SlideShare a Scribd company logo
1 of 28
A portfolio of hardware,A portfolio of hardware,
software, and services for ansoftware, and services for an
enterprise-grade Linux operatingenterprise-grade Linux operating
environmentenvironment
Marcel Mitran – DE, CTO IBM LinuxONE
mmitran@ca.ibm.com
Enterprise-grade Linux
• Freedom & Agility
• Standards Based
• Speed of Innovation
• Developer Productivity
• Community Collaboration
• Quality of Software
• Dynamic Resource Allocation
• Non-Disruptive Scalability
• Continuous Business Availability
• Operational Efficiency
• Trusted Security
• Data and Transaction Serving
The best of
ENTERPRISE COMPUTING
The best of
LINUX & OPEN
2
IBM LINUXONE SYSTEMS
IBM LINUXONE SOLUTIONS
CloudCloud DevOpsDevOps
IBM LinuxONE CommunityIBM LinuxONE Community
CloudCloud
Open Mainframe ProjectOpen Mainframe Project
Open Source ContributionOpen Source Contribution
MobileMobile AnalyticsAnalytics
IBM LinuxONE Elastic PricingIBM LinuxONE Elastic Pricing
IBM LinuxONE Announcements - August 2015
3
Ubuntu on LinuxONE Systems makes it easy to build, model, deploy and manage
Enterprise scale out clusters and scalable cloud architectures
Cloud
Fully supported, scalable cloud
services based on OpenStack
technologies
Ecosystem development
Community engagement with
emerging generation of agile
developers
Target Workloads
Optimized for leading open source,
data centric scale out workloads –
web commerce, database, NoSQL,
analytics, etc. Connect systems of
record to systems of engagement.
Juju based automation
Automated deployment, management
and scaling of next generation
workloads
Secure Scalable
Clouds
Easy
Application
Migration
Diverse
Ecosystem
#1 Linux for Cloud
and Scale-out
Why Ubuntu OS and Ubuntu
OpenStack on LinuxONE?
Providing Significant Time to Value Advantage
4
LinuxONE Elastic Pricing
On Premise Cloud Model
what you pay depends on what you use for your
LinuxONE Deployment
Move from CapEx to OpEX
as deployment models change LinuxONE has adapted
to the shift
Simple
Simplified monthly or quarterly billing based on your
usage
Low barrier to entry
Reduces risk of entry and cost of acquisition by
as much as 75%
Metered
Based on real usage data ...
5
6
Current Members
Open Mainframe Project
Create an open source, technical community that industry and community
participants may easily participate in and so that they may contribute to the creation
of assets and materials that will benefit the development of Enterprise Grade
characteristics of Linux such as:
– High Availability and Disaster Recovery
– Security
– Reliability, Availability and Serviceability
– Performance and Scalability
LinuxONE – Diagonal Scale
Extreme Virtualization and Scale
• Hypervisor partitioning built into firmware
• Complete isolation – EAL5+
• Supports as many as 85 hypervisor instances – z/VM or KVM
• 1k Linux guests/hypervisor
• +1 million docker containers
• Hypervisor communication is via fast, in-memory TCP/IP
• Hipersockets or Shared-OSA – 3x less latency than discrete servers
• Massive dedicated I/O – 640 power co-processors
• 960Meg L4 cache, 5Ghz core, dual-TLBs, crypto acceleration
Super Elastic System
• Combine horizontal and vertical scaling
• Non-disruptively add/remove resources from Linux guests
• Non-disruptively add/remove Linux guests
Compose high-performance scalable applications. Dynamically and seamlessly re-allocate
resources between guests. Provide right-time analytics and powerful engagement
HW
HiperSocket LAN / Shared OSA
Linux guest
DockerDocker DockerDocker DockerDocker DockerDocker
…
7
LPAR1 LPAR2
A single LinuxONE Emperor ran more than 1 Million
Docker containers
Workload: busybox httpd server (no NAT)
LinuxONE Emperor runs 2000+ Docker containers
on avg 2.0x better than a compared Haswell-based
system
Workload: Apache Solr
LinuxONE Emperor can host over 10k Docker
containers
Workload: 4k Apache Solr + 6k busybox httpd server (no NAT)
Extreme Virtualization with Docker®!
Docker is a registered trademarks of Docker, Inc. in the United States and/or other countries
Disclaimer:
This claim is based on results from internal lab measurements. Performance results may vary depending on the workload and other factors.Benchmark:
oApache Solr search queries driven by Apache Jmeter*
System Stack:
oLinuxONE Emperor (IBM z13): Native LPAR on 36 CPU cores with 755GB memory
oHaswell-based alternative system (Lenovo System x3650 M5 w/ E5-2699 v3 processors): Native Linux on 36 CPU cores with 755GB memory
oHeavy Docker Container: Apache Solr v4.10.0, WebSphere Liberty v8.5.5.2, IBM Java 1.8.0 SR1
oLightweight Docker Container: BusyBox
oSystem SW: Docker 1.10.0-dev**
w/ aufs storage backend, RHEL 7.1***
Note:
•Each active container is driven by a client thread in Apache Jmeter, which keeps sending the same Solr query repeatedly to the container to search documents that contain given key words in a pre-
loaded & pre-indexed 46GB Wikipedia snapshot.
**
The docker runtime was modified to increase a thread count limit, to avoid connection time-out, and to separate a dockerinit binary from a docker binary.
***
A modified Linux 4.3.0 kernel to support more than 1024 network bridge ports was installed on RHEL 7.1.
LinuxONE is designed for large, cache-intensive
business workloads
9
LinuxONE also has
960 MB
per drawer off-chip
L4 cache
Haswell chip
E5 2600 V3
(representa
tion)
Core0
Core1
Core2
Core3
Core4
Core5
Core6
Core7
Shared
L3
Cache
LinuxONE chip
(representa
tion)
Latency introduced by
(1) hop across multiple
L3 rings, and (2) CPU-to-
L3 frequency differences
Latency is minimized –
all cores have equal
access to L3, no
frequency boundaries
http://www.anandtech.com/show/6355/intels-haswell-architecture/10
http://www.anandtech.com/show/8423/intel-xeon-e5-version-3-up-to-18-haswell-ep-cores-/4
Core
Shared
L3 cache
Shared
L3 cache
Core
Core
Core
Core
Core Core Core
Core
Core
Core
Core
Core
Core
Core
Core
Core
Core
LinuxONE is designed for high I/O
bandwidth business workloads
10
Up to 141 cores for
business logic
Up to 320 I/O
channel processors –
each with
2 POWER cores
(160 PCIe slots)
Up to 24 cores
dedicated to I/O
processing
LinuxON
E
HP BL460c Gen9
24 cores for both
business and I/O
processing
ZERO I/O
cores
4 I/O channel
processors
(2 PCIe slots)
 10-40% of typical business applications
processing is estimated to be I/O processing*
 Only LinuxONE has specialty cores dedicated
to I/O processing
 LinuxONE has about 80x more I/O channel
processors than typical scale-out servers
3 Oracle RAC clusters
4 server nodes per cluster
12 total HP DL380 servers E5-2699v3
2.3GHz 2ch/36co
(432 cores)
Oracle DB
workload
Which platform
provides the lowest
TCA over 3 years?
$13.5M (3 yr. TCA)
$29.3M (3 yr. TCA)
TCA includes hardware, software, maintenance, support and subscription.
Workload Equivalence derived from a proof-of-concept study conducted at a large Cooperative Bank.
3 Oracle RAC clusters
4 nodes per cluster
Each node is a Linux guest
LinuxONE with 61 cores
Customer Database Workloads
each supporting 18.3K tps
Oracle Enterprise Edition
Oracle Real Application Cluster
Database workloads with high I/O bandwidth
requirements benefit from LinuxONE architecture
11
Lower cost54%
2x Better Data-Serving with LinuxONE
MariaDB 10.1.5
1.8x to 2.1x
throughput
improveme
nt on
Sysbench
Benchmark
PostgreSQL 9.4
1.6x to 2.2x
throughput
improveme
nt on
pgBench
Benchmark
1.9x to 2.1x
throughput
improvement
on YCSB
Benchmark
Cloudant DBaaS
MongoDB 3.0.4 (WiredTiger, no
sharding)
1.9x to
2.1x
throughpu
t
improvem
ent on
YCSB
Benchmar
k
12
Scaling-up with MongoDB on LinuxONE
Single MongoDB node on LinuxONE scales up to 2TBs with sustained
throughput and response time <5ms, while supporting
+4 Billion documents, 460,000
reads+writes/second, with no sharding required!
“We are committed to make MongoDB available on all major
platforms and are excited to add support for IBM LinuxONE
Enterprise Grade Linux and LinuxONE Platform. This
announcement is a leap forward for customers who want to
deploy modern, mission-critical applications built with
MongoDB and take advantage of the performance, scalability
and security of IBM’s LinuxONE platform hardware products.”
--- Eliot Horowitz CTO & Founder, MongoDB
13
Scaling-out with Node.js on LinuxONE
Node.js v0.12
http://www.ibm.com/developerworks/web/nodesdk
High Performance JavaScript for LinuxONE
•Highly scalable, event-driven platform with non-blocking I/O
•Thousands of concurrent connections with minimal overhead
•Up to 2.1x more RESTful web interactions with AcmeAir in node.js
“LinuxONE can scale to up to 30B RESTful web interactions/day with Dockerized
Node.js and MongoDB, driving over 470K database read and writes per second,
while maintaining response-times that are 2x better than alternative
platforms.”
14
SOR+SOI+SOE-in-a-Box with LinuxONE
Co-located SOR, SOI, and SOE for right-time insights and powerfull engagement
TPC-E Database Aggregation Query
14
MillionRows/Minute
# Spark Partitions
12
10
8
6
4
2
0
1 2 4
3.1x
2.3x
1.9x
LinuxONE x86
Co-locating Node.js on LinuxONE vs. x86 results in
60% Faster Response Time 2.5x better
Throughput
Apache Spark co-located on LinuxONE drove
up to 3x faster than Spark running
off- platform on x86
WebSphere
App Server
Business Logic
Persistent Layer
DB2
Twitter
SOE
SOR
15
256 variable workloads
Average server utilization = 77%
Utilization
Time
Statistical models show how consolidating workloads drives
up CPU utilization
16
6x Peak-to-average
Average server utilization = 17%
1 variable workload
Utilization
Time
1.3x Peak-to-average
 Service-level-guarantees require provisioning for peak utilization
 But costs are inversely proportional to average utilization
 Consolidation leads to higher average utilization levels and lower
cost/workload
Performance comparison based on IBM Internal tests comparing IBM LinuxONE private cloud with one comparably configured private x86 cloud running an aggregation of light, medium and heavy workloads designed to
replicate typical IBM customer workload usage in the marketplace.  System configurations are based on equivalence ratios derived from IBM internal studies. Price comparison estimates based on a 3yr Total Cost of
Ownership (TCO) using publicly available U.S. prices (including a 20% discount for middleware) current as of January 1, 2015. LinuxONE and x86 TCO estimates include costs of infrastructure (system, memory, storage,
virtualization, OS, cloud management), middleware, power, floor space and labor. Results may vary based on actual workloads, system configurations, customer applications, queries and other variables in a production
environment and may produce different results. Users of this document should verify the applicable data for their specific environment. 
Variable workloads consolidated on LinuxONE
have lower TCO
17
32 cores
$7.2M (3yr TCO)
264 x86 cores
$10.6M (3yr TCO)
32% Lower TCO*
*estimated
123 workloads
(219 VMs)
32 medium workloads –
77 tps; 4GB RAM & 20GB
storage per VM
(64 VMs)
27 heavy I/O workloads –
255 tps; 122GB RAM &
540GB storage per VM
(27 VMs)
64 light workloads –
14 tps; 2GB RAM &
20GB storage per VM
(128 VMs)
LinuxONE with 32 cores, 3,806
GB memory, and Storwize v7000
with 47x400 GB SSDs.
11 x86 servers, each with 24 Intel
E7-8857 v2 3.0 GHz cores, 512 GB
memory, and 7x400 GB SSDs
I/O intensive
database
Oracle
Web
Application
WAS DB2
Web
Application
WAS DB2
A breakdown of TCO shows how
software costs soar on the x86 platform
18
For LinuxONE, 78%
fewer cores results in
72% lower
middleware costs
In x86 environments, high priority workloads must
have dedicated servers…
19
Consolidation ratios derived from IBM internal studies. LinuxONE numbers derived from measurements
on similar server. Workloads running WebSphere 8.5 ND, DB210 AESE, and Monitoring software.
Results may vary based on customer workload profiles/characteristics. Prices will vary by country.
Virtualized on 3
Intel 40-core servers
(120 cores total - Linux)
32 cores on LinuxONE
$13.7M (3 yr. TCA)
$5.77M (3 yr. TCA)
High priority
workloads
Low priority
workloads
High priority online banking workloads
driving a total of 9.1M transactions per hour
and low priority discretionary workloads
driving 2.8M transactions per hour
Linux
Hypervisor
Linux
Hypervisor
High
priority
workloads
High
priority
workloads
High
priority
workloads
Low
priority
workloads
Low
priority
workloads
Lower cost58%
…to avoid the effects of ‘noisy
neighbors’ – causes increased
proliferation of x86 servers
LinuxONE is designed to ensure highest
availability and lowest downtime
20
Comprehensive,
multi-layered strategy
includes…
Error Prevention
Hardware and firmware designed to
protect against outages
Built-in redundancy eliminates single
points of failure
Extensive testing and failure analysis
at every level
Error Detection
and Correction
Error detection embedded in components
Built-in automated diagnostics; problem
determination and isolation
Non-disruptive installation, upgrades and
maintenance avoids
outages
Error Recovery
Automated failover to speed
recovery and minimize system impact
Business continuity and disaster recovery
solutions – GDPS, HyperDispatch, Call
Home, etc.
LinuxONE is designed for minimal
unplanned downtime
21 Source: ITIC 2015 - 2016 Global Server Hardware, Server OS Reliability
Survey
Source: IDC 2015
Unplanned Downtime of >four (4) hours on each server hardware platform (2015)
ITIC survey shows IBM
LinuxONE had no – 0% –
unplanned system downtime
due to inherent flaws in
hardware
89%
Unplanned
downtime instances
per year
45% 94%
Time to resolve
unplanned
downtime
Productivity hours
lost per year per
user
Recent IDC study concludes clients who
leverage LinuxONE can virtually eliminate lost
productivity caused by downtime
Core0
Core1
Core2
Core3
Core4
Core5
Core6
Core7
Shared
L3
Cache
Core0
Core1
Core2
Core3
Core4
Core5
Core6
Core7
Shared
L3
Cache
If a core fails, a spare can be “turned on”
without system or program interruption
22
• Each LinuxONE server has two cores designated as spare
• Core failover (called sparing) is transparent to applications
• Spares need not be local to the same chip, node or drawer
• Any core can failover to a spare
x86 servers
do not have core
sparing
Achieve 100% continuous availability
23
• Applications restarted in another site
when disaster event occurs
• Single point of control and automation,
reliable and rapid recovery
• Keeps data available and consistent
• Protects against disk subsystem failures
Geographically Dispersed Parallel Sysplex –
A fully integrated software solution providing continuous availability / disaster
recovery for LinuxONE virtual machines
Recover from outages
in 6 seconds instead of 2+ hours
GDPS Virtual Appliance requires LinuxONE/VM and the HyperSwap functionality.
HyperSwap is not currently available for bare partitions or for KVM on LinuxONE
Advanced cryptography is handled at multiple levels
depending on business requirements
24
• Each core has its own cryptographic co-
processor
– Optimized for encryption functions
• Crypto Express5S PCIe card (optional)
adds additional crypto capability
– Elliptic Curve, SHA3, Visa FPE, etc.
• Meets FIPS, ANSI, PKI, and DK standards
On-chip cryptography
Optional Crypto Express5S for additional advanced
cryptography features
Top to bottom security is built in, not bolted on
25
Enables application and database security without
modifying applications
Reduces security complexity and expense:
• Central security process that is
easy to apply to new workloads or
as user base increases
• Tracks activity to address
audit and compliance
requirementsHardware
Architecture
LinuxONE/VM
Networks
Data & Applications
Administration
RACF
Resource Access Control Facility (RACF) is the
backbone of LinuxONE security
Access control to all classes of resources
Integrated into LinuxONE/VM
Supports cryptographic services
Supports digital certificates
Average security incursion costs on LinuxONE are
much lower than other platforms
26
Customer surveys show LinuxONE average
incursion costs are a fraction
of distributed platforms
Source: “Tracked, Hacked and Attacked?” © 2013, Solitaire Interglobal Ltd.
Linux
ONE
Comprehensive Security:
Only LinuxONE natively covers multiple
dimensions of security; other platforms
require augmentation and additional
expense (Solitaire)
…You’re running very large I/O- or cache-intensive workloads like
database applications or transaction processing
…Your insights and engagement are time-sensitive
…You’re looking to consolidate large numbers of servers
…Your software license costs are unsustainable
…Even the smallest system failure is unacceptable
…A security breach would be catastrophic to your business
…You’re looking to reduce overall operational expenses
IBM LinuxONE is a better choice if…
Thank You
Marcel Mitran – CTO IBM LinuxONE
mmitran@ca.ibm.com

More Related Content

What's hot

Intel open stack-summit-session-nov13-final
Intel open stack-summit-session-nov13-finalIntel open stack-summit-session-nov13-final
Intel open stack-summit-session-nov13-final
Deepak Mane
 
Building a PaaS Platform like Bluemix on OpenStack
Building a PaaS Platform like Bluemix on OpenStackBuilding a PaaS Platform like Bluemix on OpenStack
Building a PaaS Platform like Bluemix on OpenStack
Animesh Singh
 

What's hot (19)

As a Service: Cloud Foundry on OpenStack - Lessons Learnt
As a Service: Cloud Foundry on OpenStack - Lessons LearntAs a Service: Cloud Foundry on OpenStack - Lessons Learnt
As a Service: Cloud Foundry on OpenStack - Lessons Learnt
 
Mesosphere and the Enterprise: Run Your Applications on Apache Mesos - Steve ...
Mesosphere and the Enterprise: Run Your Applications on Apache Mesos - Steve ...Mesosphere and the Enterprise: Run Your Applications on Apache Mesos - Steve ...
Mesosphere and the Enterprise: Run Your Applications on Apache Mesos - Steve ...
 
Delivering Infrastructure-as-a-Service with Open Source Software
Delivering Infrastructure-as-a-Service with Open Source SoftwareDelivering Infrastructure-as-a-Service with Open Source Software
Delivering Infrastructure-as-a-Service with Open Source Software
 
Red Hat Storage Day Boston - Why Software-defined Storage Matters
Red Hat Storage Day Boston - Why Software-defined Storage MattersRed Hat Storage Day Boston - Why Software-defined Storage Matters
Red Hat Storage Day Boston - Why Software-defined Storage Matters
 
Cont0519
Cont0519Cont0519
Cont0519
 
Future of Power: PowerLinux - Jan Kristian Nielsen
Future of Power: PowerLinux - Jan Kristian NielsenFuture of Power: PowerLinux - Jan Kristian Nielsen
Future of Power: PowerLinux - Jan Kristian Nielsen
 
DUG'20: 10 - Storage Orchestration for Composable Storage Architectures
DUG'20: 10 - Storage Orchestration for Composable Storage ArchitecturesDUG'20: 10 - Storage Orchestration for Composable Storage Architectures
DUG'20: 10 - Storage Orchestration for Composable Storage Architectures
 
Integrating OpenStack to Existing infrastructure
Integrating OpenStack to Existing infrastructureIntegrating OpenStack to Existing infrastructure
Integrating OpenStack to Existing infrastructure
 
Cloud Foundry and OpenStack - A Marriage Made in Heaven! (Cloud Foundry Summi...
Cloud Foundry and OpenStack - A Marriage Made in Heaven! (Cloud Foundry Summi...Cloud Foundry and OpenStack - A Marriage Made in Heaven! (Cloud Foundry Summi...
Cloud Foundry and OpenStack - A Marriage Made in Heaven! (Cloud Foundry Summi...
 
Intel open stack-summit-session-nov13-final
Intel open stack-summit-session-nov13-finalIntel open stack-summit-session-nov13-final
Intel open stack-summit-session-nov13-final
 
Considering Bare Metal
Considering Bare MetalConsidering Bare Metal
Considering Bare Metal
 
{code} and containers
{code} and containers{code} and containers
{code} and containers
 
Virtualization technolegys for amdocs
Virtualization technolegys for amdocsVirtualization technolegys for amdocs
Virtualization technolegys for amdocs
 
Deep Dive on REX-Ray, libStorage and the Container Storage Interface - Clinto...
Deep Dive on REX-Ray, libStorage and the Container Storage Interface - Clinto...Deep Dive on REX-Ray, libStorage and the Container Storage Interface - Clinto...
Deep Dive on REX-Ray, libStorage and the Container Storage Interface - Clinto...
 
Leading an Open Source community at a large Enterprise - Jonas Rosland - Open...
Leading an Open Source community at a large Enterprise - Jonas Rosland - Open...Leading an Open Source community at a large Enterprise - Jonas Rosland - Open...
Leading an Open Source community at a large Enterprise - Jonas Rosland - Open...
 
Cloud foundry Docker Openstack - Leading Open Source Triumvirate
Cloud foundry Docker Openstack - Leading Open Source TriumvirateCloud foundry Docker Openstack - Leading Open Source Triumvirate
Cloud foundry Docker Openstack - Leading Open Source Triumvirate
 
Building a PaaS Platform like Bluemix on OpenStack
Building a PaaS Platform like Bluemix on OpenStackBuilding a PaaS Platform like Bluemix on OpenStack
Building a PaaS Platform like Bluemix on OpenStack
 
Cloud foundry integration-with-openstack-and-docker-bangalorecf-meetup
Cloud foundry integration-with-openstack-and-docker-bangalorecf-meetupCloud foundry integration-with-openstack-and-docker-bangalorecf-meetup
Cloud foundry integration-with-openstack-and-docker-bangalorecf-meetup
 
Containers Docker Kind Kubernetes Istio
Containers Docker Kind Kubernetes IstioContainers Docker Kind Kubernetes Istio
Containers Docker Kind Kubernetes Istio
 

Similar to LinuxONE cavemen mmit 20160505 v1.0

Oracle day co_opensource_final
Oracle day co_opensource_finalOracle day co_opensource_final
Oracle day co_opensource_final
Manuel Contreras
 
Presentazione IBM Power System Evento Venaria 14 ottobre
Presentazione IBM Power System Evento Venaria 14 ottobrePresentazione IBM Power System Evento Venaria 14 ottobre
Presentazione IBM Power System Evento Venaria 14 ottobre
PRAGMA PROGETTI
 
Cloud Expo East 2013: Essential Open Source Software for Building the Open Cloud
Cloud Expo East 2013: Essential Open Source Software for Building the Open CloudCloud Expo East 2013: Essential Open Source Software for Building the Open Cloud
Cloud Expo East 2013: Essential Open Source Software for Building the Open Cloud
Mark Hinkle
 

Similar to LinuxONE cavemen mmit 20160505 v1.0 (20)

Linux one vs x86
Linux one vs x86 Linux one vs x86
Linux one vs x86
 
Linux one vs x86 18 july
Linux one vs x86 18 julyLinux one vs x86 18 july
Linux one vs x86 18 july
 
Introduction to Apache Mesos and DC/OS
Introduction to Apache Mesos and DC/OSIntroduction to Apache Mesos and DC/OS
Introduction to Apache Mesos and DC/OS
 
Oracle day co_opensource_final
Oracle day co_opensource_finalOracle day co_opensource_final
Oracle day co_opensource_final
 
The best private cloud in the world - hands down
The best private cloud in the world - hands downThe best private cloud in the world - hands down
The best private cloud in the world - hands down
 
AMIS Oracle OpenWorld 2015 Review – part 2- Hardware & IaaS and PaaS Cloud Fo...
AMIS Oracle OpenWorld 2015 Review – part 2- Hardware & IaaS and PaaS Cloud Fo...AMIS Oracle OpenWorld 2015 Review – part 2- Hardware & IaaS and PaaS Cloud Fo...
AMIS Oracle OpenWorld 2015 Review – part 2- Hardware & IaaS and PaaS Cloud Fo...
 
Marcelo Perazolo, Lead Software Architect, IBM Corporation - Monitoring a Pow...
Marcelo Perazolo, Lead Software Architect, IBM Corporation - Monitoring a Pow...Marcelo Perazolo, Lead Software Architect, IBM Corporation - Monitoring a Pow...
Marcelo Perazolo, Lead Software Architect, IBM Corporation - Monitoring a Pow...
 
Speed up Digital Transformation with Openstack Cloud & Software Defined Storage
Speed up Digital Transformation with Openstack Cloud & Software Defined StorageSpeed up Digital Transformation with Openstack Cloud & Software Defined Storage
Speed up Digital Transformation with Openstack Cloud & Software Defined Storage
 
State of the Container Ecosystem
State of the Container EcosystemState of the Container Ecosystem
State of the Container Ecosystem
 
Closing Keynote
Closing KeynoteClosing Keynote
Closing Keynote
 
OpenStack at the speed of business with SolidFire & Red Hat
OpenStack at the speed of business with SolidFire & Red Hat OpenStack at the speed of business with SolidFire & Red Hat
OpenStack at the speed of business with SolidFire & Red Hat
 
Techmeeting-17feb2016
Techmeeting-17feb2016Techmeeting-17feb2016
Techmeeting-17feb2016
 
Cloud Foundry and OpenStack - A Marriage Made in Heaven! (Cloud Foundry Summi...
Cloud Foundry and OpenStack - A Marriage Made in Heaven! (Cloud Foundry Summi...Cloud Foundry and OpenStack - A Marriage Made in Heaven! (Cloud Foundry Summi...
Cloud Foundry and OpenStack - A Marriage Made in Heaven! (Cloud Foundry Summi...
 
Webinar: OpenEBS - Still Free and now FASTEST Kubernetes storage
Webinar: OpenEBS - Still Free and now FASTEST Kubernetes storageWebinar: OpenEBS - Still Free and now FASTEST Kubernetes storage
Webinar: OpenEBS - Still Free and now FASTEST Kubernetes storage
 
Presentazione IBM Power System Evento Venaria 14 ottobre
Presentazione IBM Power System Evento Venaria 14 ottobrePresentazione IBM Power System Evento Venaria 14 ottobre
Presentazione IBM Power System Evento Venaria 14 ottobre
 
Oscon 2017: Build your own container-based system with the Moby project
Oscon 2017: Build your own container-based system with the Moby projectOscon 2017: Build your own container-based system with the Moby project
Oscon 2017: Build your own container-based system with the Moby project
 
Red Hat multi-cluster management & what's new in OpenShift
Red Hat multi-cluster management & what's new in OpenShiftRed Hat multi-cluster management & what's new in OpenShift
Red Hat multi-cluster management & what's new in OpenShift
 
Cloud Expo East 2013: Essential Open Source Software for Building the Open Cloud
Cloud Expo East 2013: Essential Open Source Software for Building the Open CloudCloud Expo East 2013: Essential Open Source Software for Building the Open Cloud
Cloud Expo East 2013: Essential Open Source Software for Building the Open Cloud
 
TechTalkThai-CiscoHyperFlex
TechTalkThai-CiscoHyperFlexTechTalkThai-CiscoHyperFlex
TechTalkThai-CiscoHyperFlex
 
Docker Orchestration: Welcome to the Jungle! Devoxx & Docker Meetup Tour Nov ...
Docker Orchestration: Welcome to the Jungle! Devoxx & Docker Meetup Tour Nov ...Docker Orchestration: Welcome to the Jungle! Devoxx & Docker Meetup Tour Nov ...
Docker Orchestration: Welcome to the Jungle! Devoxx & Docker Meetup Tour Nov ...
 

Recently uploaded

Future Visions: Predictions to Guide and Time Tech Innovation, Peter Udo Diehl
Future Visions: Predictions to Guide and Time Tech Innovation, Peter Udo DiehlFuture Visions: Predictions to Guide and Time Tech Innovation, Peter Udo Diehl
Future Visions: Predictions to Guide and Time Tech Innovation, Peter Udo Diehl
Peter Udo Diehl
 

Recently uploaded (20)

TEST BANK For, Information Technology Project Management 9th Edition Kathy Sc...
TEST BANK For, Information Technology Project Management 9th Edition Kathy Sc...TEST BANK For, Information Technology Project Management 9th Edition Kathy Sc...
TEST BANK For, Information Technology Project Management 9th Edition Kathy Sc...
 
TopCryptoSupers 12thReport OrionX May2024
TopCryptoSupers 12thReport OrionX May2024TopCryptoSupers 12thReport OrionX May2024
TopCryptoSupers 12thReport OrionX May2024
 
1111 ChatGPT Prompts PDF Free Download - Prompts for ChatGPT
1111 ChatGPT Prompts PDF Free Download - Prompts for ChatGPT1111 ChatGPT Prompts PDF Free Download - Prompts for ChatGPT
1111 ChatGPT Prompts PDF Free Download - Prompts for ChatGPT
 
Future Visions: Predictions to Guide and Time Tech Innovation, Peter Udo Diehl
Future Visions: Predictions to Guide and Time Tech Innovation, Peter Udo DiehlFuture Visions: Predictions to Guide and Time Tech Innovation, Peter Udo Diehl
Future Visions: Predictions to Guide and Time Tech Innovation, Peter Udo Diehl
 
Unpacking Value Delivery - Agile Oxford Meetup - May 2024.pptx
Unpacking Value Delivery - Agile Oxford Meetup - May 2024.pptxUnpacking Value Delivery - Agile Oxford Meetup - May 2024.pptx
Unpacking Value Delivery - Agile Oxford Meetup - May 2024.pptx
 
Overview of Hyperledger Foundation
Overview of Hyperledger FoundationOverview of Hyperledger Foundation
Overview of Hyperledger Foundation
 
A Business-Centric Approach to Design System Strategy
A Business-Centric Approach to Design System StrategyA Business-Centric Approach to Design System Strategy
A Business-Centric Approach to Design System Strategy
 
Portal Kombat : extension du réseau de propagande russe
Portal Kombat : extension du réseau de propagande russePortal Kombat : extension du réseau de propagande russe
Portal Kombat : extension du réseau de propagande russe
 
AI revolution and Salesforce, Jiří Karpíšek
AI revolution and Salesforce, Jiří KarpíšekAI revolution and Salesforce, Jiří Karpíšek
AI revolution and Salesforce, Jiří Karpíšek
 
IESVE for Early Stage Design and Planning
IESVE for Early Stage Design and PlanningIESVE for Early Stage Design and Planning
IESVE for Early Stage Design and Planning
 
Google I/O Extended 2024 Warsaw
Google I/O Extended 2024 WarsawGoogle I/O Extended 2024 Warsaw
Google I/O Extended 2024 Warsaw
 
How we scaled to 80K users by doing nothing!.pdf
How we scaled to 80K users by doing nothing!.pdfHow we scaled to 80K users by doing nothing!.pdf
How we scaled to 80K users by doing nothing!.pdf
 
Oauth 2.0 Introduction and Flows with MuleSoft
Oauth 2.0 Introduction and Flows with MuleSoftOauth 2.0 Introduction and Flows with MuleSoft
Oauth 2.0 Introduction and Flows with MuleSoft
 
Optimizing NoSQL Performance Through Observability
Optimizing NoSQL Performance Through ObservabilityOptimizing NoSQL Performance Through Observability
Optimizing NoSQL Performance Through Observability
 
WSO2CONMay2024OpenSourceConferenceDebrief.pptx
WSO2CONMay2024OpenSourceConferenceDebrief.pptxWSO2CONMay2024OpenSourceConferenceDebrief.pptx
WSO2CONMay2024OpenSourceConferenceDebrief.pptx
 
PLAI - Acceleration Program for Generative A.I. Startups
PLAI - Acceleration Program for Generative A.I. StartupsPLAI - Acceleration Program for Generative A.I. Startups
PLAI - Acceleration Program for Generative A.I. Startups
 
Speed Wins: From Kafka to APIs in Minutes
Speed Wins: From Kafka to APIs in MinutesSpeed Wins: From Kafka to APIs in Minutes
Speed Wins: From Kafka to APIs in Minutes
 
Measures in SQL (a talk at SF Distributed Systems meetup, 2024-05-22)
Measures in SQL (a talk at SF Distributed Systems meetup, 2024-05-22)Measures in SQL (a talk at SF Distributed Systems meetup, 2024-05-22)
Measures in SQL (a talk at SF Distributed Systems meetup, 2024-05-22)
 
What's New in Teams Calling, Meetings and Devices April 2024
What's New in Teams Calling, Meetings and Devices April 2024What's New in Teams Calling, Meetings and Devices April 2024
What's New in Teams Calling, Meetings and Devices April 2024
 
ASRock Industrial FDO Solutions in Action for Industrial Edge AI _ Kenny at A...
ASRock Industrial FDO Solutions in Action for Industrial Edge AI _ Kenny at A...ASRock Industrial FDO Solutions in Action for Industrial Edge AI _ Kenny at A...
ASRock Industrial FDO Solutions in Action for Industrial Edge AI _ Kenny at A...
 

LinuxONE cavemen mmit 20160505 v1.0

  • 1. A portfolio of hardware,A portfolio of hardware, software, and services for ansoftware, and services for an enterprise-grade Linux operatingenterprise-grade Linux operating environmentenvironment Marcel Mitran – DE, CTO IBM LinuxONE mmitran@ca.ibm.com
  • 2. Enterprise-grade Linux • Freedom & Agility • Standards Based • Speed of Innovation • Developer Productivity • Community Collaboration • Quality of Software • Dynamic Resource Allocation • Non-Disruptive Scalability • Continuous Business Availability • Operational Efficiency • Trusted Security • Data and Transaction Serving The best of ENTERPRISE COMPUTING The best of LINUX & OPEN 2
  • 3. IBM LINUXONE SYSTEMS IBM LINUXONE SOLUTIONS CloudCloud DevOpsDevOps IBM LinuxONE CommunityIBM LinuxONE Community CloudCloud Open Mainframe ProjectOpen Mainframe Project Open Source ContributionOpen Source Contribution MobileMobile AnalyticsAnalytics IBM LinuxONE Elastic PricingIBM LinuxONE Elastic Pricing IBM LinuxONE Announcements - August 2015 3
  • 4. Ubuntu on LinuxONE Systems makes it easy to build, model, deploy and manage Enterprise scale out clusters and scalable cloud architectures Cloud Fully supported, scalable cloud services based on OpenStack technologies Ecosystem development Community engagement with emerging generation of agile developers Target Workloads Optimized for leading open source, data centric scale out workloads – web commerce, database, NoSQL, analytics, etc. Connect systems of record to systems of engagement. Juju based automation Automated deployment, management and scaling of next generation workloads Secure Scalable Clouds Easy Application Migration Diverse Ecosystem #1 Linux for Cloud and Scale-out Why Ubuntu OS and Ubuntu OpenStack on LinuxONE? Providing Significant Time to Value Advantage 4
  • 5. LinuxONE Elastic Pricing On Premise Cloud Model what you pay depends on what you use for your LinuxONE Deployment Move from CapEx to OpEX as deployment models change LinuxONE has adapted to the shift Simple Simplified monthly or quarterly billing based on your usage Low barrier to entry Reduces risk of entry and cost of acquisition by as much as 75% Metered Based on real usage data ... 5
  • 6. 6 Current Members Open Mainframe Project Create an open source, technical community that industry and community participants may easily participate in and so that they may contribute to the creation of assets and materials that will benefit the development of Enterprise Grade characteristics of Linux such as: – High Availability and Disaster Recovery – Security – Reliability, Availability and Serviceability – Performance and Scalability
  • 7. LinuxONE – Diagonal Scale Extreme Virtualization and Scale • Hypervisor partitioning built into firmware • Complete isolation – EAL5+ • Supports as many as 85 hypervisor instances – z/VM or KVM • 1k Linux guests/hypervisor • +1 million docker containers • Hypervisor communication is via fast, in-memory TCP/IP • Hipersockets or Shared-OSA – 3x less latency than discrete servers • Massive dedicated I/O – 640 power co-processors • 960Meg L4 cache, 5Ghz core, dual-TLBs, crypto acceleration Super Elastic System • Combine horizontal and vertical scaling • Non-disruptively add/remove resources from Linux guests • Non-disruptively add/remove Linux guests Compose high-performance scalable applications. Dynamically and seamlessly re-allocate resources between guests. Provide right-time analytics and powerful engagement HW HiperSocket LAN / Shared OSA Linux guest DockerDocker DockerDocker DockerDocker DockerDocker … 7 LPAR1 LPAR2
  • 8. A single LinuxONE Emperor ran more than 1 Million Docker containers Workload: busybox httpd server (no NAT) LinuxONE Emperor runs 2000+ Docker containers on avg 2.0x better than a compared Haswell-based system Workload: Apache Solr LinuxONE Emperor can host over 10k Docker containers Workload: 4k Apache Solr + 6k busybox httpd server (no NAT) Extreme Virtualization with Docker®! Docker is a registered trademarks of Docker, Inc. in the United States and/or other countries Disclaimer: This claim is based on results from internal lab measurements. Performance results may vary depending on the workload and other factors.Benchmark: oApache Solr search queries driven by Apache Jmeter* System Stack: oLinuxONE Emperor (IBM z13): Native LPAR on 36 CPU cores with 755GB memory oHaswell-based alternative system (Lenovo System x3650 M5 w/ E5-2699 v3 processors): Native Linux on 36 CPU cores with 755GB memory oHeavy Docker Container: Apache Solr v4.10.0, WebSphere Liberty v8.5.5.2, IBM Java 1.8.0 SR1 oLightweight Docker Container: BusyBox oSystem SW: Docker 1.10.0-dev** w/ aufs storage backend, RHEL 7.1*** Note: •Each active container is driven by a client thread in Apache Jmeter, which keeps sending the same Solr query repeatedly to the container to search documents that contain given key words in a pre- loaded & pre-indexed 46GB Wikipedia snapshot. ** The docker runtime was modified to increase a thread count limit, to avoid connection time-out, and to separate a dockerinit binary from a docker binary. *** A modified Linux 4.3.0 kernel to support more than 1024 network bridge ports was installed on RHEL 7.1.
  • 9. LinuxONE is designed for large, cache-intensive business workloads 9 LinuxONE also has 960 MB per drawer off-chip L4 cache Haswell chip E5 2600 V3 (representa tion) Core0 Core1 Core2 Core3 Core4 Core5 Core6 Core7 Shared L3 Cache LinuxONE chip (representa tion) Latency introduced by (1) hop across multiple L3 rings, and (2) CPU-to- L3 frequency differences Latency is minimized – all cores have equal access to L3, no frequency boundaries http://www.anandtech.com/show/6355/intels-haswell-architecture/10 http://www.anandtech.com/show/8423/intel-xeon-e5-version-3-up-to-18-haswell-ep-cores-/4 Core Shared L3 cache Shared L3 cache Core Core Core Core Core Core Core Core Core Core Core Core Core Core Core Core Core
  • 10. LinuxONE is designed for high I/O bandwidth business workloads 10 Up to 141 cores for business logic Up to 320 I/O channel processors – each with 2 POWER cores (160 PCIe slots) Up to 24 cores dedicated to I/O processing LinuxON E HP BL460c Gen9 24 cores for both business and I/O processing ZERO I/O cores 4 I/O channel processors (2 PCIe slots)  10-40% of typical business applications processing is estimated to be I/O processing*  Only LinuxONE has specialty cores dedicated to I/O processing  LinuxONE has about 80x more I/O channel processors than typical scale-out servers
  • 11. 3 Oracle RAC clusters 4 server nodes per cluster 12 total HP DL380 servers E5-2699v3 2.3GHz 2ch/36co (432 cores) Oracle DB workload Which platform provides the lowest TCA over 3 years? $13.5M (3 yr. TCA) $29.3M (3 yr. TCA) TCA includes hardware, software, maintenance, support and subscription. Workload Equivalence derived from a proof-of-concept study conducted at a large Cooperative Bank. 3 Oracle RAC clusters 4 nodes per cluster Each node is a Linux guest LinuxONE with 61 cores Customer Database Workloads each supporting 18.3K tps Oracle Enterprise Edition Oracle Real Application Cluster Database workloads with high I/O bandwidth requirements benefit from LinuxONE architecture 11 Lower cost54%
  • 12. 2x Better Data-Serving with LinuxONE MariaDB 10.1.5 1.8x to 2.1x throughput improveme nt on Sysbench Benchmark PostgreSQL 9.4 1.6x to 2.2x throughput improveme nt on pgBench Benchmark 1.9x to 2.1x throughput improvement on YCSB Benchmark Cloudant DBaaS MongoDB 3.0.4 (WiredTiger, no sharding) 1.9x to 2.1x throughpu t improvem ent on YCSB Benchmar k 12
  • 13. Scaling-up with MongoDB on LinuxONE Single MongoDB node on LinuxONE scales up to 2TBs with sustained throughput and response time <5ms, while supporting +4 Billion documents, 460,000 reads+writes/second, with no sharding required! “We are committed to make MongoDB available on all major platforms and are excited to add support for IBM LinuxONE Enterprise Grade Linux and LinuxONE Platform. This announcement is a leap forward for customers who want to deploy modern, mission-critical applications built with MongoDB and take advantage of the performance, scalability and security of IBM’s LinuxONE platform hardware products.” --- Eliot Horowitz CTO & Founder, MongoDB 13
  • 14. Scaling-out with Node.js on LinuxONE Node.js v0.12 http://www.ibm.com/developerworks/web/nodesdk High Performance JavaScript for LinuxONE •Highly scalable, event-driven platform with non-blocking I/O •Thousands of concurrent connections with minimal overhead •Up to 2.1x more RESTful web interactions with AcmeAir in node.js “LinuxONE can scale to up to 30B RESTful web interactions/day with Dockerized Node.js and MongoDB, driving over 470K database read and writes per second, while maintaining response-times that are 2x better than alternative platforms.” 14
  • 15. SOR+SOI+SOE-in-a-Box with LinuxONE Co-located SOR, SOI, and SOE for right-time insights and powerfull engagement TPC-E Database Aggregation Query 14 MillionRows/Minute # Spark Partitions 12 10 8 6 4 2 0 1 2 4 3.1x 2.3x 1.9x LinuxONE x86 Co-locating Node.js on LinuxONE vs. x86 results in 60% Faster Response Time 2.5x better Throughput Apache Spark co-located on LinuxONE drove up to 3x faster than Spark running off- platform on x86 WebSphere App Server Business Logic Persistent Layer DB2 Twitter SOE SOR 15
  • 16. 256 variable workloads Average server utilization = 77% Utilization Time Statistical models show how consolidating workloads drives up CPU utilization 16 6x Peak-to-average Average server utilization = 17% 1 variable workload Utilization Time 1.3x Peak-to-average  Service-level-guarantees require provisioning for peak utilization  But costs are inversely proportional to average utilization  Consolidation leads to higher average utilization levels and lower cost/workload
  • 17. Performance comparison based on IBM Internal tests comparing IBM LinuxONE private cloud with one comparably configured private x86 cloud running an aggregation of light, medium and heavy workloads designed to replicate typical IBM customer workload usage in the marketplace.  System configurations are based on equivalence ratios derived from IBM internal studies. Price comparison estimates based on a 3yr Total Cost of Ownership (TCO) using publicly available U.S. prices (including a 20% discount for middleware) current as of January 1, 2015. LinuxONE and x86 TCO estimates include costs of infrastructure (system, memory, storage, virtualization, OS, cloud management), middleware, power, floor space and labor. Results may vary based on actual workloads, system configurations, customer applications, queries and other variables in a production environment and may produce different results. Users of this document should verify the applicable data for their specific environment.  Variable workloads consolidated on LinuxONE have lower TCO 17 32 cores $7.2M (3yr TCO) 264 x86 cores $10.6M (3yr TCO) 32% Lower TCO* *estimated 123 workloads (219 VMs) 32 medium workloads – 77 tps; 4GB RAM & 20GB storage per VM (64 VMs) 27 heavy I/O workloads – 255 tps; 122GB RAM & 540GB storage per VM (27 VMs) 64 light workloads – 14 tps; 2GB RAM & 20GB storage per VM (128 VMs) LinuxONE with 32 cores, 3,806 GB memory, and Storwize v7000 with 47x400 GB SSDs. 11 x86 servers, each with 24 Intel E7-8857 v2 3.0 GHz cores, 512 GB memory, and 7x400 GB SSDs I/O intensive database Oracle Web Application WAS DB2 Web Application WAS DB2
  • 18. A breakdown of TCO shows how software costs soar on the x86 platform 18 For LinuxONE, 78% fewer cores results in 72% lower middleware costs
  • 19. In x86 environments, high priority workloads must have dedicated servers… 19 Consolidation ratios derived from IBM internal studies. LinuxONE numbers derived from measurements on similar server. Workloads running WebSphere 8.5 ND, DB210 AESE, and Monitoring software. Results may vary based on customer workload profiles/characteristics. Prices will vary by country. Virtualized on 3 Intel 40-core servers (120 cores total - Linux) 32 cores on LinuxONE $13.7M (3 yr. TCA) $5.77M (3 yr. TCA) High priority workloads Low priority workloads High priority online banking workloads driving a total of 9.1M transactions per hour and low priority discretionary workloads driving 2.8M transactions per hour Linux Hypervisor Linux Hypervisor High priority workloads High priority workloads High priority workloads Low priority workloads Low priority workloads Lower cost58% …to avoid the effects of ‘noisy neighbors’ – causes increased proliferation of x86 servers
  • 20. LinuxONE is designed to ensure highest availability and lowest downtime 20 Comprehensive, multi-layered strategy includes… Error Prevention Hardware and firmware designed to protect against outages Built-in redundancy eliminates single points of failure Extensive testing and failure analysis at every level Error Detection and Correction Error detection embedded in components Built-in automated diagnostics; problem determination and isolation Non-disruptive installation, upgrades and maintenance avoids outages Error Recovery Automated failover to speed recovery and minimize system impact Business continuity and disaster recovery solutions – GDPS, HyperDispatch, Call Home, etc.
  • 21. LinuxONE is designed for minimal unplanned downtime 21 Source: ITIC 2015 - 2016 Global Server Hardware, Server OS Reliability Survey Source: IDC 2015 Unplanned Downtime of >four (4) hours on each server hardware platform (2015) ITIC survey shows IBM LinuxONE had no – 0% – unplanned system downtime due to inherent flaws in hardware 89% Unplanned downtime instances per year 45% 94% Time to resolve unplanned downtime Productivity hours lost per year per user Recent IDC study concludes clients who leverage LinuxONE can virtually eliminate lost productivity caused by downtime
  • 22. Core0 Core1 Core2 Core3 Core4 Core5 Core6 Core7 Shared L3 Cache Core0 Core1 Core2 Core3 Core4 Core5 Core6 Core7 Shared L3 Cache If a core fails, a spare can be “turned on” without system or program interruption 22 • Each LinuxONE server has two cores designated as spare • Core failover (called sparing) is transparent to applications • Spares need not be local to the same chip, node or drawer • Any core can failover to a spare x86 servers do not have core sparing
  • 23. Achieve 100% continuous availability 23 • Applications restarted in another site when disaster event occurs • Single point of control and automation, reliable and rapid recovery • Keeps data available and consistent • Protects against disk subsystem failures Geographically Dispersed Parallel Sysplex – A fully integrated software solution providing continuous availability / disaster recovery for LinuxONE virtual machines Recover from outages in 6 seconds instead of 2+ hours GDPS Virtual Appliance requires LinuxONE/VM and the HyperSwap functionality. HyperSwap is not currently available for bare partitions or for KVM on LinuxONE
  • 24. Advanced cryptography is handled at multiple levels depending on business requirements 24 • Each core has its own cryptographic co- processor – Optimized for encryption functions • Crypto Express5S PCIe card (optional) adds additional crypto capability – Elliptic Curve, SHA3, Visa FPE, etc. • Meets FIPS, ANSI, PKI, and DK standards On-chip cryptography Optional Crypto Express5S for additional advanced cryptography features
  • 25. Top to bottom security is built in, not bolted on 25 Enables application and database security without modifying applications Reduces security complexity and expense: • Central security process that is easy to apply to new workloads or as user base increases • Tracks activity to address audit and compliance requirementsHardware Architecture LinuxONE/VM Networks Data & Applications Administration RACF Resource Access Control Facility (RACF) is the backbone of LinuxONE security Access control to all classes of resources Integrated into LinuxONE/VM Supports cryptographic services Supports digital certificates
  • 26. Average security incursion costs on LinuxONE are much lower than other platforms 26 Customer surveys show LinuxONE average incursion costs are a fraction of distributed platforms Source: “Tracked, Hacked and Attacked?” © 2013, Solitaire Interglobal Ltd. Linux ONE Comprehensive Security: Only LinuxONE natively covers multiple dimensions of security; other platforms require augmentation and additional expense (Solitaire)
  • 27. …You’re running very large I/O- or cache-intensive workloads like database applications or transaction processing …Your insights and engagement are time-sensitive …You’re looking to consolidate large numbers of servers …Your software license costs are unsustainable …Even the smallest system failure is unacceptable …A security breach would be catastrophic to your business …You’re looking to reduce overall operational expenses IBM LinuxONE is a better choice if…
  • 28. Thank You Marcel Mitran – CTO IBM LinuxONE mmitran@ca.ibm.com

Editor's Notes

  1. When you think of Linux you think of open collaboration, fast paced innovation, agility and capability that empowers developers and operators. When you think of IBM Big Iron, you think of the qualities of service that have defined 5 decades of un-matched resilience, security, dynamic resource allocation, industry leading data serving and scalable transaction processing.
  2. On Aug 2015 we took Linux on z to the next level - we announced IBM LinuxONE. LinuxONE brings together a broad set of new capabilities such as a KVM hypervisor, Ubuntu -- cloud-enabled distribution, and an extreme expansion of the eco-system to include significant set of open source capabilities. LinuxONE introduces new elastic pricing which represents a new cloud-based pricing model – essentially moving from a capital expenditure to an operational expenditure pricing model. We introduced the new LinuxONE Community Cloud to enable open communities and open innovation on LinuxONE. We also worked with the Linux Foundation to start up the Open Mainframe Project which brings together academics, industry players and others to help drive open innovation and collaboration, bringing the best of the mainframe innovation to open communities, while also bringing mind-share for the mainframe in open communities. LinuxONE combines the power of open innovation with enterprise qualities of service to take Linux to the next level.
  3. Canonical delivers Ubuntu 16.04 support on LinuxONE and IBM z Systems the perfect platform for Hybrid Cloud Canonical’s Ubuntu provides more flexibility and choice with continued open ecosystem expansion Ubuntu allows developers who use Ubuntu on other systems to work in a familiar environment Allows CSPs and clients with applications built on Ubuntu to easily extend their environments to include LinuxONE and z Systems Lets customers get the performance, scale, reliability and security they need for Linux to run their business, using the tools they know, while getting the same economics they are use to
  4. Solr is a search engine, used my many web sites to search for terms in a document base Use case Greater business agility and lower costs by simplifying IT operations, enabling an organization to go to market quickly and solve business problems effectively. DevOps agility with Docker containers and microservices. Hybrid Cloud and Choice Require Portability: Cross Cloud Deployment – move the same application across multiple clouds Eliminate “lock-in”, become a “Cloud Broker” Key Messages Strong I/O subsystem on z13 enables a server to host a large number of containers that generate heavy file accesses. Superior I/O performance of z means more efficient paging translating to 1.5x better Docker density within a fixed core and memory footprint. Benchmark Apache Solr search queries driven by Apache Jmeter System Stack z13: Native LPAR on 16 CPU cores with 128G memory E5-2699 v3: Native Linux on 16 CPU cores with 128G memory SW: Apache Solr v4.10.0, WebSphere Liberty v8.5.5.2, IBM Java 1.7.0 SR1, Docker 1.6.2 w/ btrfs storage backend, RHEL 7.1
  5. NoSQL databases on LinuxOne performs 2x faster than other platforms. Compression hardware to save Spark Resilient Distributed Datasets runs 4.9x faster than other platforms and persists Docker containers 4x faster. LinuxOne can encrypt all the data 28x faster than other platforms. Node.js runs up to 2x faster. Spark Analytics runs up to 3x faster. Who wants to deal with sharded databases when you have a server that can handle TB databases, with over 2 billion documents and 470,000 read/writes per second. And remember all this speed and scalability is for just one LinuxOne server, you can cluster them to scale out as well to have even more scalability. This is a demo that combined unstructured data (tweets and news) with structured data, using real time analytics to provide insight. This demo was developed by 2 people in 4 weeks. They developed the code on their favorite Linux desktops and laptops and deployed it on LinuxOne for the performance and power. Think what you can do with a LinuxOne server! On average, AcmeAir transaction translates to 1.36x MongoDB operations (Insert, Query, Update, Delete). 460k r/w per sec was based on a fully saturated z13 Hartmut measured, with many instances of both AcmeAir and Mongo. The 27k for the 1TB system is simply based on 16 AcmeAir instances targetting 1 MongoDB instance (with lots of data). z13 seemless scale up to 350k transaction per sec leading to a 30B proof-point. Workload scales up ˜6x higher (thruput) than an alternative server. 460k number was a the outcome of the work done by Hartmut to load the machine to drive max throughput/box by having many instances of Node/Mongo running Encryption 28x faster – but for secure key, not for this demo
  6. Node.js scales nicely in Docker containers z Systems Connectivity Co-locate Node.js applications for reduced latency accessing z/OS data/services Up to 2x better throughput, 60% faster response time to DB2 on z/OS* Security and Dependability Leverages the trusted environments of z Systems to maximize security and uptime of critical Node.js applications. Unified Diagnostics Tooling with IBM SDKs v1.2 for Java® Monitor your application with IBM HealthCenter Debug your application using Interactive Diagnostic Data Explorer The IBM SDK for Node.js ™ provides a stand-alone JavaScript ® runtime and server-side JavaScript solution for IBM platforms. It provides a high-performance, highly scalable, event-driven environment with non-blocking I/O that is programmed with the familiar JavaScript programming language. The IBM SDK for Node.js ™ is based on the Node.js ™ open source project. It provides a compatible solution for IBM System Z products that require Node.js ™ functionality and package management. Node.js has one of the fastest growing eco-systems out there. More than +93k javascript modules can be quickly and easily downloaded from npmjs.org to build complex and sophisticated javascript applications with ease and speed. According to http://www.modulecounts.com, Node.js was the fastest growing eco-system amongst languages – the rate of modules contributed was around 3x that of any other language in August, 2014. Acme is a different benchmark than the one we measured 63% on in the chart, that one is a web-serving benchmark. Acme is an airline reservation system. acme airlines is a Node.js benchmark that a (now ex- ) IBMer wrote http://ispyker.blogspot.ca/2013/05/announcing-acme-air-performance.html, We ran it on Intel and zEC12. It was 29% faster on the latter. AcmeAir Benchmark (https://github.com/acmeair/acmeair) shows an implementation of a fictitious airline called &amp;quot;Acme Air&amp;quot;. The application was built with the some key business requirements: the ability to scale to billions of web API calls per day, the need to develop and deploy the application in public clouds (as opposed to dedicated pre-allocated infrastructure), and the need to support multiple channels for user interaction (with mobile enablement first and browser/Web 2.0 second). The AcmeAir benchmark was installed and measured: Intel Xeon E5-2670 @ 2.60 GHz using the 32-bit Joyent 0.10.30 Node.js IBM zEnterprise EC12 using the 31-bit IBM SDK for Node.js V1.1 Server side scripting Scripting to next level, write MW in script language Cloud foundry written in Ruby, so no longer use scripting lang to shift files but use them for more interesting tasks For node, used for web serving, web pg serving, data processing No differ betw java and java script Server side: SW being used to run on a server, so this is why perf is critical cause servers are expensive, so people pay attention to perf, security, so need to build the IT infra, want to do more with less – do transaction processing, data serving Lang that allows u to write apps so they run on servers as opposed to old way where scripts to automat something were on the client side. On the server side, say on bank app, goes to a server transaction with some QoS e.g has to be secure, so bank processing billions a day – infra in place with HW, O/S and MW that runs on top of it e.g a resilient database like DB2, IMS, then surface it to the mobile app in a real way – so when build app to run on server, the language is differ cause it biz logic for the app and the infra that runs this biz logic to manage that transaction envir Net is its goes from simple things to now more complex things Now enable yourself to do this on a server side In past scripting didn’t have to work all the time but in server envir has to work all the time, has to scale, be fast, be efficient – so scripting used to do something in a different way The 2x perf is mainly due to the IO capability of the platform Module counts - # modules in different repositories that cover differ languages, sharing libraries that you built that others can use Support: we r building assets with intent to support, on a customer by cust baiss, then eventually have a PID in place to have a chargeable service