Structure2010

【取材執筆メモ2010】

取材メモ

Structure 2010 †

• June 23-24, 2010 / Mission Bay Conference Center, San Francisco, CA

Day 1 - Wednesday, June 23 †

8:40 Cloudonomics; The Value of the Cloud(Joe Weinman, Strategy and Bus Dev. vp AT&T) pannel †

• seats 1,2,3,4,5,6
  • 1; Joseph Tobolski Senior Executive, Accenture// 2; James Staten Principal Analyst, Forrester Research //3; John Hagel Director and Co-Chairman, Center for the Edge, Deloitte Consulting LLP// 4; William Forrest Principal, McＫinsey and Company //5; Eric Berridge Principal and Co-Founder, Bluewolf // 6; Joe Weinman Strategy and Business Development VP, AT&T (moderator)
  • 4; Cloud is big cost effective, 30 to 40%. Sustainability and Mega scale is almost realized. Private cloud is glow effectiveness.
  • 2; Nikolas Kerr is not correct.
  • 3; Besic view from customer is another cost effective way. But this is fundamental change. Share resource is guess there are no control points. High risk of trouble you can see.CEO issue not CTO issue...this is cloud.
  • 1; Cost is maybe first order. But impact for market place is a big like small company has a big power. Who will be next our competitor? Such questions come up. Big change will be come.
  • 5; We still early stage. Salesforce, Google should be leverage technology they provide.
  • 4; some big datacenter try to build up vertical industry system like banking. Many company does not have real agility, cloud provide these issue.
  • 3; every new enterprise technology is as same. Cloud is same. It's scary.
  • 2; with AP, contents mash-up with company system. IT is tool. IT operation. Salesforce is bypass of IT.
  • 1; Centralization to cloud is (very excite) on every application, inside outside cloud make very easy IT environments. Cloud will accelerate a service consumption of IT.
  • Cloud washing; I am cloud that's it!! No it's not so easy. Share resource...cross corporate organization. 2) speed 6 weeks to 3days its great but behavior is big difference use to be. Capacity planning is opposite. Try to make minimize of capacity. Use to be plan to max capex.
  • 1; to executive. We can provide full system or share resource its 10% discount.
  • 6; <Cloud security? Cost is easy. Agility is hard. So ?>
  • 4; we do basic thing...how to selling your products like that.
  • 3; Tax is very important issue on cloud security. Be careful.
• <Cloud market is bigger than existing market?>
  • Oracle has hard time about cloud.

トップに戻るStructure2010

9:20 Storage and the cloud †

• 1; Stacey Higginbotham, GigaOM//2;Harel Kodesh pres. cloud infrastructure bus. EMC
  • 2; Last 30 years mainframe to pc, data is always somewhere on there. No hard drive anymore. Cloud operation make small margin, data cloud is good business such age. SaaS is hard. Cloud is full service network.
  • 2; All public cloud data will be bigger than all private cloud data.
  • 小池メモ；EMCの親父おもしろくない。クラウドの本質に関する言及なし。
  • Cloud is all about SLA.

10:10 Spiceworks Workshop - Room 1 †

• Building and Marketing Cloud Services to Reach 5 Million SMB IT Pros
  • IT Pros in the world's SMBs control the adoption of cloud services for nearly 200 million employees. Fully tapping this multi-billion dollar opportunity requires a simpler approach to managing and distributing cloud services. Learn how Spiceworks and Rackspace have teamed to bring the cloud into the daily workflow of IT pros worldwide.
  • Jay Hallberg Co-Founder and VP Marketing, Spiceworks siceworks.com

トップに戻るStructure2010

10:50 The state of the cloud address by Werner Vogels, CTO and VP Amazon †

• license has changed for company. What you want to get, we give...like age.
• This year, no talk, let's action. Benefit is so good, why don't you use it?
• digital supply chain...(Ryoji; new cloud definition?)

• -S3; Peak request over 120K per second.
  • FUD; fear uncertainty, doubt is now.
• クラウドはまだだと言う意見に強く反論。つまり、そうした圧力が強いということ。アマゾンはこれを跳ね返せるか。
• a rise of false cloud (Ryoji; interesting mention)
  • server utilization; 10 to 20% is big. Amazon; "reserved utilization" plus "on demand utilization".
  • strategy; see photo
• rise cloud; worldwide application is come
  • but rising the bar; security, operation continuity,
  • the rise of Innovation-as-a-Service; find>connect>innovate>reflect>find...cycle
  • drop io;

トップに戻るStructure2010

11:20 The First step to ROI; analytics and management layers †

• seats 1,2,3,4,5,6
• 1; Alex Polvi CEO and Co-Founder, Cloudkick // 2; Javier Soltero CTO for Management Products, Spring_Source // 3; Michael A. Jackson Co-Founder, President, and COO, Adaptive Computing // 4; William "Skip" Bacon VP of Products and CTO, Virtual Instruments // 5; Jagan Jagannathan Founder and CTO, Xangati // 6; Sebastian Stadil CEO, Scalr
• 6<why need cloud management system>
  • 1; Cloud introduction is making trouble for CTO. Management is new. No tool. Infrastructure of cloud has change and change all time.
  • 6; Change meal not use silo system. More comprehensive management be required.
  • 3; how to reduce daily jobs.
  • 4; my situation has little bit different...utilization maximization. we are more higher level.
  • 2; ROI is abuse word by software vender. Build new infrastructure like cloud, ROI is not simply apply. It's not all about ROI.
  • 3; 100 server integration my client did, agility mean how quick response of market. Beat my competitor. It's one of ROI.
• 6<hybrid cloud how do you think>
  • 5; Real time measurement is critic. Analytics is very important. Data of rich source and scale is very effective. Return result to application is difficulty.
  • 2; Challenge is not technical issue, it's people. People have each different, different scale, different opinion. This makes trouble. Variety of cloud vender has different tool and analytics.
  • 3; traditional datacenter has modified cloud datacenter by HP tools etc. , people should change, high touch. Create new staff, new skill set...Cloud + real-time system + SLA= cloud 2.0
  • 4; skepticism
  • 1; SLA watcher is best. We looking for build up.
  • 5; CIO challenge is how to get data. Analytic data is not deep. Not enough to decision making. How get more specific data you can get.
  • 3; Self space cloud; shrink process...3 month to a minutes. B2C market needs more agility.

トップに戻るStructure2010

11:30 Hitting reset: Reinventing core internet technologies †

• Nick McＫeown)マックキーウォン, Professor of Electrical Engineering and Computer Science, Stanford University.

11:40 Marc Benioff, Chairman CEO, Salesforce.com †

• with Mr. Om Malik
  • Om; What about cloud2?
  • Be; Just come back from Japan, new device and twitter...mixi...they going to cloud. I sift from Amazon to Facebook, MS, IBM is drop off, Salesforce will be a facebook model.
  • Be; Another sift is mobile. We have to transform. New sales force soft on this Friday. Java in core. Look jobs, iPhone 4G and iPad...new way to go. We are going chatter on application. 2008 December, get mail form member, we start to go social software company.
  • Be; Cloud of enterprise should be trusted. Cloud computing should be social with trust. Japan post is best example. They use our technology...keep competitiveness. Our mission is evangelized. Microsoft was agitating our event yesterday. Wow. Cloud + social + mobile
  • Be; 4th cloud...In Japan, cloud...rack and sw...this is a cloud. What a disappointing. This is not a cloud. Cloud is democratic, more efficient. In Singapore, biggest energy consumption is datacenter.
  • Be; google app engine, VMware ...those are more choice for user. Enterprise needs these options.
  • Be; Force.com apacs. Visual.com... We put Java into Force.com. More users can build up their own apps. Salesforce is biggest enterprise cloud computing company.
  • Be; oracle is no hardware, software. Salesforce is no software, hardware.

13:20 Evolving x86 for the cloud †

• 1; Stacy Higginbotham, GigaOM 2; Rick Bergman, SVP AMD
  • 2; use to be performance, now power power.

トップに戻るStructure2010

13:40 Keynote "At the Base of the Clouds" †

• Paul Maritz, Pres. CEO, VMware with Om Malik, GigaOM
  • Ma; 70% IT spend is going maintenance. How to existing infra go to cloud. Cloud is new pool of resources. Those two direction push companies to cloud.
  • Ma; technical issue is not only push. New business model is important.
  • Ma; Roll of OS is changing. Last year virtual server number exceed than real server number. This is new issue. Programmers are shifting to new development environment like Rube in Rail. Hypervisor is commoditizer of resources. "Be cloud" is new programming environment for cloud infrastructure.
  • Ma; we are infrastructure level development. Application protocol is important. Cloud is new hardware environment. Recently new middleware we are pay attention.

14:10 What comes after the Blade? Architectures for the cloud †

• 1,2,3,4,5,6 1; James Watters Senior Manager, vCloud Solutions, VMware // 2; Jeffrey Layton Enterprise Technologist for HPC, Dell // 3; Gary Lauterbach CTO, Sea_Micro // 4; Ian Ferguson Director, Enterprise and Embedded Solutions, ARM // 5; Anant Agarwal Co-Founder and CTO, Tilera // 6; Stacey Higginbotham Staff Writer, GigaOM (modulator)
  • 5; cloud is new game. What is new chip...x86. New multi-core chip...low power those should be need new architecture.
  • 3; massive parallel processing age; cloud...scale out...new set of carnage.
  • 1; datacenter efficiency and cloud are deferent. Cloud is scheduling is important. Datacenter is focusing a performance.
  • 2; HPC and cloud...near but different things.
  • 3; general purpose architecture is vital. Seamicro has deferent! Private memory system...
  • 4; All chip maker focus general use. HPC is as same. But cloud not same, massive scale needs some special.
  • 1; C++ is not high speed, but developer productivity is high. Programming productivity is come 1st, 2nd is program performance. Look facebook.. etc.
  • 小池メモ；昨年は分散データベースに主眼。今年はハードウェア。チップ、マネージメントなどにフォーカス。
  • 3; security and performance is 2 big issues on cloud. Horizontal chip scale out should pay attention on those issue.
  • 5; special purpose cloud...key is few small application should be optimized.
  • 6; How about Rube code optimized chip...how they compete...general purpose chip.
  • 3; Program method change give big impact on chip making. Especially cloud datacenter is.

トップに戻るStructure2010

14:50 The Exascale Grail †

• seats 1,2,3,4,
• 1; John West // Special Assistant for Computation Strategy, US Army Engineer Research and Development Center // 2; Chris Kemp, CIO, NASA Ames // 3; Dave Fellinger, CTO, Data_Direct Networks // 4; Jason Hoffman, CTO, Joyent
  • 3; HPC is tera-flop go to peta-flop world. It's getting worth. 2 to 3 peta byte photo...
  • 2; marze. Exa byte data by day...
  • 1; data generate nonstop, even decision making process. What we try to see real world, we need more computer power. Few tera byte systems are not enough.
  • 4; consumer products become big and high capacity...exa byte. Soon every person generates peta-byte per day. What is happen?
  • 1; Jager is biggest datacenter. 4 giga watt power.
  • 2; Federal datacenter has been integrated into cloud. but depend on type of project. supercomputer is not easy to integrate because difficulty of delivery.

トップに戻るStructure2010

15:55 Venture Catalysts: The VC Panel on Infrastructure Technology †

• seats 1,2,3,4,5,6
• 1; Glenn Solomon Partner, GGV Capital // 2; Michael Skok General Partner, North Bridge Venture Partners // 3; Ping Li Partner, Accel Partners // 4; Promod Haque Managing Partner, Norwest Venture Partners // 5; Satish Dharmaraj Partner, Redpoint Ventures // 6; Lawrence Aragon Editor-in-Chief, Venture Capital Journal and Private Equity Week
  • In 2007 Cloud deal are 18, in 2009 are 27, 2010 more 30 deal on cloud
  • 5; cloud is abuse word. Everything cloud. Elastic computing is basic.
  • 4; we think cloud is application and infrastructure. All service pushes to user...cloud
  • 3; mobile cloud platform is real cloud. User driven application is also cloud. Many cloud there.
  • 2; XaaS is cloud...right!
• <Capital efficiency?>
  • 2; look stable software company...cloud impact come. No startup we interest in. cloud is high scale business. Large player has advantage.

トップに戻るStructure2010

16:35 GigaOM Pro ask the analyst; The Road for the cloud †

• 1,2,3,4,
• 1; John du Pre Gauntt Founder, Media Dojo // 2; Phil Hendrix Founder and Director, immr // 3; Derrick Harris Infrastructure Curator, GigaOM Pro // 4; Mike Wolf VP Research, GigaOM Pro
  • 1; film entertainment is my specialty. Big media companies are looking for more better corporate system.
  • 2; Standards of cloud is not ready. Many group working on the issue.
  • 1; Metadata is key for connecting device and cloud. Digital camera with 4G connection, metadata is first second is data itself.
  • 3; IaaS is too early to market. So deferenceation there. PaaS is easier to market.

トップに戻るStructure2010

16:50 Lunchpad; †

• We selected 11 companies that we think ave the "right stuff" to succeed in the marketplace.
• 司会Lew Moorman, President, Cloud and Chief Strategy Officer, Rackspace Hosting
• Nimbula
  • is a stealth mode software start-up company founded by Amazon EC2 veterans Chris Pinkham and Willem van Biljon. The company is backed by Sequoia Capital.
  • Speaker: Chris Pinkham, Founder and CEO
• Cloudant
  • Offering data management, search and analytics as a cloud service, Cloudant leverages decades of "big science" experience to solve next- generation data problems.
  • Speaker: Adam Kocoloski, Founder and CTO

• Solid_Fire
  • is a next-generation block storage platform for cloud computing providers and other enterprises that need scalable, reliable storage for thousands of servers.
  • Speaker: Dave Wright, CEO
• Datameer
  • Founded by Hadoop veterans in 2009, Datameer's breakthrough product, Datameer Analytics Solution, builds on the power of Hadoop to deliver easy-to-use and cost-effective big data analytics.
  • Speaker: Stefan Groschupf, CTO

• North_Scale
  • provides a new class of data management technology that scales out, enabling lower data management costs while improving scalability and performance of web applications and cloud computing environments.
  • Speaker: James Phillips, Co-Founder and SVP, Products

• Grid_Centric
  • Infrastructure software that provides the scalability, efficiency, flexibility, and ease-of-use of next-generation cloud services.
  • Speaker: Tim Smith, CEO

• Nephosity
  • is an enterprise cloud computing solutions provider whose flagship product, "pomsets", is a computational workflow management system for the cloud.
  • Speaker: Michael Pan, Founder and CEO

• Cloud_Switch's
  • innovative software appliance enables enterprises to move their existing applications to the right cloud computing environment-securely, simply and without changes.
  • Speaker: Ellen Rubin, Founder and VP Products
• Riptano
  • provides professional support and services for Apache Cassandra.
  • Speaker: Matt Pfeil, Co-Founder
• Greenqloud
  • is the world's first truly green public compute cloud.
  • Speaker: Eirikur S. Hrafnsson, CEO and Co-Founder
• Zettar
  • creates software that seamlessly integrates an enterprise's existing applications with private and public storage clouds, enabling enterprise storage to move effortlessly to the cloud.
  • Speaker: Chin Fang, Founder

トップに戻るStructure2010

Closing Remarks †

Day 2 - Thursday, June 24 †

8:05 AM Scaling the Database in the Cloud †

• As we move our legacy applications to the cloud we are discovering that not all elements scale equally -- in particular, legacy databases. In this panel we investigate the differing approaches taken by hot new technology startups and the options customers have when it comes to choosing between scaling legacy systems or transferring to new database platforms.
• seats 6,5,4,3,2,1
• 6;Matthew Sarrel Executive Director, Sarrel Group // 5; Roger Bodamer SVP Product and Engineering, 10gen // 4; Emil Eifrem CEO, Neo Technology // 3; Mike Hoskins CTO and GM, Integration Products, Pervasive Software // 2; Paul Mikesell Founder and CEO, Clustrix // 1; Amit Pandey CEO, Terracotta //
• 6; Let's talk Scale db. How do you think.
  • 5; rmdb is not adapted to cloud. No scalable, no administration design, many
  • 3;シーコルはだめ。SQL will be survived maybe. Post relational age will be.
  • 2; Lack scalability is key. RMDB will use some part of cloud.
  • 5; Not only scalability, but also modeling structure,…
  • 4; 1st seek data structure, then looking for best database. This is best way.
  • 1; looking workload, scale, data sets…depend on. Almost DB approach is not realy new.
  • 5; RDB is not all mighty. In scalability area, no scale db. Highly connected data should be other selection but scale.
  • 2; If you are cloud, traditional db is not scalable….real deferencetion between old and new are parallel and scale in single box…distribute adaptation…
  • 1; Each case, different problem…we solve depend on selecting best db set.
  • 5; Hardware is commoditized.
  • 3; Some issue is adapted scale db. Rdb is so popular. Very early stage of cloud, there are large gaps between them.
  • 2; Don't throw away scale. If you don't relational system, we trouble.
  • 3; Gaps between existing system and cloud, should fill up …but no throw away scale reason. We are testing big table type of DB, there are no characteristic like scale.
• 6; Licensing is big issue. How do you think?
  • 1; We pay too much to oracle. There are a lot of free open source.
  • 2; Depend on issue. But you can find best license model.
• 6; Almost programmer study scale/ rdb. What they do?
  • 1; Vast most data is just transit one, not input/output db. Cash will be place for them.
  • 4; Programmer love scale and hate scale. In case of cloud, many case scale is pain.
  • 2; But scale has a lot of functionality, it's pain but good point also.
• 6; Scale will be in enterprise cloud ? Large net business concentrate on large scale db.
  • 4; There are lot of scale position in enterprise.
  • 1; Distributed system will be large scale db system not scale in house.

トップに戻るStructure2010

8:45 AM KEYNOTE: †

• Loom CTO Rackspace hosting & Om Malik Founder,GigaOM
  • We start cloud business few years ago, now 10 % of revenue come from cloud. Still 90% is traditional. Goal will half and half.
  • Cloud is huge momentum, company has flexible and new set of tools. IT dept. priority should change. Of course don't touch oracle db, but other area should change.
  • Everything is not change in cloud age, but lot of change. Dell targeted cloud customer. They adjust like us.
• Regacy hosting is high margine business, but cloud is?
  • Cloud now is workload off use…like buch process etc. Last wave of internet, open stack, now costume technology support cloud. These technology is first wave…then open system come. We support a lot of open source projects. Next generation service like that.
  • Yesterday, launch pad…private cloud come. Seamless integration with your existing system and cloud…that very important.

トップに戻るStructure2010

9:10 AM Dealing With the Data Tsunami: The Big Data Panel †

• We're hurtling into an era of machine-generated data and the real-time web, but the fact is, we have too much data already. So at what point will extracting knowledge from all that data no longer be cost-feasible? In this panel we ask what strategies and tactics can be implemented today in order to ensure that we are prepared for tomorrow. The panel also considers current state-of-the-art best practices in large data management, both their benefits and downsides.
• seats 6,5,4,3,2,1
• 1; Phil Hendrix Founder and Director, immr // 2; Amr Awadallah CTO and Co-Founder, Cloudera // 3; Damian Black CEO, SQLstream // 4; Hagan Hacigumus Department Head, NEC Labs // 5; Todd Papaioannou VP, Cloud Architecture, Yahoo! // 6; Barry Zane Founder and CTO, Par_Accel
• 1; What is big data diffinition?
  • 2; Depend on situation, purpose, case..
  • 3; It's big meal, scale cannot keep up…such as.
  • 4; Not size…how fraction, how difficult of management etc.
  • 6; data cannot use traditional method is big data. Not size, complexity.
  • 2; our client 20 gigabyte every night process…but this is big data for them.
• 1; Gove me a overview.
  • 2; Ebay is 3rd customer of Hardope, few petabyte data they has…type of algolism is complex. Univ. of Fhinx…
  • 3; Streaming data project for Australian Gov. Few minute delay has critical, so big limit. Max throughput should be guaranteed. Very exited project.
  • 6; Fidelity case….card fraud system. Pure information model, 250 million per year they save.
• 1; Yahoo how about?
  • 5; we are big user of big data. 600 million user everyday, lots of lots of data. Spam filtering, user analytics…we appreciated power.
• 1; delay, analytic, etc. What is a problem of big data world.
  • 4; Big data area is not yet be payback compare with investments….will be change.
  • 2; Archive, Data mining
  • 5; Gig data world…
  • 6; We involved large retail data…we analyzed those data…Unfortunately user not understand what they can do. 200 terabyte data…looking for cheaper method, it's need.
  • 3; Cost use to be few million dollars need, now much lower to get same results.
• 1; What type of question user ask?
  • 2; Hadope; 1) data size 2)data type, complexity 3) algorism those three should be ask. ROI is important. It is justifiable?
  • 3; Next generation data should be real time…like base ball results…it is contents. Existing system not adaput those needs.
  • 6; What we need is flexibility. Real enterprise…real money. Intelligence is about life…Not easy to ask new technology. Be careful.
• 1; New solution for big data?
  • 4; There are no single solution. Value side is business intelligence and real time analytic. Real-time information sheraling is another big issue to think about.

トップに戻るStructure2010

9:50 AM KEYNOTE: Unleashing the Business Potential of Cloud †

• Cloud computing has tremendous potential. The model can enable service innovation and provide a differentiated customer experience, while creating a lower IT cost structure with increased flexibility and resiliency. But there are very real challenges that enterprises need to overcome in order to scale and fully leverage this model across their businesses. Erich Clementi, IBM's vice president of corporate strategy and general manager of Enterprise Initiatives, will sit down with Om Malik of GigaOM for a fireside chat to discuss the adoption of cloud computing, future challenges for CIOs and the implications for business model transformation.
• Erich Clementi VP, Strategy and General Manager, Enterprise Initiatives,IBM & Stacey Higginbotham Staff Writer,GigaOM
• Stacy; making cloud practical?
  • Erich; Now cost model, new price point. Incubation program…let's talk. US we estimate more 1 billion dollars, data transfer fee is very high in clinical industry (Hospital) We try to build up clinical cloud. Make new price and cost.
  • Enterprise pay attention about privacy and security… They accumulate data themselves. It is trust me model. Hospital industry is as same as enterprise.
  • Information exchange in clinic…special purpose cloud. Multi-tenant makes work and gives benefit. New York and San Francisco, data should be share for same customer.
  • 200k people, 40 to 50 application, 1 petabyte, huge scale of data…this is one of our case.
• It's Salesforce for hospital. How do you see in five years? Who have db, server , data….
  • Cloud is next big wave. All IT will be impacted and have influenced by cloud in 5 years. Re-engineering come by cloud. Security, privacy and data integration….so many silo system existing. Those system and data should be integrated. Use to be so expensive…Service management, CRM those will be change.
  • Customer control cloud is important. (??)
• What IBM expecting enterprise cloud?
  • Work off load is main solution of cloud…then risk management work is big category. Business work outsource become horizontal. They come to your office.
  • Next year, IBM try to come more customer. We have so many technology…cloud capabilities…security, privacy, latency…we are ready to support them. We have very good middleware this is what we want to discuss with customer.
• Credit company case how about ?
  • Good case. Data should belong channel by bank. Data integration in this area are huge venefit.
  • We see cloud is big chance. Like Tyco we have new business chance can generate.

トップに戻るStructure2010

10:20 AM BREAK †

• (Optional Workshops - Located on Level 2)
  • Dupont Fabros Technology Workshop - Room 1
    • State-of-the-Art Data Center Practices for Cloud Computing
    • This workshop panel will bring together the practitioners behind some of the world's most advanced data center facilities. As the function of the data center's shifts to in-house clouds we discuss the issues they're facing and what's being done to address them. GigaOM Pro Analyst Tom Trainer moderates this workshop, which is brought to you through the generosity of Dupont Fabros Technology.
    • Lee Kestler SVP, Sales and Leasing, Dupont Fabros Technology & Mike Wolf VP Research,GigaOM Pro
• Bluewolf Workshop - Room 2
  • Sustaining Cloud Value Post Go-Live
    • You're in the cloud - now what? How do you continue the value of the cloud? Time and time again, companies implement cloud technologies and see a decline in user adoption and overall effectiveness. That is about to change. Hear agile pioneers offer battle-tested tips on creating scalable programs and sustainable profits.
    • Jesse Endo Director, Bluewolf Beyond, Bluewolf & Jeffrey Lane Senior Manager, Coverity
• Zeus Technology Workshop - Room 3
  • Best Practices for Fast, Secure, and Available Application Delivery Across Physical, Virtual, and Cloud Environments
    • Learn how to increase web application speed, security, and availability with the Zeus Elastic Application Delivery Platform. Find out how Zeus customers have been able to use cloud infrastructure to automatically burst application capacity at peak times, manage application delivery across hybrid physical data center and cloud environments from a single point, and minimize the cost of application infrastructure.
    • Kosten Metreweli CMO, Zeus Technology トップに戻るStructure2010

11:00 AM Pushing the Cloud to the Edge †

• The cloud products Akamai unveiled at Structure 2009 pre-empted the needs of the market, as they were the result of the company pushing compute and storage services to the edges of the network in order to provide the best cloud performance. Om Malik sits down with President and CEO Paul Sagan to talk about what's happened in the year since what Akamai is doing to further shape the future of the cloud.
• Paul Sagan President and CEO, Akamai & Om Malik Founder,GigaOM
• "smugmug" photo share site is a good example. Cloud should control . Cloud optimization for PaaS; ex is Starbucks site. Cloud optimization of SaaS is Bullhorn.
  • Optimization for all layer…From platform to application. AWS type is hosting mode…mean cannot control internet network…real-time performance. Mission critical application cannot work without network. General 2000 server work for deliveling data.
  • We have terabit data stream like world cup this week. Akamai is not cloud provider, but cloud player. Outside of datacenter we play.
  • Large enterprise is main customer. Our business transform…big hospital…vertical service. Optimization business is growing. But HD is big..game is also big. Internet deliver video like TV.
  • VDI(virtual desk top) come big. Mobile will be also big traffic. Rich media web on mobile…edge software optimize contents and monitoring for mobile devices.
  • Cost is big driver…device base optimization like iPad or iPhone is coming. Data privacy…gov.

トップに戻るStructure2010

11:20 AM Fatter and Faster: Reducing Latency and Increasing Capacity on Broadband †

• The cloud computing revolution, in order to realize its full potential, is dependent on one key factor: high-capacity broadband. As customers deploy the cloud they are finding that they need ever-fatter pipes and lower latency than ever before. This panel investigates the perspectives from the broadband service providers and broadband technology providers.
• seats; 6,5,4,3,2,1
• 1; Dave Asprey CTO, Pulse_Tracer// 2; Mark Day Chief Scientist, Riverbed Technology // 3; Baruch Deutsch Senior Director of Product Marketing, Cisco // 4; Kenneth Duda VP, Software Engineering, Arista Networks // 5; Michelle Munson President and Co-Founder, Aspera // 6; Lane Patterson Chief Technologist, Equinix
• 1;Network capacity…large file transfer…this is bottle neck?
  • 6; Ethernet come after T1…100M…1G now going fat pipe.
  • 4; Teleco still depend on voice not data. This economic will change…scale of data is most important.
  • 2; Key point is….SFO is good…others are bad connection. Depend on place.
• Bigger pipe…and optimization what happen?
  • 3; many optimization technologies are last century tool. Customers are looking for integration solution. Not single physical boxes.
  • 5; Our case unstructured data file transfer…especially large like life science. Deployment is important…transfer or synchronize …depend on purpose.
  • 4; Optimization by software is so complex…virtualization will change situation…switch edge router those has changed…
  • 6; All user don't like buy box instead cloud.
  • 2; All application but application become complex. How to make solution.
  • 5; Automation will take care of application complex issue..
  • 3; some service missing middle layer (optimization) …should use existing technique.
• 1; Net guy are looking for social networking type service. Does we optimize again?
  • 3; Web caching help. No problem maybe.
  • 4; Network transpalency is one way of optimization. Big fat pipe is always good.
  • 3; Public cloud, less 10 mili sec no proble. More than 10 m. sec should some kind of optimization need.
• 1; netowork capacity never pass Processing speed…so
  • 3; Cloud in enterprise is slow because security, privacy…and performance.
  • 5; Geography is not real issue, virtualizations more bigger on performance.
  • 6; 10 years later, network is not so big but connection and automation is more sophisticated.
  • 3; back to latency issue, xml, tcp, http all are make latency….accumulation is not so easy to solve.

トップに戻るStructure2010

12:00 PM Facebook: The Demands of Webscaling †

• Jonathan Heiliger architects the infrastructure for one of the most demanding applications on the web: Facebook. In a fireside chat with Om Malik, he shares some insights gleaned from working at the social networking site as well as from his stewardship of the exclusive Webmonsters group, which brings together keepers of the largest webscale infrastructures in order to learn from one another. Not to be missed.
• Jonathan Heiliger VP, Technical Operations, Facebook & Om Malik Founder,GigaOM
  • large flat homozinius
• Have you seen semiconductor infrastructure…
  • chip test…application test…they change. Large multi-core like seamicro…is interesting. Programing model change…massive parallel programming changed… we take advantage.
  • Scale out is large apply on facebook application. Start from PHP that is ease to deploy, but performance is not so good. Hip-hop PHP it's mix-up PHP and C++ now.
  • Technology is all behind curtain but is important. Important system like personalization is depend on small good staff.
  • In future, facebook infrastructure can be cloud...maybe a part of. No more back to buying serve router etc. It's so easy even cannot have 100% control.
  • Biggest lesson last 12 months are…
  • Facebook is destination web site…this is a goal. 5million access per second.

トップに戻るStructure2010

12:20 PM LUNCH(Optional Workshops - Located on Level 2) †

• EMC Workshop - Room 1
  • Supermassive Storage: Cutting Edge Practices from the Service Provider industry
  • The Service Provider industry has some of the most mammoth and demanding requirements for provisioning cloud-based storage. Their current innovations and practices will become the common place practices for tomorrow's F500 corporation and their data storage solutions. In this workshop we examine how EMC's innovative Atmos solution is being used by the biggest service providers to solve their extreme storage needs and how they cope with policy management, multi-tenancy, network intrusion prevention and data federation when serving global-sized cloud customers.
  • 1; Gary Orenstein Host, The Cloud Computing Show // 2; Tobias Ford Assistant VP, Cloud, AT&T // 3; Tim Varma VP of Product Management, PEER 1 Hosting
• Joyent Workshop - Room 2
  • The Promise ... and Reality ... of the Cloud
  • Is the cloud delivering on its promise? If not, where are its limitations? What are the best use cases for cloud deployment? Hear from one enterprise that recently moved its legacy business to the cloud, and two others that grew their businesses within it. What challenges did they face? How did they overcome them?
  • 1; Chris O'Brien Columnist, San Jose Mercury News // 2; John Igoe Director Cloud Software Solutions, Dell // 3; Rafi Jacoby Director, R&D, Context Optional
• Par_Accel Workshop - Room 3
  • Virtualized, High Performance Strategic Analytics
  • When your data holds hidden answers and your need to know is urgent, serious analytic power is essential. The Par_Accel Analytic Database delivers near real-time answers to your toughest questions in a virtualized environment. Learn how Par_Accel and VMware bust the performance-penalty myth -- strategic analytics thrive virtualized.
  • 1; Barry Zane CTO, Par_Accel // 2; Bob Goldsand Senior Technical Alliance Manager, VMware

トップに戻るStructure2010

1:20 PM SaaS and Open Source †

• What if you were to provide open source applications as a service (SaaS) ? The potential low total cost of ownership and extensive customization make the premise very attractive to many customers out there. Could it be that simple? And what of the hidden costs in customization and maintenance? We ask two of the valleys most noted entrepreneurs how they are pushing ahead with a vision of free software delivered as a service and what value is provided to the customers and the coders.
• Seats; 1,2,3
• 1; Kim Polese // 2; Larry Augustin CEO, SugarCRM // 3; Marten Mickos CEO, Eucalyptus Systems
  • 2; SaaS should be open…will be open. Intuite become SaaS platform. Lock in become price up.
  • 3; Destructive technology…open source…new arena oss is creating. Oss builds up infrastructure. Oss user have own benefit and return to community.
  • 2; Google, facebook…open source…participation become more natural to take benefit.
• 1; Open source first "code open" to "API open" How you think?
  • 3; Bank is safer than home to keep money. SaaS is as same. Data, Application…SaaS is better than In-house datacenter.

トップに戻るStructure2010

1:40 PM Guru Panel: Scaling Fast and in Real Time †

• We assemble a panel of some of the top innovators -- and biggest infrastructure customers -- in the web world to talk about their experiences pushing the boundaries of cloud computing. Core topics will include how to scale quickly, considerations that need to be made when it comes to real-time web experiences and what technologies they want to see developed.
• seats 5,4,3,2,1
• 1; Jonathan Heiliger VP, Technical Operations, Facebook // 2; Matthew Mengerink VP of Customer Quality, Engineering Services and Site Operations, Pay_Pal // 3; Tom Mornini CTO and Co-Founder, Engine Yard // 4; Jay Parikh Director of Engineering, Facebook // 5; Mark Williams VP Network Operations, Zynga Game Network // 6; Todd, Yahoo
• 1; how products design.
  • 4; Store user data and apply new application…facebook.

    -1: how provisioning;

  • 5; human is always bad predictor.
  • 4; 400million user realtime monitoring.
• 1; Lots of data, lots of appli,…more more…How to avoid application bottle neck.
  • 3;

トップに戻るStructure2010

2:20 PM Rethinking the Homogenous Cloud †

• Many clouds are based on some uniform, consistent building block scaled out to a homogenous infrastructure. While this approach supports scalability and interoperability, it can also result in inefficiency. This presentation will highlight the tradeoffs of conventional thinking and address the implications for future technology and standards to deliver cloud computing that is open, secure and efficient.
• Jason Waxman GM, High Density Computing, Intel Corporation

• Delever best service…
• Goal of Cloud datacenter…3 area; Federated and Automated and Client Aware.
• Economies of Scale; Tigh coupling across software, hardware, datacenter. Cloud…50 % are servers, 23 % is power…
• homogenous cloud (today) vs. Best breed cloud (tomorrow). Security; trusted compute pools.
• Cross Platform Cloud Computing. Secure multi-tenancy, Federation…

トップに戻るStructure2010

2:40 PM Working with the Hyperscale Providers, Lessons from the Cutting Edge of TCO and Innovation †

• Four years ago a group was formed within Dell to target the emerging hyperscale market. This group, led by Forrest Norrod soon developed into the infrastructure partner for some of the world's largest cloud computing providers and hyperscale data center operators. Come hear what Forrest has learned working with these customers who operate at the cutting edge of efficiency and how he sees the cloud and hyperscale market evolving going forward.
• Forrest Norrod VP and GM, Server Platforms, Dell & Stacey Higginbotham Staff Writer, GigaOM
  • Dell is 3rd in Server. Behind HP and IBM. So vertical team organized…try specific industry needs. How isolate our team from others…important. 2006..2007…How measure efficiency? Power, maintenance, operation datacenter.
  • what type of venture, dell wish to working with? Basic power, memory, scalable, software building block…these areas.
  • Critical thing is optimized system we provide.
  • within dell, start mobile devices which need grow.

トップに戻るStructure2010

3:00 PM BREAK (Optional Workshops - Located on Level 2) †

• Dell Workshop - Room 1
  • Two Paths to the Cloud - Are You Looking for a Revolution or an Evolution?
    • There are two complimentary paths to cloud building, evolutionary and revolutionary. One focused on infrastructure and traditional enterprise applications the other, platform focused and based on cloud-native apps. While evolutionary dominates today, is the revolutionary approach where the future lies? Come learn more about these approaches and how each can benefit you.
    • John Igoe Director Cloud Software Solutions, Dell
• Zuora Workshop - Room 2
  • Deploying Cloud Commerce: Learn How to Meter, Price, and Bill in the Cloud
    • How do you properly monetize your cloud service offering when legacy systems cannot support the new world of pay-as-you-go, usage-based billing? Learn how EMC, Sun, and others succeeded in launching new cloud business models in record time. Based on Zuora's platform and experience deploying cloud commerce solutions, this session will walk through multiple demos, use cases, and best practices for deploying cloud commerce.
    • Madhu Rao Director, Solutions Engineering, Zuora & Lew Tucker Former CTO, Cloud Computing Business Unit, Sun Microsystems
• Equinix Workshop - Room 3
  • Creating an Enabling Architecture for Cloud Services
    • Having great application software and fast hardware are only half the battle. A scalable deployment architecture is equally necessary to achieve the economic and operational advantages of mainstream cloud services. This workshop explores key elements of cloud deployment architecture and calls attention to the considerations necessary for a high quality end-user experience.
    • Darryl Brown Cloud and IT Services Vertical Marketing, Equinix

トップに戻るStructure2010

3:40 PM The Platform-as-a-Service (PaaS) †

• PaaS is your choice when you need more flexibility than SaaS but do not want to manage your own infrastructure. In this panel we will investigate whether focused propositions will win out, or whether the more broad efforts aimed at the Enterprise will take the cake. Will interoperability efforts lead to their commoditization or will verticalized plays hold their premium? The panel brings together distinct viewpoints to debate these and other questions about the future viability of PaaS.
• seats; 5,4,3,2,1
• 1; Derrick Harris Infrastructure Curator, GigaOM Pro // 2; Yousef Khalidi Distinguished Engineer, Windows Azure, Microsoft // 3; Lew Moorman President, Cloud and CSO, Rackspace Hosting // 4; Byron Sebastian CEO, Heroku // 5; Mike Piech Senior Director, Product Marketing, Oracle // 6; Tien Tzuo CEO, Zuora
• 1;What is PaaS
  • 2; Definition is broad. Package software should take care. We only compeat. Scale and location free.
  • 4; Customer relation software. PaaS
  • 5; Private cloud…agility, flexibility…siro system to integrated system.
  • 3; IaaS is new computing stack…PaaS is new この質問にみんな混乱した答えを出すのは意外。
  • 5; Single application run on promise or public cloud…
  • 6; We familier mainframe stack…Serever/client stack…now cloud. Cloud stack not only you but also your customer can reach easily to your system. This is new.
  • 4; PaaS is more wide range reach, more scale…
• PaaS priority…Cost or performance, capability…
  • 4; Customer focus on capability. Cost is not our focus.
  • 1; Cost is very important. But number one is agility.
  • 6; Elasticity is important. It's cloud.
  • 5; If you are looking for specific use like huge scalability or massive processing…you can easy to chose.
• 1;
• 2; Easy management is important. And cost is important like pay as you use. SLA yes it needs. But anyway how keep your past investments.
• 3; Marketing is key, to easy maintenance…user should understand.
• 4; Value of PaaS is share knowledge of users. Price is important but value also need.
• 5; User select right server, cpu, memory, transaction…
• 6; PaaS too wide range from Microsoft azure to SaaS like Salesforce….too early.
• 1; lock in language, platform…
  • 3; in PaaS, it is adaptation
  • 2; Data import export is important.
  • 4; Open standard, open source….what else? No proprietary.
  • 6; Anyway platform is lock-up whatever you are.

トップに戻るStructure2010

4:20 PM Hybrid Clouds -- The Best of Both Worlds? †

• Mix together an organization's existing infrastructure with that of a cloud service provider and you have what's known as a hybrid cloud. But while hybrid clouds solve numerous problems -- namely by mitigating risk, capital expenditure and uncertainty -- are we diluting the real underlying value that each approach offers on its own?
• seats; 7,6,5,4,3,2,1
• 1; James Urquhart, Senior Market Strategist, Cloud Computing, Cisco // 2; Michael Crandell CEO, Right_Scale // 3; Marten Mickos CEO, Eucalyptus Systems // 4; Tom Roloff SVP, EMC Consulting, EMC Corporation // 5; Randy Rowland GM Managed Hosting and Cloud Computing Services, Terremark // 6; Sinclair Schuller CEO, Apprenda // 7; Joe Wildman, AT&T
• 1; what is hybrid cloud?
  • 2; internally siro sytem, outside integrate.
  • 7; Appli mash-up, Capacity assist etc. many type. One sigle platform bridge enterprise and cloud is wrong. Frontend data storage type use is mission continuity not involved. Many type.
  • 5; Legacy and new cloud conbine.
  • 4; Many enterprise start journey of cloud, Private cloud is first…integrate public cloud is second stage…maybe.
  • 2; Opposite case like start AWS and go back to private cloud. ＡＴＴのジョーは既存大企業のクラウドを頭に描いて話し、ライトスケールのマイケルはAWSのようなグリーンフィールドを描いて話す。
• 5; Security, internal rule.. makes hybrid cloud.
• 7; Cloud need big pipe (like wide area ether network)

トップに戻るStructure2010

5:00 PM Different Clouds, Different Purposes: A Taxonomy of Clouds †

• When web upstarts started offering cloud services, they stirred the sleeping enterprise giants and hosting companies into action as well. The buyer is now spoiled, for from the bedroom to the boardroom there is a vendor that can offer it the right cloud. But how, exactly, are they different from one another? Will they happily co-exist together? Is interoperability even desirable? Or is it a path to creating commodity and mediocrity? You decide.
• 6,5,4,3,2,1
• 1; Stacey Higginbotham Staff Writer, GigaOM // 2; Val Bercovici Cloud Czar and Chairman of SNIA's Cloud Storage Initiative, Net_App // 3; Joseph Crawford Executive Director, IT Solutions, Verizon Business // 4; Carl Meadows Senior Product Manager, Cloud Services, The Planet // 5; Shelton Shugar SVP, Cloud Computing, Yahoo! // 6; Denoid Tucker VP of Technology, Strata_Scale
  • 5; yahoo has a big legacy system and internal cloud. We decide a cloud is main infrastructure.
  • 6; mid-tea (untie few thousand server) is not need cloud. Really big system you should be cloud.
  • 3; our customer chose agility of cloud.
  • 4; lot of outside of us. They looking for web service provide near to user. It's a cloud they chose.
  • 2; transparency of cloud is sales point. Need standards.
  • 3; PCI compliance is very important for our customer. Certification of compliance is should necessary.
  • 2; very integrated application environment….it is not your core competency, please take it.
  • 5; Simple and automation is another sales point of cloud.
  • 2; global company especially hosting company, change is huge issue like fiber channel. Now cloud come, and so easy to change. Customer is so happy.
  • 6; Dark side is…cloud interface is looks like easy…but customer made so many mistake. Wrong port, not correct set-up…this is make big trouble in cloud.
  • 3; virtual machine tune on is so easy and now real…4 to 5 years from now, you have a cloud commodity environment. We talk cloud like now we talk windows now.

トップに戻るStructure2010

5:40 PM CLOSING REMARKS †

[2010-06-23]