Digital Walkabout

Happy New Year!

Half a day left of 2011, and I have a few moments to pen my thoughts of the year that has passed. 

2011 has been a year which I would never exchange for any other. I have spent it with the most important people in my life. I have survived the end of one startup, and embarked on another. I have faced reversals of fortune, and taken a pig-headed willingness to look facts in the face and turn things around. I have never experienced such highs and lows in the same timeframe. This has been the most satisfying and uplifting time of my life. 

As such I have been quieter than in past years, yet with your support, I have had the opportunity to turn my focus and energies to what really matters. Much of this has been behind-the-scenes work, in place of promotional activities. Much of this also has been taking very difficult decisions and swallowing the bitter pill. The plaudit "no pain, no gain" never has had as much meaning. Growth and development comes at a cost. 

As I’ve been thinking about this past year, I realized I owed my general sense of well-being to all of you. Without you, your support, your feedback (even the harshest sort), and insights, I wouldn’t have the great pleasure of my gig, and being a part of the technologies, models, information and mindshare that are changing our very lives.

Thus, to all of you out there, thank you, truly and sincerely.

No matter where you are or who you are — whether you’re Tory or Labour, Mac or Windows, iOS or Android; Brazilian, Indian, British, American or a citizen of another country; chocolate lover or sorbet fan - have a Happy New Year and a great 2012.

And, yes, I do hope the Maya were off by a fantastic scale. :)

 

Countdown.

I haven't blogged for a bit for a reason. More soon.

A walk in the iCloud

With its roll-out of iCloud, Apple has moved past MobileMe to offer fuller-fledged Cloud Storage, putting into practice their vision of Cloud-based device synchronisation. Yet, what does iCloud really do?

iCloud enables iOS and MacOSX apps to store documents (content) and key-values (metadata) to Apple's Cloud, such that users can view or edit content from any device, without having to sync or transfer files. 

Most apps will use iCloud document storage to share documents, enabling photos, videos, and documents to be accessible across devices; the key-value data store manages metadata, such as storing application state, settings, and other important information that delivers a better user experience. The iCloud service handles storage searches, change notifications, version control, conflicts, and security for applications that integrate with a user’s iCloud storage account.

This architecture is nothing new, but implemented as part of the iOS and Mac OSX platforms, it removes the MacBook or PC as the central hub for managing content, enables over the air (OTA) synchronisation, and is tightly integrated into a slew of Apple apps. 

What is novel is the extension of metadata-driven content management via a Cloud-based repository onto devices to provide de-duplication. Rather than require the upload of data that already exists elsewhere, such as in the iTunes store, iCloud enables the matching and de-dup of such data.

This is the first step in realising a long-desired model for the consumer market. Can and will this extend to enterprises? I see several challenges to overcome in order to do so:

• De-duplication of arbitrary data requires that all content carry enough contextual metadata to be efficient and accurate. This requires either a normalised view of metadata for arbitrary data formats to be supported by iCloud, or that Apple’s iCloud key-value formats be opened up so that they can be adopted. I find it difficult to see either happen. 
• The intelligent synchronisation of Mail, Calendar, Contacts, Documents, etc. has been further advanced in other SaaS and PaaS platforms such as Google’s. Apple’s model has the right idea in mind - the apps work with local data and sync when there’s connectivity. The problems of conflict resolution, caching, and offline devices/modes are not trivial, and in all likelihood, Apple will address them through tight integration and control over what goes into iCloud. This will be hard to sell into the enterprise. 
• Transparent, automatic sync or streaming/push of arbitrary content across devices is a difficult problem. It has taken years of work and certain technological developments to address the performance and security concerns in this space. Apple will certainly encounter these difficulties as it develops iCloud for prime time. 
• iCloud can only be fully experienced within the Apple iOS ecosystem. Both history and Apple’s model indicate that it won’t be opened up to any competing content platforms anytime soon. 

Whilst it will guide the direction of Cloud Storage and re-enforce the validity of the device-centric model, iCloud is clearly focussed on personal content, which plays to Apple’s strengths. This leaves the enterprise market wide open. 

 

Build it and they will come

Jevon's Paradox is the proposition that technological progress that increases the efficiency with which a resource is used tends to increase (rather than decrease) the rate of consumption of that resource. When that increase is not accompanied by any more than organic growth in resource management, it can result in a resource crunch. Simon Wardley of the CSC Leadership Forum has often pointed out that, whilst cheaper operations costs and reduced capital spending should signal a reduction in TCO, the truth is quite the opposite. This argument is rather interesting on its own, but I see the greater issue for Cloud to be resource starvation. There are several parallels which reflect that, despite observed reductions in TCO, the resource crunch has resulted in both increased costs and reduced service.

We have witnessed this effect in the mobile space, where mobile bandwidth has been consumed at an alarming rate with the emergence of data services. AT&T's iPhone service in the US the last few years has only highlighted this, with the network brought to its knees in Manhattan, San Francisco, Los Angeles, Chicago and other major metropolitan centres. Advances in the utilisation of mobile network resources saw their near exhaustion in less than a decade. This has been accelerated and exacerbated by the common practise of planning infrastructure after the fact. Other regions have fared better, such as South Korea, Japan, and Europe, yet at the price of resourcing and implementing broad infrastructure projects.

Will the Public Cloud be the next usage model that only increases resource consumption? The present rate of adoption, for example, by small businesses and social network games of Amazon Web Services, seems to suggest so. These are early adopters, yet the benefits of Cloud, together with these trends, are significant enough to strongly indicate only an increase in consumption.

Despite all the buzz to the contrary, compute and storage resources are finite, andnetwork resources continue to provide the principal bottleneck. Presently, dominant Public Cloud campaigns gloss over, or entirely ignore, this finite nature. The more compute, storage and network resources transform into utilities, the greater the exponentiation of consumption. Virtualization only accelerates the consumption rate, given the increase in server density. Without understanding this, increased Cloud adoption will be pursued with poor management of resources, and, resultingly, costs will increase to a point where they are uneconomical.

Resource management will grow in importance as inter-connectivity and cloud consumption increase. Although billing and metering have developed to a certain degree within Cloud service providers, this has not been approached in a manner that would truly manage the use of resources. Much has been developed to support this at the virtualization layer, yet not at the service layer. Resource management aims to address this gap, by enabling actions to be taken with workloads based upon service level agreements and policies, resource profiles and how they map to workload execution within resource pools and individual nodes.

This impacts both TCO and ROI. TCO can be maintained at a consistent level of efficacy through managing resource availability and constraints, to optimise the cost-vs-service ratio. ROI can be targeted and achieved through assurance of resource availability. Without resource management, costs can easily increase beyond efficacy, and pathfinding activities suffocated, as resources are heavily consumed and capacity exhausted.

Whilst a high-volume, low-margin strategy suits a number of business models, those looking to the Cloud to execute their mission-critical or highly-regulated processing need more than an optimistic RAS model to assure their operations. Tenant visibility into the Cloud (more on this in my next post) combined with fine-grained resource management will make this actionable.

Build it and they will come

Jevon's Paradox is the proposition that technological progress that increases the efficiency with which a resource is used tends to increase (rather than decrease) the rate of consumption of that resource. When that increase is not accompanied by any more than organic growth in resource management, it can result in a resource crunch. Simon Wardley of the CSC Leadership Forum has often pointed out that, whilst cheaper operations costs and reduced capital spending should signal a reduction in TCO, the truth is quite the opposite. This argument is rather interesting on its own, but I see the greater issue for Cloud to be resource starvation. There are several parallels which reflect that, despite observed reductions in TCO, the resource crunch has resulted in both increased costs and reduced service. 

We have witnessed this effect in the mobile space, where mobile bandwidth has been consumed at an alarming rate with the emergence of data services. AT&T's iPhone service in the US the last few years has only highlighted this, with the network brought to its knees in Manhattan, San Francisco, Los Angeles, Chicago and other major metropolitan centres. Advances in the utilisation of mobile network resources saw their near exhaustion in less than a decade. This has been accelerated and exacerbated by the common practise of planning infrastructure after the fact. Other regions have fared better, such as South Korea, Japan, and Europe, yet at the price of resourcing and implementing broad infrastructure projects. 

Will the Public Cloud be the next usage model that only increases resource consumption? The present rate of adoption, for example, by small businesses and social network games of Amazon Web Services, seems to suggest so. These are early adopters, yet the benefits of Cloud, together with these trends, are significant enough to strongly indicate only an increase in consumption.

Despite all the buzz to the contrary, compute and storage resources are finite, andnetwork resources continue to provide the principal bottleneck. Presently, dominant Public Cloud campaigns gloss over, or entirely ignore, this finite nature. The more compute, storage and network resources transform into utilities, the greater the exponentiation of consumption. Virtualization only accelerates the consumption rate, given the increase in server density. Without understanding this, increased Cloud adoption will be pursued with poor management of resources, and, resultingly, costs will increase to a point where they are uneconomical. 

Resource management will grow in importance as inter-connectivity and cloud consumption increase. Although billing and metering have developed to a certain degree within Cloud service providers, this has not been approached in a manner that would truly manage the use of resources. Much has been developed to support this at the virtualization layer, yet not at the service layer. Resource management aims to address this gap, by enabling actions to be taken with workloads based upon service level agreements and policies, resource profiles and how they map to workload execution within resource pools and individual nodes. 

This impacts both TCO and ROI. TCO can be maintained at a consistent level of efficacy through managing resource availability and constraints, to optimise the cost-vs-service ratio. ROI can be targeted and achieved through assurance of resource availability. Without resource management, costs can easily increase beyond efficacy, and pathfinding activities suffocated, as resources are heavily consumed and capacity exhausted. 

Whilst a high-volume, low-margin strategy suits a number of business models, those looking to the Cloud to execute their mission-critical or highly-regulated processing need more than an optimistic RAS model to assure their operations. Tenant visibility into the Cloud (more on this in my next post) combined with fine-grained resource management will make this actionable.

 

Cloud made easy

The world is exploding with a growing number—and variety—of computing devices. By 2015, we expect to see 15 billion devices and an additional one billion users worldwide. This growing user base will continue to demand, and expect, a wide assortment of applications and on-demand services—all with a fantastic user experience.

How will we get there? Client-aware cloud computing: Internet-based computing that provides shared resources, software applications, and information to computers and other devices on-demand—similar to your electric service.

'Client-aware’ defined 

Imagine you’re low on battery power; no problem. The client-aware application knows and has already saved your critical data. Intel Software and Services Group (SSG) pioneered the Web application programming interface (API) that adds this benefit to the cloud. 

In a client-aware environment, the cloud takes advantage of client capabilities to optimize application delivery and end-user experience across a range of devices in a secure fashion. Client-aware solutions adapt seamlessly to your devices regardless of the type of client system you are using. You’re in control. You are no longer tethered to a single computer or network. Change computers, and your applications and documents follow you through the cloud. Move to a mobile device, and your applications and documents are still available. There's no need to buy a special version of a software program for a particular device, or to save your document in a device-specific format. 

Intel Cloud Builders: The path to Intel’s ‘Cloud 2015’ vision 

With many IT organizations evolving to meet these user expectations of a wide assortment of applications and on-demand services, Intel SSG and Data Center Group (DCG) formed Intel® Cloud Builders—an industry initiative where leading systems and software solution providers provide best practices and practical guidance on how to deploy, maintain, and optimize a cloud infrastructure. The program’s goal is to provide a path to Intel’s “Cloud 2015” vision.

Intel Cloud Builders will deliver proven, detailed reference architecture solutions, and are designed to address IT challenges and the requirements of the Open Data Center Alliance. These best practices can be used now to build and enhance clouds that are more simplified, secure, and efficient.

“Engaging with the program benefits Microsoft and our mutual customers because we jointly create reference architectures so our customers can get a tried and true solution,” says Bill Laing, Microsoft corporate vice president, Server and Cloud Division. “Together, we're really about delivering more value to our customers at lower cost over time.”

“HP is a proud member of the program,” says Steven Dietch, vice president, Marketing, Cloud Infrastructure, Hewlett-Packard. “We're collaborating very tightly, particularly from an engineering perspective where ultimately the end goal is to make it easier for customers to deploy cloud solutions whether that is in a private or public domain.”

Simon Crosby, CTO, Datacenter and Cloud Division, Citrix Systems, Inc., said, “Our participation in the program will provide customers access to a stable, tested starting point for deploying cloud infrastructures built on state-of-the-art Intel® Xeon® hardware, and cloud solutions from Citrix. Citrix will build on these reference architectures to enable our customers to deliver scalable cloud services."

Taking the ‘crazy’ out of cloud design, deployment, and operation 

In short, Intel Cloud Builders is the vehicle by which Intel provides proven solutions that enable leading IT organizations and system and solutions providers across the industry to make the transition to cloud computing.

The goal with Intel Cloud Builders is to show an enterprise or service provider infrastructure IT engineer how to design and operate a cloud, and make the transition to cloud computing simpler, safer, and more cost-effective. 

In addition to addressing the challenge of getting a basic cloud up and running, our expanded scope includes usage models that show IT how to build a power-efficient cloud, a set of trusted pools, on-boarding, secure cloud access, scale out storage, unified networking, client-aware computing, and more.

The result is a set of detailed reference architectures—designed to address the evolving usage requirements of the Open Data Center Alliance, which should make for a few less “crazy” IT people in the world.

Software optimization makes Intel solutions shine 

For the datacenter, SSG works closely in optimizing many large scale datacenters and internet applications—Web apps, social media, video, etc. Examples include the work done with Facebook and software optimizations that led to a six times reduction in image processing at Alibaba—a global leader in e-commerce for small businesses—and a 50 percent increase in power density at Baidu—a Chinese search engine for websites, audio files, and images. 

 

Smart TV coming to a home near you

For almost 55 years, television has lived in our homes as the centerpiece of our living rooms. Today there are more than 2 billion TVs worldwide. People around the globe spend an average of 3-4 hours per day watching television.

Our favorite domestic appliance is deeply embedded into the fabric of who we are and what we do. It’s simple, easily understood and we do not want it to go away.

But that doesn’t mean TV isn’t changing. Over the years, color TVs replaced black and white screens. Then came Atari games on TV, cable, video cassette recorders, digital video recorders, high-definition screens, video-on-demand, 3-D—all have been part of TV’s evolution.

Today TV is at another big inflection point, with the advent of “smart TV,” but exactly what is “smart TV”?

What is smart TV?

Smart TV is much more than just Internet-connected TV, it’s a new TV experience, that lets users:

• Search online and personal content as well as broadcast programming all from their TV
• Access downloadable applications
• Connect to social networks while watching favorite programs or movies
• Control their TV with either their smartphone or unique remote controls with voice commands

The term, “smart TV” is an Intel-generated term, says Brian David Johnson, Intel Labs futurist and author of Screen Future: The Future of Entertainment, Computing, and the Devices We Love. It’s based on consumer research, our marketing push and a new breed of a consumer electronics device.

Johnson says that smart TV is not a product but rather a whole new category of TV.

High-profile partners in smart TV

Intel is off to a grand start in this brave new world, with recent high-profile announcements with partners Google, Sony, and Logitech. Logitech last month became the first company to unveil a device that uses Google TV, which is powered by the Intel Atom CE4100 processor . The Google TV platform allows users to surf the Internet and watch TV at the same time. Sony unveiled a few weeks ago its HDTV and Blu-Ray player, both equipped with Google TV and Intel’s CE4100.

Google TV melds the hundreds of channels available on pay television with the virtually unlimited content—video, games, and otherwise—available on the Internet. It runs on Google’s Android operating system and utilizes Google’s Chrome web browser. It uses picture-in-picture technology to display your search results on top of live TV, so you never need to switch between experiences. 

Analyst Rob Enderle commented on Google TV to ComputerWorld, "In the right format and with the right usage model and customer experience, people could adopt this.” He added, "Folks probably won't be doing much browsing on their TV. But consuming Internet media? Certainly. Up until now it has been too difficult for most to do that.”

What smart TV is not (or shouldn’t be)

The message from analysts, reporters, and tech gurus could not be clearer: Do not make TVs into PCs.

"Do you want to live in a world where your Tivo says, ‘I'm terribly sorry, before you can see this next show, I have to defrag myself?’” asked Intel Fellow Genevieve Bell at a recent consumer press event.

“People don't want their televisions to turn into a computer. People actually love their televisions because it turns out they're nothing like computers.”

Bell, leader of Intel Labs’ Interaction and Experience Research group, is adamant that Intel’s smart TV push will work. This go around, she says, Intel really “gets it” because smart TV is really about the TV experience and not the Internet.

“If you just have the TV emulate the PC experience, then I think that approach will fail,” says Tim Bajarin, a research analyst and president of Creative Strategies, in PCMag. “On the other hand, if you turn the various Internet sites that might work on a big screen into channels, with viewing at the heart of the experience, and deliver an experience that consumers are used to on a big screen, then the chance of success is better.”

The concept of bringing the Internet to the TV set has been a hot topic for over a decade, adds Bajarin, but only now are we “starting to see products that deliver the promise in a way that consumers may actually find interesting.”

Learning what works, and doesn’t, is one of the hard-won lessons from our experience with Viiv, which was Intel’s platform initiative similar to Centrino and vPro that provided a collection of computer technologies with a combination of Intel ingredients to support a "media PC" concept.

What did we learn from Viiv?

“We failed with Viiv,” says Johnson. “We learned a lot of things, but the most important being we tried to turn the television into a computer.”

Gary Palangian, Google TV program manager in DHG, agrees that the Viiv approach was about putting a PC in the living room, while smart TV is about putting your Internet on the TV.

He points out, for example, that Viiv was powered by PC chips, while the CE4100 was designed specifically for consumer electronics, offering home theatre quality audio/video performance, signal processing, surround sound, and 3-D graphics.

Another key factor in our better grasp of what users want from smart TV came from an Intel research project led by Bell’s user experience group in DHG called “The Social Lives of Television.”

Intel Labs anthropologist Dr. Alex Zafiroglu, Johnson, and other anthropologists and ethnographers visited hundreds of people in their homes in India, the U.K., the U.S. and China to learn how they engaged with their TVs so that Intel could better understand what consumers actually wanted.

“When we started working on the concept four years ago, we figured the number-one thing people would want in the future is movies-on-demand,” explains Johnson.

“But our focus groups revealed that what people really wanted on their TVs was Internet access. People saw the Internet as a way they could get whatever they wanted on demand. Watching what they wanted, when they wanted it, and where they wanted was a profound and liberating experience.”

These findings were invaluable when Google approached Intel looking for hardware to run Google TV.

Why smart TV is a must-win for Intel

The consumer electronics market is “an essential part of Intel’s strategy to expand beyond the traditional computer segments and it represents a growth opportunity for our company and shareholders,” explains DHG leader Brad Daniels.

Whilst Intel CEO Paul Otellini points out that the smart TV segment is just getting started, he told an external audience in October, he’s bullish on its prospects. “TV is about to change more in the next year than it has in the last 50 through the seamless integration of the microprocessor and the Internet into our viewing experience,” Paul predicted.

And success in smart TV will help us in other adjacent markets. “Smart TV represents a great new category to grow our business,” says Chief Marketing Officer Deborah Conrad. “But it’s also a fantastic opportunity to extend our brand and associate Intel with exciting, innovative devices and services.”

What’s ahead for smart TV?

In addition to Google TV, our CE4100 is also in D-Link’s new Boxee Box. We’re also beginning work with several other smart TV players for product releases in the coming months in the U.S and in EMEA later in 2011.

“People are used to seeing Intel in the PC space and not the CE space,” says Kevin Patterson, ASMO campaign manager for smart TV in SMG. “It’s a big opportunity for Intel.”

Intel enjoys a competitive advantage in terms of hardware over the next few years, says Wilfred Martis, manager of retail consumer electronics in the Digital Home Group (DHG), and we have a “whole lot” of OEMs lined up to bring Google TV to the market early next year. But the competition isn’t standing still. As in the smartphone world, ARM-based SOCs will gain traction, and companies like Samsung are busy designing CE chips.

“In this smart TV space, we are ahead, the first to come to market,” points out Martis, “It’s ours to lose.”

“I don’t sleep anymore, that’s how exciting this is,” he adds. “It’s been a long time coming. We can now walk into stores and say, ‘We made that happen.’”

 

My interview with MSDN on Cloud

I was recently interviewed by Robert Duffner of MSDN on Cloud Computing for the Windows Azure team - you can read it in full at http://bit.ly/9jMHtC #cloudcomputing

Cloud computing to help foster a greener planet?

A recent Microsoft-commissioned study - focussed on Microsoft business applications - bears out what environmentally-conscious companies have hoped for all along: cloud computing has the potential to reduce energy consumption and carbon emissions by 30 percent or more. 

According to the study, conducted by Accenture and WSP Environment and Energy, large data centres, such as those run by tech giants Microsoft and Google, do benefit from economies of scale and operational efficiencies; small businesses, of about 100 users, moving business applications away from on-site servers into the cloud can see net energy and carbon savings of more than 90 percent; and mid-sized organisations, of about 1,000 users, the could see savings were between 60 to 90 percent. 

“The cloud has the ability to deliver business value for customers in an age where corporate responsibility is critical to business success," said Rob Bernard, chief environmental strategist at Microsoft. 

Larger operations are generally more efficient managing server utilisation and capacity, the researchers said, noting that they can increase the server capacity for an application during peak periods and decrease when not in use. Large public cloud environments are better at dynamic provisioning and can serve “millions of users” across many companies simultaneously on one “massive” shared infrastructure, according to the study. 

The researchers concluded that while independent organizations can incorporate dynamic provisioning, server utilization and other strategies into their own data centre, the benefits will not be as great as it would be for a large cloud provider. 

The study focused on three commonly used Microsoft business applications: Microsoft Exchange Server 2007 for e-mail, Microsoft SharePoint Server 2007 for content sharing, and Microsoft Dynamics CRM for the customer relationship management platform. Each on-premise product was compared with cloud-based equivalents: Microsoft Exchange Online, Microsoft SharePoint Online, and Microsoft Dynamics CRM Online.

Researchers calculated the carbon footprint of server, networking and storage infrastructure for three different deployment sizes, of 100 users, 1,000 users, and 10,000 users, according to the writers. The study relied on actual data provided by Microsoft to estimate cloud-based energy use and carbon footprint. 

Choosing the cloud option can enable a significant reduction in carbon emissions, and the smaller the organization, the larger the benefits, said the study. In fact, the research team calculated that 32 percent of emissions could be saved by moving 50,000 e-mail users in North America and Europe off individual Exchange servers and on to Microsoft Exchange Online. 

Companies are generally considering the move to cloud computing to reduce costs and increase business flexibility. Along with hoped for reductions in the electric bill, companies are interested in minimizing the number of physical infrastructure, such as servers and switches, needed. Companies can also outsource some of the performance and management tasks.

 "The benefits of cloud computing are clear: Increased productivity, reduced costs and lower management overhead of products have become key considerations for organizations evaluating Microsoft's cloud services,” said Bernard. 

The study focused on select Microsoft applications, but the study authors felt it was “likely” that “similar advantages” would be seen for other business applications and cloud providers. 

 

 

MeeGo v1.1 released

The MeeGo open source project hosted by the Linux Foundation, and supported by Intel and Nokia, announced the project release for MeeGo v1.1. MeeGo provides a solid baseline for device vendors and developers to create software for a range of devices such as netbooks, tablets, handsets, media phones, smart TVs and in-vehicle infotainment (IVI) based on the Intel® Atom™ processor. Optimized for the Intel Atom processor, MeeGo delivers an exciting new internet experience on these devices. MeeGo was designed as a common software platform across various devices with adaptable, yet consistent, user experiences. SSG believes open source is the best method to engage the broad developer community and advance innovation in these new device categories. With MeeGo, companies can deliver a differentiated user experience that reaches across device categories and increases the opportunity for OEMs, application developers and service providers to reach more customers. The MeeGo v1.1 release includes: Core Operating System (OS) v1.1 – consolidated common base operating system for all UXs Netbook User Experience (UX) v1.1 - complete set of core applications for netbooks In-Vehicle Infotainment (IVI) UX v1.1 – includes a sample IVI home screen and taskbar built with Qt 4.7 and speech recognition Handset UX v1.1 - technology snapshot implements basic development UX for voice calling, SMS messaging, web browsing, music and video playback, photo viewing, and connection management SDK v1.1 Beta - SDK for MeeGo Core OS and supported UXs will be released later this month. MeeGo code is expected to be released to the open source community twice a year. For more information please visit www.meego.com or www.meegozone.com.