Cloud

Oracle OpenWorld Keynote: Renee James President of Intel

So far Oracle Open World is full of pomp and circumstance with a lot of messaging around how a combined set of platforms with database, cloud, hardware, and software on all of it positions Oracle better than anyone else.   We can’t play cloud bingo because EVERYTHING is moving to the cloud but as Oracle notes, an optimized cloud.  Everything Oracle will run better and faster on Oracle Cloud or Oracle hardware (which includes cloud of course).   That can be a compelling message when you are talking about 100-1,000 times the speed in some cases.

But anyway, here’s the stream from the opening keynote.

  • Starts with a video hitting the cloud, Internet of things, data, and the bottom line.
  • tag line: We see what’s coming and it looks amazing
  • Judy Sim, CMO of Oracle kicks off the commentary
    • Note: #oow2014 is the official hashtag

  • Stats
    • 2,700 sessions
    • 60,000 attendees
    • 5 exhibit halls
    • 34,512 chairs
    • 17 venues being used
    • 460 Oracle product demos
    • 200 session labs
    • 7.1 million online attendees
    • 1 activity feed in the Open World App (iOS and Android)
  • Who is speaking?
    • Larry Ellison will speak tonight and will speak on Tuesday at 1:30 PM
    • Mark Hurd will keynote tomorrow morning with companies like Fedex, Intel, GE, etc.
    • Thomas Kurian, EVP of Product Development speaks on Tuesday morninig
    • Renee James, President of Intell will speak tonight
    • Dr. Didier Bonnet of Cap Gemni will speak tomorrow

Renee James, Intel President

Intel has done a huge amount of innovative work with Oracle to make their technology “fly”.

  • Will hear a lot about cloud and big data this week. Renee wants to put this into perspective from an Intel point of view and why it’s a big deal now.
  • It’s not just about building big data centers. Want to push the frontiers of computational science.
  • We are now producing enough data to create incredible insights into the human condition
  • 90% of the data that exists today was produced in the last 2 years
  • by 2020 we will have produced enough data to match the number of stars in the universe
  • Mobile devices produce 27% of data
  • Internet of things today accounts for just 2% of that data and will grow to be about 10% of data by 2020.  It will be REALLy big data
  • The history of computing
    • Mainframe – data measured in kilobytes
    • Client Server – measured in megabytes. More democratic set of systems.  Local area networks were a huge innovation.  Meaningful insight existed but it was beyond what we could store and manage
    • Age of the Web: data measured in gigabytes.    Another big set of changes.   It’s the beginning of the big bang we see now.
    • Mobile Web: data measured in terabytes.  Now we have a flood of data.  Data is no longer IT supporting the business. It could even become the business.
    • It’s about deriving insights in this cloud computing or cloud architecture world
  • Take a look at one IT company
    • MD Anderson Cancer Center.  They view themselves as being in the age of information
    • Their mission is to eradicate cancer….. very ambitious
    • They started their moonshot projects to understand cancer from a molecular basis and move it from the data center to the patient
    • Leverages the Oracle Healthcare data model running on Exadata. Uses cloud storage and computing as well.
    • Believe they can make a substantial impact on Cancer with this approach
  • 75% of the data center buildout has been around the public cloud in the past few years
  • Quote: a private cloud infrastructure can yield the same results around costs, performance, etc. We believe that we will see a hybrid solution.
  • Private cloud grows faster than the public cloud.  47% of workloads are expected to be running in a hybrid mode by 2017 (per Gartner)
  • In the end it will be about compliance, security, and where your workload runs best.
  • Big question is can I improve my efficiency and lower my private cloud costs.  Is it possible for private cloud costs to be similar to public cloud costs?  Intel says “YES”
  • Phases
    • Standardization
    • Virtualization
    • Automation: eliminate the manual processes. (think Dev Ops
    • Orchestration: workloads come as requests, get their storage, networking, etc.   It’s a pool based approach and lets you get optimal performance of your infrastructure. May involve a burst to the cloud when necessary.
    • Intel has found that by increasing server utilization rates, they can get private cloud costs to be competitive with public cloud over time.
    • Security has become a significant investment to Intel.  It includes trusted resource pools, taking into account privacy, putting root of trust and asset tags.  They are working with Oracle on several security optoins
  • Westjet
    • Uses Exadata 3 and 4
    • 50% of $4 billion comes from the web site
    • Used Intel tokenization technology to secure credit card data at the first layer.
  • Next Generation firewall: Steve Grobman, CTO Intel Security Platforms
    • VM’s are exposed to bad neighbor attacks to get around firewalls
    • Intel is working with Oracle to further secure this.
    • Put the Mcafee next gen firewall within the network at the vm level.
    • Take the firewall policies with you as you move VM’s around
    • Also working to improve the traditional firewall.  Both approaches will protect you from the outside and inside.
  • Performance
    • This is Intel’s lifeblood. Now building 14 nanometer transistors
    • Now creating workload optimized platforms.
    • The second part of performance is what you can do with it.  They’ve embraced the challenge.
      • In memory computing. Xeon E7 lets you run the entire db in memory and run at semi-conductor speeds
    • Silicon Photonics: builds optical devices that can move data faster over longer distances. it’s better than traditional copper.
      • Renee showed the first Silicon photonics chip
        • Chuck Duvall demoed it.
        • Large data sets are always growing
        • It’s not always possible or practical to run it all in memory.  When that happens, you have to go outside the server to external storage. The cable to do this is large but it can’t be very long.
        • Showed  three 800 GB SSD’s geting 3.3 million transactions per minute in that server.
        • Then showed a second server outside the rack. It’s “far” away and typically hits your performance. Using the silicon photonics cable gives you almost the same transaction output.
          • The limit is measured in hundreds of meters rather than 1-2 meters.
    • Major investment in non-volatile memory
      • More info to come.
    • Intel and Oracle are working together on Exalytics
      • Balaji Yelamachilii, SVP of Oracle Business Analytics
      • Uses in memory computing
      • Also uses sbase
      • Daytime objective is to run applications with as many users / planners / analysts as they can fit.  Need to get more concurrency.  Worked with Intel to find the inefficiencies and remove the contentions and take advantage of cores to achieve 100% linear scalability
        • This newer system will allow almost double the number of transactions with the same concurrency ….. or allow more users
      • Nightime objective workload needs change.   Now they have a tight window to do the batch jobs. It’s complex processing and are more sequential in nature. Processing speeds becomes more important now.
        • Worked with Intel to further optimize the machine code.
        • The newer system shows a 1.66 X improvement.
      • Same processor but the software and silicon optimization has huge implications.

Follow Perficient & Oracle on Twitter for the latest posts in our Oracle blog and more.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Michael Porter

Mike Porter leads the Strategic Advisors team for Perficient. He has more than 21 years of experience helping organizations with technology and digital transformation, specifically around solving business problems related to CRM and data.

More from this Author

Subscribe to the Weekly Blog Digest:

Sign Up
Categories
Follow Us
TwitterLinkedinFacebookYoutubeInstagram