3. Listening to Cloud Providers & Users
Range of Cloud Service Providers
With Broadest Array of OEM and ISV Collaborators
4. Accelerating Adoption of Cloud
Private
Private Cloud Public
Cloud Public Cloud
Cloud
Today: 14% Hybrid Cloud (Public + Today: 7%
2014: 42% Private) 2014: 23%
>40% of IT operations1 35% by 20152 >40% of IT operations1
1. ODCA global member survey, Oct 2011, N=63
2. Gartner, Dec 2010, N=55 The Road Map From Virtualization to Cloud Computing (G00210845)
7. Customer Security Concerns
Public & Private
Traditional Data Center
Cloud Data Center
Reduced physical control
Increased multi-tenancy
Reduced effectiveness of
existing security tools
IT Pro survey of key concerns:
61% 55% 57%
Lack of visibility inhibiting Lack of control over data key Avoid putting workloads with
private cloud adoption1 concern for public cloud adoption1 compliance mandates in cloud1
1 source: McCann “what’s holding the cloud back?” cloud security global IT survey, sponsored by Intel, May 2012
8. The Need for End-to-End Security Solutions
Secure Secure Secure
Datacenters Connections Devices
Interoperable, Open Industry Standards
9. Securing the Datacenter
Need for Greater Drives New
Controls & Auditability HW and SW Capabilities
Endpoint Aware Integrity
Real-time Integrity
Location Compliance
Control Reporting Security Stack Integrity
VM Integrity
Location & Asset Control
Real-Time Malware Host Integrity
Monitoring Hardening
External Assessment
and Reputation
Digital Certificates
11. Example of Collaborative Security
for a Virtual Cloud World
Virtualized and Private Public Cloud Data
Cloud Data Center Extended Security Policy Center
Isolate, Protect, Control Vms
Company
Provide Visibility & A
Mfg
HR Sales Reporting
Company Company
B C
VMM Apply Security Policy At
Multiple Control Points
Monitor Workloads Across
Cloud Infrastructures
McAfee ePO1
Intel Trusted Execution
Intel Trusted Execution
Technology is run:
Technology is run: Server
“Do not migrate to
“known good”
unknown server”
1 Integrating McAfee ePolicy Orchestrator (ePO) with Intel TXT requires custom integration work
*McAfee MOVE AV = McAfee Management of Optimized Virtualized Environments Anti-Virus
13. MANUFACTURING
HEALTHCARE TRANSPORTATION
ENERGY RETAIL
Cloud 2015: 15 Billion
Connected Devices1
PC, CE, PHONES COMMUNICATIONS
1 “Worldwide Device Estimates Year 2020 - Intel One Smart Network Work” forecast
14. Device Proliferation Challenges
The Perimeter is now
the User and Data
Customers Want
Seamless Interoperability:
Increasing Security &
Liability Risks
Services Span
Range of Devices
May 2010 May 2011
IDC information worker custom survey, sponsored by Unisys, May 2011 and May 2010 Q1 (2011); Q4
15. Client Aware Cloud
Compute HD or SD
CAPABILITIES
Graphics
CONTENT
Flash or HTML5
Media Compressed or
Connectivity uncompressed
Security Secure Content
Location & Commerce
Personalized
CLOUD Balanced Optimized
Services Compute Content
17. What is Big Data?
Traditional Data Big Data
Volume Gigabytes to Terabytes Petabytes and beyond
Velocity Occasional Batch Processing Real-Time Data Analytics
Variety
Centralized, Structured Distributed, Unstructured
i.e. Database multi-format
Untapped Information
18. Why is Big Data Important?
US Healthcare Europe Public Sector
$300 Billion in value/year 250 Billion Euros value/year
~ 0.7% annual productivity growth ~ 0.5 annual productivity growth
Global Personal Location Data Manufacturing
$150 Billion+ value for service providers Up to 50 percent decrease in
Up to $700 Billion value to end users product development, assembly
costs
Source: McKinsey Global Institute Analysis
19. Big Data Solutions: Volume
Traditional Storage Distributed Storage
Architecture
Application Servers
application
storage client
Metadata Storage
Servers Servers
SAN storage
(Storage Area metadata
services services
Network)
20. Big Data Solutions: Velocity
In Memory Analytics Network Edge Analytics
Stream Processing Analysis & Decision Support Applications
Search and Analysis of 53 Million Analyze data as its
Customer Records:
From 2-3 Hours to 2-3 collected to make near
Seconds!1 real-time decisions
1: Hilti Corporation case study
21. Big Data Solutions: Variety
Unstructured Distributed File
Multi-format Data System & Map
Reduce Tools
110100101010100
101010010101010 10101001 Data
010101010 1 Analytics Tools
0
1
0
1
0
110100101010
0
Structured Data Database 1
0
1
110100101010100 0
101010010101010 1101001
010101010
22. Intel’s Big Data Strategy
• Contribute code, promote development
Enable the Apps and new research
• Develop benchmarks; software
Make it efficient optimization libraries
Develop data • Continued compute, network and
center architecture storage innovation
Make it easier to • Publish reference architectures
deploy
Grow the • Education, common usage models,
community open source standards bodies
24. Cloud 2015 Vision
FEDERATED AUTOMATED
Share data securely across IT can focus more on
public and private clouds innovation and less on
management
CLIENT AWARE
Optimizing services based on
device capability
25. Goal of Datacenter Optimization
Cost of Typical Major Internet
Datacenter
Facilities Other IT
Address Networking
6%
5% 3%
~75% of Labor
Ownership 13% Servers
50%
Costs…
Power
23%
Source: Intel internal analysis, 2008 of 3 yr TCO
…via optimized platforms & technologies
Optimized Optimized Scalable Optimized Datacenter
Silicon Technologies Software Systems Optimization
26. Energy Challenges
Age Distribution Energy Consumption Performance Capability
of Servers of Servers of Servers
>4 Years Old
<4 Years Old
4%
>4 Years Old
<4 Years Old 32% 35% <4 Years Old
68% 96%
>4 Years Old
65%
Source: Intel analysis, 2012
27. Value of Time-to-Market for Cloud
Difference in early deployment
Previous TTM
Cost per Web Transaction
Accelerated TTM
Savings
~$750,000,000 Estimated Savings by
Accelerating Latest Generation Servers
Source: Intel analysis, 2012
28. Enabling Open, Interoperable Solutions
Standards Technology Partner Solutions
Learn More at www.livestream.com/cloudslam
& www.intel.com/datacenter
* Other names and brands may be claimed as the property of others.
Presenter: Jason Waxman, GM, Cloud Infrastructure GroupTitle: The Future of Cloud SecurityAbstract:Security is the #1 cited customer concern preventing the broad adoption of Cloud Computing. Intel is changing the game with hardware enhanced security, combined with ecosystem software and services. Come hear Intel’s Cloud Infrastructure Group GM speak about Intel’s vision for the future of Federated, Automated, & Client Aware Cloud Computing environments, better secured via hardware enhanced security. Jason will also provide a rare view into what we hear customers asking for in the Cloud Security space, and the broader industry trends that will shape Cloud Security moving forward. Immediately following this keynote, make sure to attend a customer panel discussion led by Intel.
Key message: Amazing things are happening today thanks advanced Intel technologyGlobal IT Leaders are offering innovative new services that help consumers optimize and personalize their experience This helps consumers navigate a growing array of choices to makes businesses easier to do business with than ever before. Evolving usage models mean that instead of having a bank in every town, we have ATMs in every grocery store and now thanks to smart devices and the cloud we can have a bank in our pocket, right next to our own music store, library and movie theater. We’ve gone from having a wrinkled out of date map of one city stuffed into our glove box, to having an in dash navigation system that not only covers the world, but gives us real time traffic updates and helps us avoid construction. Beyond the sheer usefulness of these new services, this also gives business the ability to form deeper relationships with consumers to drive revenue and margin growth. Business must offer amazing experiences not just to consumers, but to their employees. Researchers, product developers, doctors and content creators all can benefit from the new era of the cloud. Businesses must increase the productivity of their workforces and unlock creativity with powerful workstations that offer near real-time rendering of complex images, and high performance cluster to mine the explosion of data to draw new insights into their customers to offer personalized services. The cloud and increased analytical capabilities available today help them achieve this. Beyond this business can offer more automation to speed application deployments and offer automated provisioning of services and line of business chargebacks based on usage to improve the impact and efficiency of their infrastructure. In total these experiences offer exciting new opportunities for businesses, for IT departments and for telecommunications service providers.
Intel has the broadest base of cloud providers, OEM and ISV partners who we listen to , so that we can understand what the challenges IT is facing today which we then build into our product design plans and research goals.
One of the big things we are hearing from customers and analyst is the continued excitement around the agility and efficiency that a cloud architecture can provide which is driving rapid adoption in the cloud, as well as increasing experimentation in a hybrid model to link public and private clouds.
But to achieve this rapid growth we as an industry need to focus on addressing several challenges. First one of the top priorities in IT today is security concerns, and this is often viewed as a key hurdle to moving to a hybrid cloud model. Another key challenge is the explosive increase in the number of devices connecting to the cloud, and sheer amount of data being generated today. Finally IT shops of all sizes continue to be under tight budget pressure. Intel is focused on addressing these challenges and I’d like to spend some time today explaining our view on these major trends.
Let’s start with secuirty…
Key Message: Cloud security remains one of the top barriers to the adoption of cloud computing and drives a need for new and broader security measures that go beyond traditional enterprise IT security tools and practices. In traditional IT environments, IT infrastructure sits behind the organization’s firewall and equipment that maybe virtualized, or partitioned to handle multiple applications on a single servers, are typically dedicated to a specific line of business. IT professionals can choose from an arsenal of mature security tools that give them a high degree of control over the security environment and the organization’s compliance with regulatory mandates.  With cloud infrastructure, in contrast, servers are typically virtualized and shared across multiple lines of business or even among multiple organizations rather than dedicated to specific lines of business. When IT wants to link multiple cloud data centers together to gain efficiencies—linking a public cloud data center based in Singapore, for example, with their private cloud based in the UK—the tools to secure this far-flung infrastructure are still evolving. The result is that IT loses a degree of control and visibility into workloads and data. This makes it difficult for IT professionals to determine whether the organization is meeting specific compliance requirements, which, in turn, lessens their confidence in security across cloud environments. And it’s not just about the challenges in managing cloud datacenter security, but IT also has other concerns: are the right people accessing the right data, are the devices accessing cloud services protected against new types of malware, is my email and web traffic safe when, for example, I send sensitive email or publish documents onto a public cloud storage service.
Key message: Data must be protected not only in the data center, but in transit over the network, and while on end user devices. Also need the interoperability of open industry standardsThere are 4 areas we are focused on in delivering better security for cloud environments – securing cloud datacenters, securing the network connections, and securing the devices that connect to cloud services. The 4th area is focused on accelerating the development of unified standards for cloud security and enabling a broad range of open, interoperable security solutions via industry collaboration. We are focused on delivering hardware-enhanced security that combines with software and services to deliver stronger, more tamper proof security for cloud environments. I’ll give some examples today and we will be adding more over time. 1. In terms of securing cloud datacenters, our focus is on improving the integrity of the infrastructure while focusing on protecting data, and delivering capabilities to make it easier to audit cloud environments to meet compliance requirements.2. Securing the connections is about protecting data, email and web traffic and applications in flight being transferred between devices and datacenters.3. Securing the devices focuses on ensuring the right people are accessing the right information. It also means knows if the devices are secure and free of new forms of malware when accessing cloud services and it’s about protecting data that resides on the device. 4. Finally, Accelerating the development of consistent security standards that will make it easier for businesses to deploy cloud computing and compare cloud service security levels, along with our collaboration with the broad ecosystem of systems and software providers to enable open, interoperable solutions for cloud security.
Key Message: In terms of securing cloud datacenters, our focus is on improving the integrity of the infrastructure while focusing on protecting data, and delivering capabilities to make it easier to audit cloud environments to meet compliance requirements.IT managers need their datacenter infrastructure to be protected from malicious attacks, and be able to meet all regulatory and audit compliance requirements, which often mean identifying where data is residing and to keep track of all this real time. To the right you can see some of the specific technologies to help meet this challenge. The bottom 2 items are somewhat common today—digital certificates & you if have a tool like McAfee SiteAdvisor that can tell you if a site has been examined for malicious activity. Today, this is essentially all the information you can have when making a connectivity decision. Over time, we will be looking for capabilities to use silicon enhancements to provide stronger, more efficient tools for enforcing controls and allowing more visibility into the protection profiles of cloud-based devices.Feature reference:Server Host integrity – the ability to determine if the cloud-based server was launched into a “known good” state. Intel can provide this today with hardware-based surety via Intel TXTLocation and Asset Control – ability to create location or other physically descriptive boundaries based on system characteristics. Useful for boundary control compliance requirements. Intel can provide this today with hardware-based surety via Intel TXTVM Integrity – The ability to verify that a VM has not been compromised by comparing it to “known good” instances at launch time. McAfee can provide a software whitelisting capability that supports this today and there is potential future benefit for having hardware capabilities to enhance measurement and reporting of the VM imagesSecurity Stack integrity – the ability to assure that the critical security systems protecting infrastructure, workloads and data is operational and up-to-dateReal time integrity – provides continuous monitoring of the critical software in the run time environment. Meets needs for continuous monitoring capabilities specified in a number of environments, such as FedRAMP government recommendationsEndpoint aware integrity - establishes a mechanism for mutual verification of trust and integrity between client and cloud to allow for bi-directional control of access to cloud resources
Key Message: Securing the devices focuses on ensuring the right people are accessing the right information. It also means knows if the devices are secure and free of new forms of malware when accessing cloud services and it’s about protecting data that resides on the device.
Key Message: Intel & McAfee are partnering to provide more points of visibility and control -- across hardware, software and cloud – to enable good security decisions and establish confidence in the cloud as infrastructure.Ideally, IT and security managers can have common capabilities that allow them to have common security capabilities that can span their IT infrastructure types and give them the required controls to manage it all with a uniform security policy framework. Some of the new and extended capabilities they need to put into security management capabilities to address new virtualized or cloud-based multi-tenant models include technologies such as Intel Trusted Execution Technology (TXT) and Intel Virtualization Technology (VT) that can provide hypervisor and virtual machine protection and control points. Intel TXT can also be used to expose platform trust status for reporting dashboards and audit tools to increase visibility into the virtual/cloud infrastructure. Similarly, McAfee has optimized Antivirus via the Management of Optimized Virtualized Environments Anti-Virus, or MOVE AV, solution to enhance the efficiency of anti-malware for the unique needs of virtualized environments. We find traditional AV is often in use in these environments because conventional IT policies state anti-malware solutions as a requirement. The problem is that it results in unpredictable spikes in CPU utilization and high memory utilization. MOVE AV provides optimum protection from malware and advanced threats without the overhead of traditional based anti-malware solutions, helping organizations to better maximize server density and cloud efficiency gains. McAfee ePolicy Orchestrator (ePO) provides a control point for evaluating and enforcing security rules and policy in server and client platforms.These ingredients come together to highlight the complementary nature of the Intel and McAfee portfolio. Today Intel & McAfee will demonstrate a cloud-to-client security solution that shows how Intel hardware security and McAfee technologies can be used to implement security policy based on platform integrity. In this demonstration, a cloud security integrator called Trapezoid Digital Security Services used commercially available platforms to build a solution that controlled access to sensitive resources based on either client or cloud host integrity.
Non traditional devices are becoming connected and delivering real time data and solutions across many key segments. “if it consumes power,  it will compute,  if it computes, it will be connected.”Â
It’s not just about “a” platform anymore. It’s about the overall experience and services the user wants.Consumption is key. If you make it too painful or cumbersome, your consumerization efforts will fail.Key Point: IT consumerization isn’t going away … it will continue to grow and will continue to be a challenge for IT. Cloud provides a platform to provide access to corporate data and applications from anywhere at anytime to meet the challenges of IT consumerization. (1) Devices are changing; (2) Software on these devices are changing. IT consumerization will continue to be a challenge for IT. Computing is moving beyond the PC into a wide range of “smart” devices, driven by the Internet and the need and desire to be “always connected.” New device categories are emerging at a rapid pace to bring computing experiences into every aspect of our lives. Employees ultimately want to have consistency and interoperability across all of their devices, from phone to PC to tablet to TV to gadget, making computing a seamless experience regardless of where you are, what you are doing or what your needs at the time may be. To support IT consumerization, IT is shifting focus to deliver services to any device. By taking advantage of a combination of technologies and trends—such as ubiquitous internet connectivity, virtualization, and cloud computing—we have an opportunity to redefine the way we provide services to meet changing user requirements.
A key element of Intel’s cloud 2015 vision is that cloud services will be client aware. Client aware means that cloud services will be able to both recognize and where appropriate, capitalize on the combined capabilities of the cloud data center and the client device to optimize the deliver of cloud services. Intel’s is delivering client aware technology that will enable web developers and cloud service providers to optimize their services by detecting the compute, graphics as well as network connectivity and remaining battery life for a given users device. This new breed of Client Aware cloud is based on Intel software and hardware on the device, across the network, and in the cloud. By taking advantage of these capabilities, cloud developers can enable services that capitalize on the on compute, context and capabilities making the cloud services secure and smart. Benefits can be measured in an enhanced end user experience, increased efficiency of enterprise resources and stronger security for cloud based services. Examples applications can benefit from taking advantage of client performance include. Personalized Services – 4x (Java Script) -15x (web GL) compared to current browser performance) Faster achieved with multi-core enhanced web delivery to overcome traditional sequential processing as demonstrated at IDF 2011http://software.intel.com/en-us/videos/faster-web-apps-with-data-parallel-java-script-reasearchintel-days-2011/Balanced Compute - Cloud Builders RES - 5xFaster HTML Rendering (local rendering verses cloud) some apps make more sense to run locally (graphical web pages). redir. http://www.intelcloudbuilders.com/docs/Intel%20Cloud%20Builders_RES_July2011.pdfOptimized Content - 2x (compared to prior generation Core) faster media transcoding Intel® Quick Sync as demonstrated @ CES 2011http://www.youtube.com/watch?v=TfANycNU7h8
“Big Data” is a term that captures three differences from traditional data:Volume – the sheer scale of datasets is now measured in Petabytes vs the more traditional Gigabyte sizedVelocity – The pace of business is accelerating so IT needs to provide analytics in near real time vs being able to do batch runs overnight Variety – A lot of interesting data about customers is in social media or other non traditional formats that are not easily captured in transactional database
This is important as it will create opportunities in every industry and will become a key differentiator between competitors.Other ideas – A/B testing analytics increase Obama campaign donors 30% http://www.wired.com/epicenter/2012/04/ff_abtesting/all/1
Intel data: Data Compression 80% Reduction w/ new Intel based storage solutionsAmplidata:Policy-Driven Storage Durability• “Ten 9’s” of Durability (99.99999999%) and beyond through policies• Eliminates the reliability exposures of RAID on high-density disk drives• Eliminates data corruption or loss due to bit errors• • 50-70% reduction in TCO• Storage footprint (Capex), power, data center space & management costsLustre source: Terascala marketing material provided by Greg Scott1000s of Nodes & >200GB’s/sec performance
In Memory Analytics – SAP HANA example of putting an entire database into a large memory systemNetwork edge – Guavus example of pushing data analytics to the edge devices to allow for more rapid responses
Two paths for data today – traditional transaction data moves into standard relational databases for analysis, but now unstructured data from a variety of sensors and sources can be sorted/reduced in a distributed file system to also feed modern analytics engines.
 Keep it open: contribute opens source and ity and enable development and research on the hadoop distribution and uses (like machine learning / mahout)Make it efficient: develop benchmarks and work to improve the scalability and tuning of the code – take advanteage of Xeon, platform featuresDevelop datacenter architecture for hadoop: Big vs small core, fabric integration, flash in platform, etcMake it easier to deploy: development of reference architecturesGrow the community: education, usages model development and integration with traditional data connectorsMake it easy to deploy: Develop reference architectures and POCsMake it efficienct: Drive industry benchmarking suites (such as Hi-Bench for Hadoop)Create usage models to help define requirements; Security, manageability and interoperabilityGrow the community: Address skill set and implementation challenges in enterprise IT environments
Key Message: Intel continues to execute to our cloud vision 2015 To help IT achieve the scale they need to meet the challenges of today, Intel is focused on our vision of the next generation data center. Intel’s cloud vision 2015 is grounded on three key concepts:Federated: allows you to share data securely, spanning public and private clouds. Automated: Dynamically allocate resources, including automation of provisioning, resource monitoring Client aware: optimize cloud services based on the combined capabilities of the cloud data center, communications infrastructure and the device accessing those services. To achieve this vision requires advancements in performance, security, energy efficiency and I/O capabilities. For example as we’ve discussed before every 600 new smartphones or 122 new tablets require 1 server be added to cloud to provide the compelling user experience that consumers demand.Intel believes a similar evolution to cloud will occur in the communications network with vendors moving away from proprietary architectures to more open platforms, that systems will utilize Intel's cloud technologies to consolidate applications on these platforms, and that new services and applications will be deployed to enhance the user experience and increase revenue to the service providers.
Additional resources for high ROI, large scale deployments*Software tuning (SSG), Power mgmt (EPI), Datacenter design consulting (DCAT), HW customization (DCG - HDC)Criteria for add’l resources / supportSize of project >1k servers and/or >$2M Intel revenuesFuture growth opportunity – projected volumes/revenues in next 2-3 yearsIntel share of wallet opportunity – deep engagement with tipping point customerIntel ROIUse the request form at http://xyz.comContacts if want to submit request for additional resources:Hosters / Telco service providers – Adarsh SogalTop 5 internet companies – Sanjeev KhannaEmerging internet companies – Brad ShaferGeo contacts:ASMO: Eric Doyle; EMEA: Alan Priestley; APAC: Dave Chee; China: Evan Zhao; IJKK: Eiji Taguchi
The energy challenges faced by datacenters worldwide are well known and well documented. But among the most advanced datacenters and IT experts in the world – there is significant room for improvement in addressing energy challenges. Here is just one datapoint from a recent energy audit Intel completed with a Fortune 100 company – which we believe to be fairly typical of even technically advanced datacenters. In this study – we found 32% of servers were over 4 years old. But these servers were consuming up to 65% of the total server energy. But even more significant – these older servers were only contributing 4% of the compute performance to the server infrastructure. It’s a great example of why energy efficiency is so important to every CIO – this translates directly into dollars and cents…Talking Points: Core Silicon: Relentless improvements in performance/watt & low-power SKUs - 51% improvement in perf / watt in one generation1Software & Firmware: System instrumentation & value-added middleware to monitor & manage node, rack & row power - Increase rack density up to 40%2System Design & Datacenter Operations: Products and system designs optimized for high-ambient temperature operations. Est. 4% cooling cost savings for every 10C increase in ambient temp31 Performance comparison using SPEC_Power results published as of March 6th, 2012. See back up for configuration details. For more information go to intel.com/performance2 Intel/Baidu Case Study3 Source:http://www.datacenterknowledge.com/archives/2008/10/14/google-raise-your-data-center-temperatureBackground on source of pie chart data:We had inventory details for 2009, 2010, and 2011 from the customer.  There was ~ 1200 servers/blades in their population.  The oldest servers dated back to 2000. Using the identification names for each server, we were able to estimate the date that the servers were installed. We were also able to look at the change in inventory from year to year to determine the installation dates. Using information from Spec.org and other web sources, we estimated the peak performance and peak power consumption for 2S servers/blades for each year.   Because we were interested in computing their capacity, so we used peak numbers for both power and performance.  At the same time, we had an electrician take measurements of every circuit break (~400) in all the PDU’s so we had actual power consumption data for each rack. We did not use the actual readings because the racks were populated with more than 1 server and server type, so we couldn’t get a 1:1 correlation, but we could use the data to cross check our assumptions.  For example, there were 4 ALPHA systems of 2000 vintage installed in 2 racks. The two racks measured 2500W each, so we knew the ALPHA systems were consuming about 1200W each and we could take a judgment for those systems.  I would expect this data set to be different than our own IT data- in the slice labeled “2007 and earlier”, of the 387 servers, 155 were 2000-2002 vintage which had poor relative performance and terrible power efficiency.  The “2008-2009” slice represents 50 servers of the 1200 in total, therefore the power and performance numbers look much smaller simply because of the qty.    This skews the pie slices, but at the same time, it reflects reality.  Corporate data centers have old HW that they can’t shutdown because of the nature of the app, and companies were cautious with investments in 2008 and 2009.     In contrast, Intel IT decommissions anything older than 6 or 7 yrs, they have a more uniform population yr to yr, and they have a more uniform mix of equipment than this corporate data center. Â
Intel recognizes the challenges and opportunities of cloud computing – but we approach these in unique way. Intel is often perceived as semiconductor company – but we recognize that it takes far more than products to address industry challenges such as cloud computing. Intel approaches these issues holistically across requirements, technology enabling, and solution level reference architectures with partners. Specific to the cloud – we are rallying the ecosystem to enable and deliver open, interoperable cloud solutions. We start by driving deep engagements with IT and service providers to understand their requirements and often this means partnering to develop industry standards. Intel responds to these usage models and others that we identify through our end user engagements to deliver products and technologies that meet the requirements of the usage models. And to turn concept to reality – we collaborate with ecosystem leaders to bring solutions to market and define best practices for making these usage models deployable today and optimized for Intel Architecture..