，和IT环境
扩展和链接 水平地延伸和链接企业的业务过程和业务应用 — 从企业供应商到企业客户。
将适应变化的基因植入企业架构中
流程，组织机构的变化
信息质，量，流向的变化
应用修改，变动，集成（EAI）
资源的扩展部署
达尔文参考体系结构:
设计准则
面向服务的体系结构(SOA) /虚拟化
保护原有系统投资
避免厂商锁定
权衡行业投资
使能并行创新, 快速响应变化
多资源, 外包/内包的灵活性
模型驱动的自动化(MDA)
变化成本低 & 可预见
使服务生命周期交付和管理自动化
模型获得, 共享和使经历最佳实践的应用自动化
无需重写服务而修改策略
复合SOA,虚拟化, MDA实现动成长目标
集成的服务视图
IT 业务管理
服务交付管理
服务交付
业务服务
Business Services
应用服务
Application Services
基础设施服务Infrastructure services
业务目标和策略
达尔文参考体系结构: 执行视图
先进的数据中心IT基础体系结构
达尔文参考体系结构
达尔文参考体系结构为数据中心的远见和策略提供技术框架
Integrate & Orchestrate`
业务策略
Demand
Infrastructure
虚拟化资源
Resources
Virtualization
Sourcing
Sharing
Pooling
Clients
Imaging/
Printing
Servers
Storage
基础设施服务
Web/Grid
services
…
Document/Imaging
services
Network
Software
Environ-
ment
ERP
CRM
Collaboration
Productivity
…
Supply
业务流程
Information
应用服务
Operational Reference Models
Business Rules and Policies
Business Process Automation
Data Stds
Meta data
Product
…
Business Docs
Master Data
Mobility/Network
services
Database
services
Registry
services
Trust & Security
services
Collaboration
services
Key Performance Indicators
IT Business Support
Mgt Integration Channel
IT Operations Support
Manage & control
Lifecycle
Policy- based control
适应性IT基础设施
适 应 性 管 理
一个开放的体
系结构，它为硬
件和软件定义了
一套规范使能实
现虚拟化，异构
资源的管理和控
制，和实现基础
设施流程自动化
以支持业务流程
的动态管理。
业务应用
新的IT基础体系结构的特点
工业标准 & 模块化基础结构
刀片服务器
模块化存储
多操作系统能力
灵活的分区
Windows and Linux and UNIX
.NET and J2EE
存储器/服务器 网络标准
Itanium and IA32
服务器
多操作系统
连续和安全运行
集群高可用性
容错/容灾
动态资源优化
按需供应
分区技术
公用数据中心(UDC)
零延迟企业(ZLE)
自动的 & 智能管理
OpenView套件
System Insight Manager
Utility Controller
WLM
创新技术
标准化 虚拟化 集成管理
运用业务导向型信息化规划方法论
－ 匹配和转化
愿景勾画
创建业务
情形
规划和设计
执行
变化
推进
对现有问题和将来方向的普遍理解
产生一个公认和具有动力的远景
定义变革的路线/计划，以及解决方案
建立并验证适应变化的业务模式和框架
执行并监控项目，体现业务利益和知识获取
TO-BE business plans
AS-IS situation
BITA alignment+gap analysis
Quick wins
Transformation opportunities
Readiness for change assessment
Business & IT Roadmaps for change
Transformation scenarios, strategies & choices
COULD-BE: scenarios + business benefits & goals
BTO (measures + CSFs)
Stakeholder needs analysis
Shared vision + buy-in
Team building
Business case & transition strategies
New business model, principles & processes
Business Info. architecture
New organization structures
IS Application portfolios
IT architecture framework
Benchmarks/ technical POC/Prototype
Change plan (skill & competence gap assessment)
Transition strategy & roadmap
Program charter & plans
Solution Blueprints
Communication plans
3Rs, HR & education plans
Benefits dependency models
Measures & targets
‘RAID’ assessments
Project plans
Business applications.
Business information systems
IT Infrastructure
Trained users & learning communities
Live managed services
Quality & benefits management and measurement systems
Business transformation knowledge & learning's
探索和分析
动成长企业的优势:纲举目张
管理成本:
降低的运营和购置成本
管理固定成本
降低风险:
确保业务运营的安全性
与连续性
提高灵活性:
使 IT 环境能适应业务需求变化
使企业能够响应不断变化的市场和竞争压力
利用可变能力作为一种竞争优势
优化固定成本与可变
成本间的平衡
管理变化成本
投资回报最大化
降低创新风险
降低技术实施影响
提高动态环境下的业绩和满意度
改善业绩:
端对端跨企业改善业务运营业绩
增加客户和雇员的满意度
HP适应性IT基础架构
是动成长企业达尔文参考体系结构的重要组成
虚拟
管理
Integrate & Orchestrate`
企业业务策略
Demand
Infrastructure
虚拟化资源
Resources
Virtualization
Sourcing
Sharing
Pooling
Clients
Imaging/
Printing
Servers
Storage
Infrastructure Services
Web/Grid
services
…
Document/Imaging
services
Network
Software
Environ-
ment
ERP
CRM
Collaboration
Productivity
…
Supply
业务流程
Information
应用服务
Operational Reference Models
Business Rules and Policies
Business Process Automation
Data Stds
Meta data
Product
…
Business Docs
Master Data
Mobility/Network
services
Database
services
Registry
services
Trust & Security
services
Collaboration
services
Key Performance Indicators
IT Business Support
Mgt Integration Channel
IT Operations Support
Manage & control
Lifecycle
Policy- based control
适应性IT基础设施
适 应 性 管 理
标准
议程
HP“动成长企业”及IT架构
HP资源虚拟化技术
HP与中国同步成长
虚拟化技术:
关注服务,共享资源, 按使用付费 (PPU)
网络虚拟化
服务器虚拟化
存储虚拟化
应用虚拟化
数据中心 虚拟化 (ITU)
服务器
集群
分区
刀片
客户机
虚拟服务器 环境
基于服务器
基于阵列
基于网络
集线器/路由器
交换机
VLAN
整体
客户机/服务器
Web 服务
OGSA
网格
通过池化和共享IT基础设施资源使能实现资源优化利用和按需自动供给
HP 虚拟服务器环境
释放IT 投资的价值 优化资源利用率
提高业务灵活性 动态资源分配
保证服务水准 实时检测，提示和行动
智能策略引擎
HP 虚拟服务器环境 for HP-UX 11i
实时自动扩展和收缩虚拟服务器
HP-UX 11i – 动成长企业的关键性基础: 可管理，高可用，安全
服务器虚拟化技术:
资源管理
软硬分区
即时增容
群集技术
HP-UX虚拟服务器环境
Services Delivery to Partner Apps in the Adaptive Enterprise
Services Delivery
Legend
CPU
Application Load
HP-UX 11i Operating Environment
Systems Management
HP Virtualization Software
HP VSE QSS: provision a complete stack
HP VSE QSS: a complete stack, including consolidated third party apps
Batch Jobs
Low Priority
BEA WebLogic Server
Highest Priority
Oracle Database
Medium Priority
HP VSE QSS: allocate initial CPUs
High Availability
HP VSE QSS: Apply Initial Load and distribute additional CPUs
HP VSE QSS: Increase Oracle Load
HP VSE QSS: Increase Other Workloads and redistribute CPUs
HP VSE QSS: Reduce Oracle Workload and redistribute CPUs
BEA Load
Services Delivery of CPUs across applications based on application load and priority
Oracle Load
Batch Load
HP-UX虚拟服务器环境
In production use today
Services Delivery Management
Serviceguard
WLM
iCOD
Services Delivery
100%: SAP
Unpurchased processors
X
80% Oracle CRM
20% Security
50% SAP
40% Oracle CRM
10% Security
Legend
Oracle CRM
Security
SAP
Services Delivery of CPUs across all systems
Time
Consolidated Server
(5 CPU system)
(5 CPU system)
(6 CPU system)
HP 虚拟服务器环境的优势
最丰富的，集成的虚拟环境
Unix市场上唯一基于目标的负载管理能力
第一个完全虚拟化高可用的跨数据中心的Oracle 9iRAC解决方案 (SGeRAC)
扩展服务器虚拟化至中间件和数据库层-HP VSE Quick Start Solution for BEA & Oracle
Licenses shipped:
10000+ Instant Capacity on Demand
160000+ Serviceguard
115000+ Workload management tools
20000+ Virtual partitions
HP Integrity and HP 9000 Server Roadmap
2004
2005
HP 9000 Superdome
HP Integrity rx4640-8
All upgrades “in-box” except as noted
Current offering
CPU: Itanium2
Madison 9M
CPU: PA-8900 (dual-CPU)
CPU: Montecito
New Chipset
HP 9000
rp7410-8, rp8400-16
rp7420-16, rp8420-32
CPU: Itanium2 Madison9M
CPU:
PA-8900 (dual-CPU)
CPU: Montecito
New Chipset
*Not available at initial processor release
HP 9000 rp4440-8
CPU: PA-8900 (dual-CPU)
CPU: Montecito
HP Integrity rx2600-2
HP 9000 rp3440-4
CPU: PA-8900 (dual-CPU)
CPU: Montecito
Timeframes not to scale
Revision July-04
HP Integrity rx1600-2
CPU: DP Itanium
4-128P
2-32P
2-16P
1-8P
1-4P
1-2P
2006
New Chipset
CPU: PA-8900
New Chipset
CPU: PA-8900
CPU: Next Gen & Chipset
DP Itanium
HP Integrity Superdome
HP Integrity
rx7620-16, rx8620-32
Plans subject to change
PCI-E
PCI-E
PCI-E
PCI-E
PCI-Express
New Chassis Intro.
New 8p Server & Chipset
CPU: Montecito
New 4p Server & Chipset
CPU: Montecito
DDR-II
DDR-II Memory
DDR-II
DDR-II
DDR-II
DDR-II
DDR-II
DDR-II
CPU: Itanium2 Madison9M
CPU: Itanium2 Madison9M
CPU: Montvalle
CPU: Montvalle
HP关键业务系统的发展趋势：
安腾芯片一统天下 操作系统百花齐放
HP Servers
HP NonStop Servers
HP Blade Server
HP Servers
HP NonStop Servers
HP e3000 Servers
HP AlphaServer Systems
HP Blade Server
server platform
HP-UX (with Tru64 features)
Linux®
HP NonStop Kernel
HP OpenVMS
Windows® /64
HP-UX
HP Tru64 UNIX®
HP MPE/ix
Linux®
HP NonStop Kernel
HP OpenVMS
operating system
processor technology
future
today
Itanium™ processor family
PA-RISC
Itanium™ processor family
Alpha
MIPS®
HP Integrity server 在腾飞:
Q4 FY04
HP Integrity server revenue
60% Q3 to Q4
244% FY2004 over FY2003
HP-UX 11i on Integrity servers
64% Q3 to Q4
341% Q4 FY2004 over Q4 FY2003
Windows on Integrity servers
68% Q3 to Q4
186% Q4 FY2004 over Q4 FY2003
(1/3 of all Integrity Superdome shipments are with Windows)
全球最大100个客户中 40% 使用基于Itanium 服务器
HP 占有基于 Itanium 服务器50％ 以上的市场份额
Integrity 平台上目前拥有超过3,000 个 ISV 应用软件 ( 比去年增加 200% ）
过去180天中 企业用户安装了180个正式运行的 SAP系统
ISV wave of momentum for HP Integrity
HP Integrity midrange announcement
HP Integrity announcement
3195
February 25th, 2005
4500
October 31st, 2005
SuperDome无与伦比的投资保护特性
PA-8800
Itanium
Cabinet 1
Cabinet 0
X B A R
X B A R
X B A R
X B A R
CELL 0
CELL 1
CELL 2
CELL 3
CELL 4
CELL 5
CELL 6
CELL 7
CELL 8
CELL 9
CELL 10
CELL 11
CELL 12
CELL 13
CELL 14
CELL 15
I/O Chassis
I/O Chassis
I/O Chassis
I/O Chassis
I/O Chassis
I/O Chassis
I/O Chassis
I/O Chassis
I/O Chassis
I/O Chassis
I/O Chassis
I/O Chassis
I/O Chassis
I/O Chassis
I/O Chassis
I/O Chassis
Remote Link
Remote Link
Remote Link
Remote Link
HP 服务器 体系结构
HP 高端服务器基础单元 Cell Block Diagram
S
D
R
A
M
To crossbar
PCI-X Sys Bus Adapter
Cell Controller
8 GB/sec
PCI-X Host Bridge
...
倍提升
GB/sec
4倍提升
16 GB/sec
4x 288 bits
125 MHz
16 GB/sec
P
P
P
P
16 individual
“ropes”
PCIX
2x
Mem buffer
2x
Mem buffer
2x
Mem buffer
2x
Mem buffer
2倍提升
8GB/sec
市场上颗粒度最细的扩容方式
内存槽多,升级颗粒度细,投资保护好
SX2000芯片组
Cell
Controller
PCI-X System
Bus Adapter
2X MEM
BUFFER
PCI-X
Host
Bridge
32 SDRAM
DIMMs
2X MEM
BUFFER
2X MEM
BUFFER
2X MEM
BUFFER
PCI/PCI-X
16 Links
Itanium® 2 Buses
* Accepts single or dual processor modules
** Not used in 8 socket systems
CPU
Socket*
CPU
Socket*
CPU
Socket*
CPU
Socket*
PCI-X
Host
Bridge
…
…
Crossbar **
Cell Board
I/O Subsystem
Maximizes application performance from 4 to 128 processors
DDRII
PCI-X
PCI Express
New
New
New
New
New
Plus:
Link level retry
In-box networking
Reduced Latency (>25%)
CPU BW
2-3x IO BW
3-4x Fabric BW
~2x Mem BW
4x Memory Capacity
Note: Futures listed are not commitments, can change any time
利用分区技术构建数据中心
分区是在一个或多个服务器内的物理或逻辑结构 为隔离运行环境 并提供 IT资源动态再配制的灵活性 和确保应用能享有 避免无关事件的影响 否则它将引起破坏,中断, 或性能下降.
城市内的 建筑物
建筑物内的 单元住宅
单元住宅内的房间
房间内的 壁橱
多节点
物理硬分区
单节点
物理硬分区
虚拟逻辑分区
资源软分区
性能 / 处理器
#1
#16
#32
分区越多，性能越高
CPUs
HP 独特的分区技术
没有开销，完全隔离
使用 nPAR 分区
增加的可伸缩性
不使用 nPAR 分区
性能 / 处理器
#1
#16
#32
CPUs
HP 独特的分区技术优势
没有开销，完全隔离
分区越多，性能越高
OpenVMS
并且支持多操作系统
利用分区技术构建数据中心
HP分区技术
多节点物理硬分区 (at least 1 OS image per node)
在一个节点内的物理硬分区 (multiple OS images with HW isolation)
OS image with HW
isolation
OS image with HW
isolation
OS image with HW
isolation
在一个物理硬分区
内的虚拟分区 (multiple OS images with SW isolation)
在一个操作系统内的资源软分区 (resource allocation by app)
OS image with SW isolation
OS image with SW isolation
OS image with SW isolation
hard partition
1 OS image
Application 1
with guaranteed compute resources
Application 2
with guaranteed compute resources
Application n
with guaranteed compute resources
Based on SLOs or percentages
HyperPlex 城市内的建筑物
nPartitions 建筑物内的单元住宅
vPartitions 单元住宅内的房间
PRM (Process Resource Manager)
HP-UX WLM (Workload Manager) 房间内的壁橱
基于大型服务器的资源整合
多个操作系统和应用系统整合在一个大型机中
web server
Linux®
application server
Windows® 2000
database server
HP-UX
物理硬分区 1
物理硬分区 2
物理硬分区 3
Oracle资源数据库, 统计应用系统， 数据仓库, …
电子邮件和 文件服务， 前端应用服务
Web 服务, 防火墙, 负载均衡
增加系统利用率
增加灵活性: 多操作系统
增加可靠性: 集群
利用分区技术实现整合
SuperDome的物理硬分区和虚拟逻辑分区 配置案例
有关键硬件隔离的4 个不同客户的生产环境
有硬件物理隔离 (电隔离)的测试和生产环境
nPartition 1 为测试环境
nPartition 2 为生产环境
Virtual partition 1 为客户 1 有 1个或多个应用
Virtual partition 2 为客户 2 有 1个或多个应用
Virtual partition 4 为客户 4 有 1个或多个应用
Virtual partition 3 为客户 3 有 1个或多个应用
在同一服务器中有多个应用实现硬件隔离
伸缩自如 大小由之的硬分区技术
SuperDome Partitions
16
1
1
4
安全性最佳又兼具灵活性的资源配置手段
真正的电气物理隔离
硬分区可在线增加/替换
可以对硬分区在线维修
没有开销问题
1
32
腾挪灵活的软分区技术
高度灵活性资源配置手段
无电器隔离，安全性差。
一个软件分区的硬件故障会影响到其它分区。
不推荐使用在关键应用的生产环境，可使用于开发测试环境。前端应用服务器
需占用主机资源
HP-UX 分区的威力:
混合硬分区的隔离度和软分区的灵活性
nPartition 2 “dev & test”
nPartition 1 “production”- HPUX 11i
cpu0
CC
memory
I/O
cpu1
cpu2
cpu3
memory
cpu0
CC
memory
I/O
cpu1
cpu2
cpu3
memory
cpu0
CC
memory
I/O
cpu1
cpu2
cpu3
memory
cpu0
CC
memory
I/O
cpu1
cpu2
cpu3
memory
cpu0
CC
memory
I/O
cpu1
cpu2
cpu3
memory
cpu0
CC
memory
I/O
cpu1
cpu2
cpu3
memory
cpu0
CC
memory
I/O
cpu1
cpu2
cpu3
memory
cpu0
CC
memory
I/O
cpu1
cpu2
cpu3
memory
vPar 2
HPUX11i + web
vPar 3
HPUX 11i, large I/O buffer cache
vPar 4
HPUX 11i, tuned for large memory
vPar 1
HPUX 11i + patch
cpu0
CC
memory
I/O
cpu1
cpu2
cpu3
memory
cpu0
CC
memory
I/O
cpu1
cpu2
cpu3
memory
cpu0
CC
memory
I/O
cpu1
cpu2
cpu3
memory
cpu0
CC
memory
I/O
cpu1
cpu2
cpu3
memory
cpu0
CC
memory
I/O
cpu1
cpu2
cpu3
memory
cpu0
CC
memory
I/O
cpu1
cpu2
cpu3
memory
nPartitions provide electrical hardware isolation - separate systems, I/O, boot, CPU, memory, etc.
vPars are software isolated (OS, applications) for different customers, different OS tuning - each tailored by number of CPUs, amount of memory and I/O per customer and application mix
软件隔离
动态配置
单CPU粒度
硬件电气隔离
Server consolidation example
vPars 的进步. “vPars/VM”
vPars
vPars/VM
PA support
Negligible
overhead
Maturity
(end ’01)
Before Sept. ‘05
Before vPars/
VM &:
- heavy I/O
workloads
- Metro/
Continental
clusters
- direct I/O control
Integrity mid-range,
high-end support
Dedicated resources
Separate OS
instance
CPU granularity
Dynamic CPU
migration
VSE co-existence
1-8 way Integrity
support
Sub-CPU
Shared I/O
Virtualized resources
CPU, I/O over-
subscription
Dynamic CPU, I/O
resourcing
Security isolation;
separate mgmt. OS
Backplane comm.
Multi-OS design
HP restricted; customer viewable under NDA; subject to change without notice
在sub-CPU 粒度下构建虚拟 SMP
text
CPU
text
CPU
text
CPU
text
CPU
text
CPU
text
CPU
Higher resource utilization
VM4
VM5
VM6
VM7
VM8
VM2
VM9
VM3
VM1
Uni-processor virtual machines with sub-CPU granularity -
Virtual machine scalability
VM1
VM3
VM2
Hardware partitioning, virtual SMPs -
All available in one package – HP Integrity Virtual Machine
VM
2
VM
1
VM
4
VM
5
VM
7
VM
8
VM
3
VM
6
Virtual machine scalability and Higher resource utilization
Dynamic I/O Sharing
Host (Integrity VM + platform OS)
Virtual server’s I/O packets directed to I/O cards by the Platform Manager
Virtual Server 1
Virtual Server 2
Virtual Server 3
I/O card can be dedicated to a virtual machine for performance isolation
HP-UX
Sun (Solaris)
IBM (AIX)
nPartitions
Virtual Partitions
HP-UX WLM
PRM
HyperPlex
SRM
$$
AIX WLM
Domains
SP $$
No high speed connect
Hard partition with multiple nodes
Resource partitioning
Virtual partitions with S/W isolation
Hard partition within a node
Goal-based resource partitioning
HP 服务器: 软硬兼备 文武双全
DH Brown: HP shines in UNIX resource management
HP StorageWorks 产品家族
存储虚拟化
Resource efficiency
Scalable capacity
网络存储使能提供
更大的可扩展性
有效的资源共享
更高的系统可用性
更短的备份时间和空间
实时信息服务
实时可扩展性
存储网格使能提供
新的信息服务
Storage
Grid
Network
Storage
DAS
将巨大的存储物理容量，变成单一的大型、可管理存储数据池， 提供给数据中心内各种服务器使用。
存储虚拟化:分类
在单个存储阵列中虚拟化存储容量, 扩展存储阵列的容量
适用于大数据卷，异构主机 / 单个存储阵列环境
Intelligence 驻留在存储阵列控制器中
实现方法 - EVA
适用于异构环境
为网络中任何主机虚拟化任何存储阵列的容量
Intelligence 驻留在网络中
存储公用的先决条件
实现方法 - SANworks CASA/VersaStor 非对称存储池
跨异构阵列虚拟化存储容量, 扩展存储阵列的容量
适用于单服务器 / 单集群环境
Intelligence 驻留在主机
实现方法 - 虚拟复制软件 (SANworks Virtual Replicator)
基于网络的虚拟存储 (Fabric Level)
基于服务器的虚拟存储 (Server Level)
基于阵列的虚拟存储 (Storage level)
本地的虚拟化管理—(centralized management)
IP, FC, ATM
remote replication
异构存储环境
异构服务器环境
remote
SAN
基于网络的虚拟化管理
存储网络(SAN)
存储小家电
HP 虚拟阵列解决服务器洪水泛滥问题
传统阵列难以应付对局部数据的频繁访问
虚拟统阵列对局部数据的频繁访问游刃有余
Server1
Server2
Server3
Lvm1
Lvm2
Lvm3
Lvm4
Lvm5
Lvm6
Lun1
Lun2
Lun3
Lun4
VA/EVA
五大绝技融于EVA一身，难怪笑傲江湖
虚拟化RAID技术 Virtual Raid
虚拟化数据智能均摊技术 VLeveling
虚拟化数据智能修复技术 VActive Hotspare
虚拟化数据快照技术 VSnapShot
智能快照式克隆技术 VSnapClone
HP-UX Workload Manager 负载管理器(WLM)
基于目标的策略引擎管理下的虚拟服务器环境
HP-UX 11i WLM automatically reconfigures CPU entitlements to satisfy SLOs in priority order
application A
application B
application C
Priority 1
response time SLO:
交易须在1秒内完成
交易响应时间由应用响应时间衡量器application response measurement (ARM) 提供或由其他方式提供
Priority 2
job duration SLO:
批处理须在1小时之内结束
处理时间由操作系统核心提供
Priority 3
CPU entitlement SLO:
分配 50% of CPU 资源
50% Real Memory
50% Disk I/O
25% Real Memory
25% Disk I/O
25% Real Memory
25% Disk I/O
WLM 的下一步HP global Workload Manager
PRM:
static resource management
WLM:
dynamic workload management
gWLM
dynamic, multi-system workload management
50%
50%
App A
App B
App A
App B
App G
App H
App E
App F
App C
App D
App A
App B
Policies
Monitoring and Reporting
HP restricted; customer viewable under NDA; subject to change without notice
Site A
Site B
Dark Fiber
DWDM Device
DWDM Device
Cache Fusion
Replicated and Synchronized Data for one single RAC database
via SLVM with MirrorDisk/UX or CVM
Load Balancing (access layer)
WLM
WLM
100 kms
容灾和高可用的完美结合
HP Serviceguard extension for Oracle RAC
SAN
SAN
Active/active environment with transparent access to applications and data!
Ultravailable Networks
News
数据中心管理
HP OpenView 新一代数据中心管理平台
Service
Information Portal
(服务信息门户)
Application Smart Plug-Ins
(应用管理)
Insight Manager
Operations(运营)
Services Desk
(服务台)
Storage Area Manager(SAN管理)
Internet Services/ Transaction Analyzer
Data Protector(数据备份)
Performance(性能)
Network Node Mgr
(网络节点管理)
动成长企业信息管理中心功能结构
基于ITSM的管理体系结构
运营管理
报表管理
故障管理
性能管理
资源管理
安全管理
应用管理
综合事件管理
配置管理
问题管理
变更管理
服务级别 管理
工作任务管理
事件管理
服务管理层
管理功能层
组件管理层
被管理对象
存储器
网络
服务器
应用/Web服务
数据库
备份/恢复
流程管理
帮助台 管理
服 务 器 管 理
网 络 管 理
存 储 管 理
中 间 件 管 理
数 据 库 管 理
备 份 管 理
信周 息期 生管 命理
非管 IT理 设 备
业 务 应 用 一
业 务 应 用 二
业 务 应 用 三
业 务 应 用 四
业 务 应 用 五
业 务 应 用 六
数据中心信息管理
ILM将解决以下三大信息管理挑战
数据保管
公司和政府法规要求制定保管政策
在对公司提出的法律诉讼中，公司可能会接到传票，要求提供电子邮件通信和相关文件
数据管理
信息量加速增长
需要显著降低管理成本，同时维持相应的服务等级
提高文件服务器的性能
减少备份时间
参考信息管理
参考信息（静态内容） 未获得充分利用，如果能够适当利用，将获得潜在的巨大商业价值
通过实现整个生命周期中信息管理的自动化，并根据其相关性、使用和服务质量目标，ILM有效解决了这些问题
数据中心信息管理
信息生命周期管理(ILM)计划
从创建到删除
使数据存储、数据保护和数据可用性与业务需求保持一致
并自动执行特定应用政策
符合当前和未来的数据保管和审计法规
删除
保护和恢复
存档和再调用
复制与分布
创建和修改
议程
HP“动成长企业”及IT架构
HP资源虚拟化技术
HP与中国同步成长
HP现状
成立于1939年,是美国硅谷第一家 高科技公司
2004财年营业收入799亿美元
2004年财富杂志美国500强第11位, 全球500强第24位
2004年美国商业周刊全球100最佳 品牌第12位
60%的收入来自美国以外,是一家 融合不同文化的全球性公司
所有产品及服务在全球IT市场中 都位列前三位
财务状况
800亿美元年营业收入
Personal Systems
$
Enterprise Servers & Storage
$
HP Services
$
HP Financial Services
$
Software
$922M
Other
$449M
FY04 revenue: $80B
Imaging & Printing
$
Q2 FY05 Revenue
Personal Systems
30%
Enterprise Servers & Storage
19%
HP Services
18%
HP Financial Services
2%
Software and Other
1%
Imaging & Printing
30%
Revenue by segment
Revenue by region
EMEA
42%
(+10% Y/Y)
Asia Pacific
17%
(+9% Y/Y)
Americas
41%
(+4% Y/Y)
HP核心优势:成长力
0
10000
20000
30000
40000
50000
60000
70000
80000
90000
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
m$
revenue
HP营业额
HP在中国
销售及IT服务: CHP中国惠普成立于1985年, 为第一家中美合资高科技企业, 中国惠普在大陆设九大区域总部,28个城市办事处,37个支持服务中心, 超过200个金牌服务网点,惠普商学院及 IT管理学院
金融服务: 1997年经国家特许在上海成立惠普租赁公司
生产研发: 以上海为基地,设有两座工厂,一个信息产品开发中心及一个软件开发中心
全球运营支持中心: 以大连为基地, 设有业务流程外包中心, 软件开发外包中心及客户响应中心
在华采购金额每年超过100亿美元
上海
广东
深圳
成都
昆明
福州
杭州
长沙
武汉
南京
郑州
西安
天津
沈阳
济南
哈尔滨
太原
青岛
黑龙江
吉林
辽宁
内蒙古
河北
山东
山西
河南
陕西
宁夏
甘肃
新疆
江苏
安徽
江西
浙江
湖北
湖南
四川
贵州
云南
西藏
福建
广西
海南
广州
乌鲁木齐
青海
北京
Primary Office ( 9 )
Support Only Office ( 35 )
Plan to Build SOO ( 4 )
Northern
North West
North East
Eastern
South West
Southern
ShenZhen
合肥
厦门
长春
苏州
兰州
贵阳
南宁
重庆
石家庄
大连
海口
宁波
南昌
拉萨
西宁
银川
呼和浩特
中国惠普服务网点及支持服务体系
覆盖率149个城市
35个服务分支机构
中国惠普服务全国备件库(实施)
Hainan
Xizang
Qinghai
Xinjiang
Sichuan
Chengdu
KMSC
Shanghai
Beijing
TaiWan
HongKong
Guangdong
Anhui
Guangxi
Zhejiang
Jiang
su
Fujian
Jiangxi
Guizhou
Gansu
Ningxia
Shannxi
Hubei
Henan
Shanxi
Shandong
Inner
mongolia
Hebei
Hilongjiang
Jilin
Liaoning
Guangzhou
CSZL
两个备件中心，备件命中率98%
七个副中心， 备件命中率90%
38个小库， 备件命中率95%
(high-level Same Day Service)
HP的使命,市场及业务
HP是一家面向个人消费者,大中小型企业和政府机构的全球技术解决方案提供商.
HP的使命是以创新的科技及服务,带动企业发展,创造社会 效益并且为人们带来更佳的生活品质和体验.
HP的业务涵盖了IT基础设施解决方案,信息产品及接入设备,成像,打印以及全球服务等领域。
HP在市场的定位及对客户的价值
客户 + hp = everything is possible 惠普科技, 成就梦想.
HP在市场的定位及对客户的价值
2. 变革 + hp = 动成长 HP专注 “建设及管理适应性IT基础设施的解决方案”,并且与合作伙伴紧密合作,为客户实现最佳的投资回报.
HP在市场的定位及对客户的价值
3. 您 + hp = 生活的愉悦 HP与你分享科技带给生活的愉悦! HP通过丰富多彩, 轻松易用的消费类IT产品和个人解决方案, 随时随地为人们带来既个性化又完全互联的数码体验, 让用户在先进科技的帮助下享受更多生活乐趣.
四大业务集团
小公司的灵活性,大公司的资源优势
客户解决方案集团
信息产品集团
打印及成像系统集团
专业及支持服务集团
台式机
工作站
笔记本电脑
新兴技术
消费类/商业打印
数字影像
数字出版
服务器
存储设备
软件
解决方案
客户支持
管理服务
咨询与集成
业界专家
Global Business Units
个人电脑系统
笔记本电脑
平板及掌上电脑
台式机
工作站
影像及打印系统
照片打印机
彩色激光打印机
单色打印机
多功能打印机
广受赞誉的各类耗材
存储及服务器
工业标准服务器
关键业务应用系统
数据管理及存储系统
……
管理软件
业务
IT 服务
架构
应用
信息生命周期
……
业界最完善的IT产品线
Notebooks
PDAs
Desktops
Ink
Paper
Printers
Shared Printing
Digital Publishing
Servers
Storage
Bus. Critical Servers
Software
Solutions
Services
HP
IBM
Dell
Sun
EMC
Lexmark
Sony
Xerox
Canon
Hitachi
Heidelberg
Gateway
HP助您创建适应性IT架构 十大解决方案
IT适应性
稳定的
动态的
可管理与集成的
可管理与集成的资源
企业集成解决方案
IT整合解决方案
管理解决方案
动态和自动化的资源;
虚拟化解决方案
按需供应解决方案
管理服务解决方案
集成支持解决方案
金融服务解决方案
提供稳定、可扩展的基础
业务连续与高可用 解决方案
安全解决方案
业
务
灵
活
性
被
动
应
变
预
见
应
变
前
瞻
性
的
主
动
应
变
随
机
应
变
,
获
得
竞
争
优
势
创新改革力
多元化能力
成长力
惠普之道
提供企业IT运行永续的动力
谢谢！
Presentation Title
This is not an adaptive enterprise presentation, but is intended to show how HP servers support it.
Please see separate AEs overview presentation for details on this.
This presentation provides an overview of HP storage and servers from the business critical systems, industry standard server, NonStop enterprise division and network storage solutions groups, and how they fit into ESG’s strategy
Further slide sets are also available to provide a deeper level of detail on BCS, ISS, NED and NSS products
Presentation Title
Presentation Title
Presentation Title
Customer might use 1 application per server and over provision based on peak work load.
High availability contributes to under utilization as well, for example: stand by systems are very common.
Customers want to address the utilization issue – in other words address this cost issue and get better RoIT, but they definitely don’t want to do it with increased risk and at the expense of quality of service.
HP definitely considers in the server virtualization solution space all of these typical CIO issues, . cost, quality and risk. In addition, HP delivers the additional dimension of agility – the enablement to fast anticipate and react to change.
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Within the infrastructure quadrant, where you find storage and servers – the subject of this event – we are innovating in many ways, but are particularly emphasizing the three core areas of standardization, virtualization and management.
Standardization is the embracing of industry standards, the use of standardized technologies and components as far as possible to provide flexibility and value for customers. Examples of this are our move to industry standard platforms, which Scott spoke about and our leadership in driving industry standards in areas such as storage tape technology and Linux.
Virtualization is a major enabler of the adaptive enterprise. It allows resources to be allocated according to need, as well as help achieve greater utilization, providing greater degrees of efficiency and agility. Examples of our investments here include UDC – the utility data center – and virtual server environment.
After the standard, modular resources are virtualized to respond to business needs, they can be managed to anticipate needs and achieve the synchronicity between business and IT that is the characteristic of an adaptive enterprise.
Some of the key investments we are making in the area of management include systems insight manager and HP OpenView.
You will see more examples in all of these areas – and the benefits they are delivering for customers today – during the other sessions at this event and we have significant marketing campaigns for each one.
Now let’s take a little closer look at each one.
Presentation Title
Presentation Title
Presentation Title
HP Virtual Server Environment delivers dynamically sized virtual servers where each server intelligently grows or shrinks in size based on business priorities and service level objectives (SLOs). At the heart of the Virtual Server Environment is the intelligent policy engine, HP-UX Workload Manager, which performs real time assessment of the resource utilization, and then advises and acts in accordance with the preset SLOs and business priorities.
HP is the only UNIX vendor in the industry, which offers with HP-UX Workload Manager a goal-based policy engine. For example, you can ensure goals such as specific response time for applications.
HP Virtual Server Environment addresses the key challenges customers face with their IT infrastructure:
Better return on IT investments through optimized server utilization
Increased agility through the capability to allocate resources on the fly
Highest Quality of Service through continuous real time assessment, advice, and action
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Itanium being a standards, it is also much easier to create ISV enthusiasm and get a huge and unique application ramp-up in a short time.
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Presentation Title
对HP来说， nPAR 分区技术创立真正的电气隔离，它允许把一个服务器一分为二，从而大大增加整个系统的性能。在本例中，处理器 #17 几个能够提供与处理器 #1相似的功能。
Presentation Title
此外， HP nPAR 分区允许同时运行建立在不同操作系统环境的多操作系统环境。
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Let’s take a closer look at HP’s storage network solutions
Presentation Title
THIS SLIDE ANIMATED
Storage has evolved from being a peripheral device to the server to being the central shared resource in the data center.
HP pioneered this shift in the late 90’s with innovations in Network Storage.
SANs and NAS have enabled Greater scalability, Efficient resource sharing, Higher system availability, and reduced backup windows. These are all critical for storage as data continues to grow.
However: While SAN and Gateway NAS were important 1st steps, This move to Networks Storage has not accomplished everything. They are challenged to fully keep pace with the unpredictable rate of change today, and all to often there is still more to manage and greater operational complexity.
Connectedness alone is not enough for the new dimension today
The notion of Real-Time information services is still needed
CUE ANIMATION
Building on the foundation of Network Storage, Today we are sharing with you the next step on the path to achieve the Adaptive Enterprise vision for meeting real-time demands for information services and having business and IT synchronized to capitalize on change.
HP will be delivering this vision through the Storage Grid.
The Storage Grid will use Storage “Smart Cells” in addition to today’s Network storage environments to deliver storage with new axes of Real-time Scalability and Real time Information Services.
The Grid approach to storage is the enabler of new services for information in the enterprise.
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Presentation Title
Presentation Title
A stable, efficient and adaptive computing environment that is available only from HP.
Virtualization with SGeRAC, elevates RAC solutions to a new level, imagine a RAC solution distributed across two data centers up to 100 kms apart with a single data base that is replicated and synchronized and all servers are VSE enabled. What would its attributes look like:
Clients are load balanced between data centers across multiple instances of the same application or applications connected to a single virtual database.
Resources are automatically shifted based on Service Level Objectives between data centers, servers, partitions, and applications to manage work load shifts and changing computing demands.
High Availability is maintained in typical SGeRAC fashion designed to survive component and server failures and even a complete data center failure.
Disaster Tolerance is provided between data centers through distributed applications and a replicated and synchronized data base. An active-active DT solution that maintains continuous application availability while recovering from the complete loss of a data center.
Virtual SGeRAC provides an excellent ROI by leveraging all available computing resources and applying them were and when they are needed.
Presentation Title
hp OpenView delivers incredible value to your customers by managing, optimizing and lowering the costs of their existing infrastructures and data centers
Each of these hp OpenView solutions automatically monitor and manage the functions that are critical to the success of your customers business.
HP OpenView anticipates problems and pinpoints the cause before services are impacted.
Application uptime and server performance are significantly increased.
Customer satisfaction cannot be matched by the competition.
Proactive management offers the technician context sensitive console integrated troubleshooting tools for problem resolution via hp OpenView operations. In addition auto actions can be launched.
Service management starts looking at root cause and customer impact information that is retrieved from hp OpenView service desk or Remedy and loaded into the customer message attributes (CMA) and allows them to do simple work prioritization via hp OpenView operations.
Problem management involved the incident management and service level management customer facing folks via hp OpenView service desk for a managed triage process.
Value management makes visible the Service level status of the managed service via hp OpenView
HP OpenView delivers true end-to-end management, enabling IT operations to scale across infrastructure, processes, and geographies
-- in order to elevate your paradigm from infrastructure management to service management, a core prerequisite is the ability to manage the entire infrastructure. When you manage services, you do not get the opportunity to choose which pieces of infrastructure you manage -- you need to manage them all. HP OpenView provides the industry's most comprehensive multivendor infrastructure management offerings across networks, systems, applications, and storage. Infrastructure is managed not only from an availability, or fault management perspective, it is also managed from a performance management perspective. Some vendors may manage either faults or performance across the range of infrastructure. Some vendors may manage both performance and faults of the given kind of infrastructure, but only HP OpenView has the ability to measure the entire infrastructure from a fault and performance perspective, bringing it altogether for your operations and planning organizations.
But success ultimately requires more than a collection of software tools ... it's requires an understanding of how these tools work together and in conjunction ... and the ability to deploy and optimize these tools for an existing customer environment ... technology, people, and processes.
___________________________________________________________
OpenView has solutions that address these (explain each one) domains from the back end infrastructure (bottom of the slide), tool that actually touch customers (SIP and Service Desk). Each one of these tools can be used on it’s own, except the SPI’s, to address a particular pain-point (explain building block approach).
At the heart of it all is the operations console, which acts as the nerve-center for the entire solution….consolidating and correlating events, performance data, etc. from all of the toolsets and then making business sense of it all with service views. Appropriate decision can be made, root cause can be identified and business impact assessed…keeping IT more inline with Service Level Objective and Agreements.
--------------------------------------
Animated:
So, we’ve looked at some of the depth that OpenView provides in areas such as Customer Experience/Internet Mgmt…and the ISS speaker went over Server Management with Insight Manager in detail…
OpenView actually has solutions that address these (explain each one) domains from the back end infrastructure (bottom of the slide), tool that actually touch customers (SIP and Service Desk). Each one of these tools can be used on it’s own, except the SPI’s, to address a particular pain-point (explain building block approach).
At the heart of it all is the operations console, which acts as the nerve-center for the entire solution….consolidating and correlating events, performance data, etc. from all of the toolsets and then making business sense of it all with service views. Appropriate decision can be made, root cause can be identified and business impact assessed…keeping IT more inline with Service Level Objective and Agreements.
Click
Here are the actual products, and all of these solutions shown run on a Windows console.
Presentation Title
Purpose of Slide
To introduce the overall ILM concept – and show that RISS is a part of this – BUT ONLY a PART of ILM – RISS is not the ONLY ILM offering
Speaker Notes
To quickly explain hp’s Information Lifecycle Management framework, and set the context of the business problems this approach addresses:
Today, many storage vendors offer a portfolio of products that manage the creation & modification, replication & distribution, protection & recovery, archival & recall and, removal of business critical information. These solutions exist within a particular stage of the information lifecycle. Solutions that are offered within a particular stage include:array-based software for replication, tape and disk-based backup and restore solutions, and archival to tape, WORM or other media.
An adaptive enterprise requires that information be managed throughout it’s lifecycle. This means utilizing policy-based storage services to automate the management of information based on that data’s business value. And, it’s important not only to manage the information, where and when its needed, but also at the appropriate cost. We know that every piece of data has some attribute assigned to it, IO for example, and those related to performance, availability, security, geographic location… But, the ability to truly take advantage of this important information is the next step to run business efficiently, with business agility.
Presentation Title
Presentation Title
Presentation Title
In today’s fiercely competitive marketplace, customers and partners want to conduct business with a company that will be able to support their needs for the long term. They want a company with the financial strength and stability that is to be expected of a strategic partner.
With $80 billion in revenue for fiscal 2004, $7 billion in organic growth (that’s equivalent to growing a company the size of EMC in one year), $13 billion in gross cash at the end of fiscal 2004 – in addition to generating more than $5 billion in cash from operations annually, HP is one of the strongest IT companies in the world.
Presentation Title
Presentation Title
Presentation Title
Presentation Title
In achieving the three FY04 objectives, we will also be accelerating our ability to carry out our business strategy: delivering high tech at a low cost with the best customer experience
Presentation Title
To deliver the greatest value to our customers, HP has structured its products and services offerings around the following categories: Personal Systems, Imaging & Printing, Storage & Servers, Services, and Software.
Personal Systems refers to our award-winning personal computing products for business and consumer users, such as notebooks, Tablet PCs, handhelds, desktops and workstations.
Imaging and Printing represents a category of HP’s portfolio that is unrivaled across the industry, with over 40 percent of worldwide market share. Our offerings within imaging and printing include: photo printers, color laser printers, mono printers, multi-function printers, supplies and media.
Presentation Title
This chart reflects the breadth of HP’s portfolio that no other company can match.
Presentation Title
Presentation Title
