IBM zEnterprise System

basesprocketΔιαχείριση Δεδομένων

31 Οκτ 2013 (πριν από 4 χρόνια και 12 μέρες)

206 εμφανίσεις


© 2010 IBM Corporation

IBM zEnterprise System


June, 2012


Jeffrey Frey

IBM Fellow, System z Architecture and Design

CTO, System z Platform




© 2010 IBM Corporation

2

zEnterprise:

Integrated IT Infrastructure for Optimized Service Delivery

zEnterprise Unified Resource Manager


Integrated, dynamic, intelligent
resource management


Workload
-
Aware Resource
Optimization


Enabled for Infrastructure as a
Service



Optimized to host large
scale database,
transaction, and mission
critical applications


The Most efficient
platform for Large
-
scale
Linux consolidation


Capable of massive

scale up

zEnterprise BladeCenter
Extension (zBX)

IBM zEnterprise


196 (z196)


Selected IBM POWER7


blades and IBM System x
®

Blades* for AIX
®

, Linux,
and Windows applications


High performance
optimizers and appliances
to accelerate time to
insight and reduce cost


Integrated high
performance private
network


© 2011 IBM Corporation

The Netezza TwinFin™ Appliance

High
-
performance database

engine streaming joins,

aggregations, sorts, etc.

SQL Compiler

Query Plan

Optimize

Admin

Processor &

streaming DB logic

Slice of User Data

Swap and Mirror partitions

High speed data streaming

SMP Hosts

Snippet
Blades™

(S
-
Blades™)

Disk Enclosures

© 2011 IBM Corporation

The Netezza AMPP™ Architecture

Advanced
Analytics

Loader

ETL

BI

Applications

FPGA

Memory

CPU

FPGA

Memory

CPU

FPGA

Memory

CPU

Hosts

Host

Disk

Enclosures

S
-
Blades™

Network

Fabric

Netezza Appliance

© 2011 IBM Corporation

Now expandable to 960 cores and 1.28 petabytes

TF3

TF6

TF12

TF24

TF36

TF48

TF72

TF96

TF120

Cabinets

1/4

1/2

1

2

3

4

6

8

10

Processing
Units

24

48

96

192

288

384

576

768

960

Capacity
(TB)

8

16

32

64

96

128

192

256

320

Effective
Capacity
(TB)*

32

64

128

256

384

512

768

1024

1280

.......

1

10

Capacity


= User Data space

Effective Capacity = User Data Space with compression

*: 4X compression assumed

Predictable, Linear Scalability throughout entire family

Accelerator Platforms

Low Latency, High Capacity Update

© 2011 IBM Corporation

DB2 Analytics Accelerator


Accelerating decisions to the speed of business

Blending System z and Netezza
technologies to deliver unparalleled, mixed
workload performance for complex analytic
business needs.



Get more insight from your data


Fast, predictable response times
for

right
-
time


analysis


Accelerate analytic query
response times


Improve price/performance for
analytic workloads


Minimize the need to create data
marts for performance


Highly secure environment for
sensitive data analysis


Transparent to the application


Low Latency, High Capacity Update

© 2011 IBM Corporation

7

Performance & Savings

Accelerating decisions to the speed of business

Queries run faster



Save CPU resources



People time



Business opportunities

Actual customer results, October 2011

DB2 Analytics Accelerator: “we had this up and
running in days with queries that ran over 1000
times faster”


DB2 Analytics Accelerator: “we expect ROI in
less than 4 months”


Advance to 32 minute mark for DB2
Analytics Accelerator section of keynote

© 2011 IBM Corporation

IBM DB2 Analytics Accelerator (IDAA)

DB2 for z/OS ODS members

DB2 for z/OS OLTP

members

Tightly integrated solution optimized for workloads

Simplification

Single entry point to the true data

High fidelity data

Performance

Fit The Solution To The Workload


Access to all Data

Business Analytics / ODS

OLTP


Complex queries against a data warehouse


Read only


Star schema design often used to speed up queries


Different modes of operation



Transactions against
operational data


Reads and writes


Normalized schema design to
eliminate redundancy


Multiple user throughput
operation

© 2010 IBM Corporation

9

Cloud Computing

....service sourcing and service value


ORGANIZATION


CULTURE

GOVERNANCE

Flexible Service Delivery and Consumption


Delivers select set of
standardized business
process, application, software
platform, and/or infrastructure
services on a flexible price
per use basis


Service Providers and
Service Consumers



Access by subscription


Drives efficiency,
standardization and best
practices while retaining
greater customization
and control


Formal Service Lifecycle
approach


Cloud Services

Cloud Computing

Model

.… Customization, efficiency,
availability, resiliency,
security and privacy

.…
Standardization, capital
preservation, flexibility and
time to deploy

Subscribe to Service


Request a service


SLA Negotiation


Service Personalization


“Sign“ Contract

Offer Service


Register Services


Add to Service Catalog

Service Design & Creation


Service Composition


Resource Modeling Tools


Service Level Choices


Topologies, Best Practices,
Management Templates

Deploy Service


Request Driven Provisioning


Management Agents and Best Practices


Application / Service On Boarding


Self
-
service interface

Manage Operation of Service


Service Composition and
Context


Monitoring, Reporting and
Optimization


Operational Control


Change Management


Security Management


Event handling


Process Management


Metering and Billing

Terminate Service


Controlled Clean
-
up

Service Lifecycle

10

© 2010 IBM Corporation

11

Taxonomy

Firmware and Hypervisors

z/OS, zLinux, …

AIX

Middleware

APP

APP

APP

APP

APP

APP

x86 Linux

Virtual Servers, Storage, Network


Server, Storage, Network, Accelerators, Appliances

Infrastructure
Management

OS and
Middleware
Management

Application
Management

Service Management

Workload
-
Aware Optimization

Virtualization Management

Hardware Management

SaaS

PaaS

IaaS

Service Management
& Orchestration

Self Service Portal

Service Catalog

Subscriber Mgmt

Resource Provisioning

Placement Advice

Billing

Monitoring

SLA Management

Image Management

© 2010 IBM Corporation

12

zEnterprise System

Z CPU, Memory and IO

PR/SM

Z CPU, Memory and IO

SE

SE

Virtual Machine

zOS

Linux

Linux

Linux

Virtual Machine

zOS

Virtual Machine

zOS

Virtual Machine

zOS

z Blade Extension

ISS

DP

Cell

DWA

pHyp

AIX

AIX

AIX

pHyp

AIX

AIX

AIX

pHyp

AIX

AIX

AIX

xHyp

Linux

Linux

Linux

xHyp

Linux

Linux

Linux

xHyp

Linux

Linux

Linux

AMM

z Blade Extension

ISS

DP

Cell

DWA

pHyp

AIX

AIX

AIX

pHyp

AIX

AIX

AIX

pHyp

AIX

AIX

AIX

xHyp

Linux

Linux

Linux

xHyp

Linux

Linux

Linux

xHyp

Linux

Linux

Linux

AMM

z Blade Extension

pHyp

AIX

AIX

AIX

pHyp

AIX

AIX

AIX

pHyp

AIX

AIX

AIX

xHyp

Linux

Linux

Linux

xHyp

Linux

Linux

Linux

xHyp

Windows

Linux

Linux

x86

Power

AMM

Virtual Machine

zOS

Linux

Virtual Machine

zOS

Linux

AIX

Data Power

Workload Resource Groups: SLA Mgmt and Reporting (Deployment, Performance, Availability, Security, Energy)

HMC (with Unified Resource Management)

Private Data Network (IEDN)

zHyp

© 2010 IBM Corporation

13

What is a zEnterprise Ensemble?


A zEnterprise Ensemble is a collection of 1..8
zEnterprise Nodes managed as a single
virtualized pool of server resources


A zEnterprise Node is a single zCEC with 0..4
racks with up to 1..2 blade centers per rack


Ensemble defines the scope of a zManager
management domain


zEnterprise nodes are deployed within
“synchronous” distances on L2 Network


Integrated Advanced Virtualization
Management


Implements well
-
defined external interface to
Data Center Service Management functions


Virtual Resource Management and
Optimization

zHMC
-

Ensemble Manager

© 2010 IBM Corporation

14


BC

Chassis


TOR

Switches

zBX Frames

z Frame

zEnterprise Node

Customer Managed

Management Network

Customer managed


Data Networks

Intra
-
Ensemble


Data Network

Intra
-
Node

Management

Network


OSM & OSX


OSAs


OSD OSAs

Mgt Interfaces

© 2010 IBM Corporation

15

RDMA


Aimed at providing very low latency/minimal CPU overhead interconnect between virtual
servers within zEnterprise


Utilizes RoCE (RDMA over Converged Ethernet)


New protocol that allows existing TCP applications to transparently benefit from RDMA
for transferring data


Builds upon reliable write access to a remote host’s memory in a controlled manner

© 2010 IBM Corporation

16

Dynamic Provisioning

Compute

Memory


Capture a Virtual Machine Image



The bootable disk image is copied into
the image library.



Image meta
-
data describing the Virtual
Infrastructure is captured and included
as part of the image.

Virtual

Machine

IO / Network

Virtual

Machine

Image Library


Deploy a Virtual Machine Image



The Image meta
-
data is used to create
VM container, allocating the required
server, storage, and network resources



The bootable image is copied to the
selected virtual host system.



The virtual machine is started from the
bootable disk image.

meta
-
data

SW

OS

meta
-
data

SW

OS

meta
-
data

SW

OS

meta
-
data

SW

OS

SW

OS

SW

OS

SW

OS

meta
-
data

SW

OS

© 2010 IBM Corporation

17

Virtual Resource Optimizations

Virtualization

Compute

Memory

Virtual

Machine


Dynamically adjust virtual machines resource allocations.


Allowing unallocated resources to be used by a virtual machine.


Allowing resource allocation adjustments to be made between virtual machines


Virtual server mobility between host systems.


Allocate resources on the target host.


Move the virtual machine in
-
memory state to target host.


De
-
allocating resources on the source host.

Virtual

Machine

Virtualization

Compute

Memory

IO / Network

Virtual

Machine

Virtual

Machine

IO / Network

SW

OS

SW

OS

SW

OS

SW

OS

SW

OS

SW

OS

© 2010 IBM Corporation

18

Virtual Machine Placement Intelligence


What is the most appropriate virtual
host within the ensemble to run a given
virtual machine image?


When deploying


When moving


When restarting



Need to consider…


Constraints (Configuration, Licensing..)


Optimizations (Performance,
Availability, Energy, Isolation)



Placement Advisors help to provide
optimized placement for the virtual
machines within the Ensemble

Mobility

Optimized for ….



Availability



Performance



Energy

© 2010 IBM Corporation

19

zManager

Interfaces


Resource Discovery


Monitoring and Events


Resource Provisioning


Automation


Security


Deployment and Mobility


Placement Advice


Service Level Policy

zManager

Interface

zHMC
-

Ensemble Manager

© 2010 IBM Corporation

20

zManager Interface


API is a new implementation in the HMC


New API includes existing SNMP/CIM function plus new Unified Resource Manager
capabilities


HMC SNMP and CIM interfaces not being extended with Unified Resource Manager
function


Existing SNMP and CIM capabilities remain in place and supported



Architecture/design based on current industry design practices


Requests and responses structured as RESTful web services


Status and property change notifications delivered via JMS


HMC provides an embedded JMS broker configured to support API
-
specific use only


JSON Payloads


Modern scripting languages (e.g., Java, Perl, Python) have HTTP support libraries



TCP/IP Sockets/HTTP is underlying network transport with SSL for connection security



Fully documented and supported for customer and third
-
party use

© 2010 IBM Corporation

21

Linked Pages

© 2010 IBM Corporation

22

IBM System x Blade Support


HX5
with
Westmere

processor


Intel 8
-
core Processor


2 Processor sockets


2.13 GHz @ 105W


16 DIMM 1066
Mhz

memory


Different memory configurations up to 256GB


Two internal 50GB
SSDs


8Gb
Fibre

Channel


10
GbE

Ethernet



Hypervisor is KVM


Kernel
-
based Virtual Machine


Implemented as Linux device driver


Requires a hosting Linux system/distribution (Selected
RHEV
-
H)



Operating System certifications: Linux (RHEL and SLES)
and Windows



OS Support (64 bit only):


RHEL 5.5, SLES 11 (SP1),
Additional versions to follow


Windows Server 2008 (SP2) for Data Center


Windows Server 2008 R2 for Data Center



Provide KVM
virtio

drivers from
RedHat

for all supported
versions/editions

© 2010 IBM Corporation

23

DataPower Integration


Deep
-
content routing and data aggregation


SOA Stack Acceleration


Application
-
layer security and threat protection


Protocol and message bridging



Blade Hardware Management


Energy Monitoring and Management of DP Blades


DP Firmware Load and Update


HMC Console Integration


PPM Assisted Dynamic Load Balancing (Future)


IEDN Virtual Network Provisioning


Monitoring and Reporting


DP Failure Recovery and Restart

© 2010 IBM Corporation

IBM Confidential

IBM zEnterprise System

24



1 to 80 configurable for client
use

IFL, zIIP, zAAP, ICFs and optional
SAPs

Up to 3 TB RAIM memory

15 subcapacity settings

Cryptographic enhancements

Optional water cooling and/or HV
DC Power

Upgradeable from z10 EC and z9
EC

Improvement for traditional
z/OS workloads

Improvement in CPU
intensive workloads via
compiler enhancements

40%

60%

Total capacity
improvement

30%

The Value begins at the heart of z196…….

zEnterprise 196 (z196)

Machine Type: 2817

Models: M15, M32, M49, M66, M80


Technology and Performance


Enhanced Availability and Serviceability


Workload Optimized Simplified Management



Dynamic Energy optimized


© 2010 IBM Corporation

IBM zEnterprise System

25

One to four


42u racks


capacity for 112 blades

No MIPS/MSU rating

Configured for high
availability

Integrated 10Gbe data
network

Optional rear door heat
exchanger

Optimizers

IBM Smart Analytics Optimizer

IBM WebSphere DataPower
(XI50z for zEnterprise)

…… and the value extends to heterogeneous workloads ……


Integrated IBM Certified Components driven by System z
order


Standard parts


TOR switch, BladeCenter Chassis,
Power Distribution Units, Optional Acoustic Panels


System z support


Problem reporting, hardware and firmware updates


Expanding operating system support for z196


AIX on POWER, Linux, and Windows on System x


Simplified management


Improved time to implement new applications


Central point of management for heterogeneous
workloads


No change to applications

IBM zEnterprise BladeCenter Extension (zBX)

Machine Type: 2458


Model 002

… managed by the

zEnterprise Unified Resource Manager

Select IBM Blades

IBM BladeCenter PS701 Express

IBM BladeCenter HX5 7873

© 2011 IBM Corporation

26



zEnterprise 114 (z114)

Machine Type: 2818

2 Models: M05 & M10

zEnterprise technology designed for small and mid
-
sized businesses

The Value Begins At the Heart with the z114



Improvement for traditional
z/OS workloads
1

Improvement in CPU
intensive workloads via
compiler enhancements
2


18%

12%

Up to

Up to

Total capacity improvement
1

Up to an
ADDITIONAL

25%

26
-

3100

MIPS

0
-
2

IBM provided spare cores


256
GB

RAIM fault tolerant memory

From

Up to

130
available capacity settings

Scales

From

Up to

Fully Upgradeable from the IBM System z10
Business Class

(z10 BC) & IBM System z9
®

Business Class (z9 BC); and to the z196 M15

1
-
10

configurable cores for client
use includes CPs,
IFL, zIIP, zAAP,
and ICFs


From


New technology in a new package


Modular 2 drawer design for lower cost of entry


Granularity for right
-
sizing your system


Additional Scale for consolidation and growth


Improved data center efficiency


Same Qualities of Service as the z196


Hybrid enabled to drive workload expansion and
integration



Improved Platform Economics


New Software Curve


Lower Hardware Maintenance


Lower specialty engine and memory prices


Upgradeability for investment protection