Tony Parham's Blog

Cloud computing continues its inexorable march towards increased mainstream usage.  Today’s Wall Street Journal described the competition between Google and Microsoft to become the cloud e-mail provider for the US General Services Administration (GSA), the agency which oversees government procurement and which manages federal property.  This migration of the GSA’s 15,000 employee e-mail accounts (currently on Lotus Notes) will influence both private sector and public sector computing.

One does not typically think of most governmental agencies (excluding the research/scientific laboratories) as early adopters of information technology. So GSA’s position may point out how widely embraced cloud computing concepts have become.  Indeed, (as pointed out in one of my previous posts), the GSA already has a dedicated website (apps.gov) which sings the merits of cloud-based computing.

This is just one of many initiatives which have sprung up during the tenure of Vivek Kundra, the United States’ first Chief Information Officer.  These US CIO initiatives (such as a federal IT dashboard) have been launched to  address Kundra’s 5 stated priorities: (1) Open and transparent government, (2) Lowering the cost of government, (3) Cyber-security, (4)Participatory democracy and (5) Innovation.  Indeed, many other governments around the globe are moving rapidly to embrace the benefits of these new information technologies (as shown in the G-Cloud initiative detailed in the Digital Britain Report).

Cloud computing still has challenges to address.  For example, security is always a key concern for externally-hosted computing and data.  “Wins” on this front include the GSA’s certification that Gmail and Google Apps meet GSA security requirements.  (Microsoft is apparently close to receiving GSA security certification for its web-based version of Exchange.)  “Setbacks” include last week’s missed deadline by Google re: security concerns for the email migration of Los Angeles’ 34,000 employees.

Setbacks notwithstanding, cloud computing will continue to become more mainstream in business and governmental usage around the globe.

Additional TKG contact info:

Web:  www.TKGweb.com,  Twitter: @tonyparham

More than 800 professionals attended yesterday’s 2010 MIT Sloan CIO Symposium.

Some notable quotables from throughout the day:

• Batting .500 with innovative ideas is hitting the ball out of the park.  Batting .200 isn’t bad.   Fail fast. (Roy Rosin, Intuit)
• With stability at the center of your systems, you can invite “instability.”   (Jeanne Ross, MIT CISR)
• Enterprise 2.0 collaboration can become its own silo if not integrated into other systems (Greg Hansen, AMD)
• Keys to data interoperability: (1) data availability, (2) ease of data access, (3) relationships to other data.  (Dr. Ed Curry, DERI)
• Standardization of technology + streamlining/simplifying organization structure, led to innovation and strategic vision.  (Anne Margulies, CIO, Commonwealth of MA)
• Are your organization’s decisions based on data? Or are they based on HiPPO (the Highest Paid Person’s Opinion)?

The Internet of Things (IOT) panel was perhaps the most interesting session.  Panelists included: Dr. Michael Chui (McKinsey Global Institute), Robert LeFort (Ember), Sanjay Sarma (MIT Auto-ID Center), Bob Metcalfe (Polaris), Mark Roberti (RFID Journal).  Some interesting factoids:

• Ember:  Key applications are usually an intersection of  “saving money” and “saving the planet.”
• Ember:   Just shipped its 10 millionth node, and recently became profitable after multiple years.  Took them seven years to discover that smart readers will likely be the killer app for ZigBee.  But putting intelligence into other household objects will likely be an even bigger opportunity.
• IOT will allow us to see, track and manage things which we cannot see, track and manage today.
• StickyBits allows every physical object to have its own website, and allows people to add comments regarding that object.
• Infinite Power Solutions creates those tiny batteries which may be necessary to power IOT tags.
• GPS, RFID and Barcodes (and other tagging technologies) will all coexist and each be useful/applicable for different applications.
• There is always an organizational challenge for the deployment of this (or any other) technology.  Existing players/workers may be threatened, and resist the new processes.
• The most important thing to do with RFID data is to throw it away.  It’s value has a very short shelflife.  Use it or lose it.  Only need to act on the exceptions.
• Adding RFID metadata to things like video will yield new value.
• It was mentioned (unverified) that one retailer (Harrods?) already allows a consumer to type in an ID for a purchased meat product, and see the entire value chain history of how that product was produced, back to an individual cow.  (Another example might allow a consumer to track the history of a bottle of wine, and verify that the temperature of the wine bottle stayed within a specific temperature range for its entire lifespan.)

The cloud computing panel had a number of interesting insights:

• Even Washington is  on the cloud bandwagon:  See the GSA’s apps.gov site.
• Yahoo has 350,000 – 450,000  servers in its cloud.
• SalesForce.com needs only 2,000 servers to support its 72,500 customers (which nets out to about 36 customers per server).
• InvoiceCloud is an example of a service which should be useful to firms in all industries
• In comparing the real costs of  SaaS to other solutions, it should be remembered that you cannot depreciate SaaS expenses.
• 50%  of the transactions that come into salesforce.com come via API calls  rather than from the salesforce user interface.
• For some applications, you need to be aware of where (i.e. what country) your cloud data is stored, since, in some countries, your data may be subject to search and seizure.  (Some vendors address this issue by ensuring that the data primarily resides in one specified country, and is then distributed as encrypted data when access is needed.)
• To address the organizational transformation issues which may be encountered when deploying cloud solutions: start with small pilots, advertise the new roles which will be created.  (By moving to the cloud, the business can change its focus from the underlying technologies to the actual business processes.)
• Line of business managers, not CIOs, are often the catalyst to get a company/application onto the cloud.  (For example, Gartner’s Darryl Plummer indicated that, of 50 CIOs he interviewed, only 2 knew that an application from their firm was going into salesforce.com, prior to its appearing in salesforce.com.)
• Cloud is not a sku you are going to buy, but a state you are trying to achieve (Sanjay Mirchandani, EMC)

The e-health panel had some interesting discussions regarding the application of information technology in the health sector:

• It’s difficult to balance privacy concerns with the benefits that can be realized by leveraging data.   (William Fandrich, Blue Cross)
• Mention was made of the Federal information security management act
• Discussion regarding management of personal health records, including discussion of individuals’ roles / rights; and the difficulty of transporting patient records between health institutions (such as when a doctor moves to a new hospital) even when patient consent is available.
• In addition to issues of ACCESS to records, it may also be a challenge to ensure that the data is UNDERSTANDABLE.  This may require a sharing of semantic meaning to avoid drowning consumers in meaningless data.
• Processes must be put in place to allow (and enforce) chart access control: must declare (and verify) reader’s  relationship to patient.

More info about the conference can be found at the MIT CIO website, twitter stream and LinkedIn group.

Additional TKG contact info:

Web:  www.TKGweb.com,  Twitter: @tonyparham

I recently attended a webinar given by several cloud computing vendors and thought I might jot down a few observations.

As described in Simon Wardley’s clever OSCON 2009 keynote, “Cloud computing is a generic term to describe the disruptive transformation in IT towards a service-based economy, driven by a set of economic, cultural and technological conditions.”  (An NIST definition provides more details.)

Cloud computing services are becoming more widely adopted in the marketplace.   Indeed, a few charts from the webinar graphically portray the scalability-cost benefits of clouds vs.  on-premises.  (The charts also include “quick polls” of the webinar audience regarding their anticipated: cloud use, data barriers and application barriers.)

Amazon’s Elastic Compute Cloud  (Amazon EC2) continues to be  a market leader, however there are numerous vendors vying for attention.  Vendors typically focus on a specific layer of the cloud “as a service” (“_aaS”):

• Software (SaaS) –  User-facing-applications, business logic:  Examples: Constant Contact, Salesforce.com, NetSuite, Bullhorn, etc.
• Platform (PaaS) – API’s,  pre-built containers, etc.:  Examples: Amazon EC2,  Google app engine, Force.com, Intuit, etc.
• Infrastructure (IaaS) – Hardware, virtualization, storage/backup:  Examples: RackSpace, EMC, Unisys, Zmanda, VMware etc.

Judith Hurwitz further segments the PaaS layer as:

• Lifecycle –  Examples: Amazon EC2, Microsoft Azure, Google app engine, etc.
• Anchored – (Initially anchored by a specific vendor’s user-facing-application) Examples: Intuit, Force.com, etc.
• Enabling technology –  (Such as management) Example: RightScale, etc.

Services such as Rightscale, focus on management of application deployment to clouds in a manner which decreases effort/cost, and which also prevents lock-in to a particular cloud vendor’s proprietary formats.  Indeed, there is a broad effort afoot (sometimes called the intercloud) to increase interoperability among cloud vendors.

Some publications (such as this McKinsey report) have questioned cloud computing’s proposed economic value.  However, many others (such as  TechCrunch, CIO.com and  RightScale) have countered with pro-cloud perspectives.  One of McKinsey’s themes asserted that cloud computing was vastly over-hyped.   However, such hype is typical of the lifecycle of any new technology, as depicted by the “Gartner Hype Cycle.”   Indeed, Gartner’s 2008 hype cycle depicted cloud computing as rapidly heading towards the “peak of inflated expectations.”   That dubious distinction notwithstanding, cloud computing is here to stay.

Not surprisingly, when I attended the 2009 MIT Sloan CIO Symposium, cloud computing was a key topic.  Irving Wladawsky-Berger (a brilliant ex-IBM’er and conference panelist with whom I chatted briefly) re-capped some of the cloud discussions in his blog entry.   He confirms many of the cloud benefits (virtualization, shared infrastructures, highly disciplined systems management, flexible deployment and scalability) but posits that “Pretty much everyone agreed that cloud computing should not be used for mission critical applications, or applications involving sensitive information.”

Au contraire! Although several of Irving’s co-panelists gave those depictions, quite a few of the forward-thinkers in the audience felt that the panelists may have missed the mark on those comments.    Nonetheless, the panelists agreed that bursty applications and new projects will be attractive candidates for the cloud.  (Indeed, many of us know of numerous new applications whose business are extremely reliant on the cloud, such as  Animoto,  Sonian, ShareThis and many others.)

In addition to affecting IT/business decision-makers, cloud computing has (obvious) implications for end-user-application vendors, many of whom are re-thinking their go-to-market strategies.  These on-premise software vendors are evaluating if (or when) their portfolio should move to SaaS/cloud models.   Although the migration may not be easy (as evidenced by the SaaS migration checklist from Ariba’s CEO), some market changes are undeniable.  (Indeed, some say 2009 is the tipping point re: SaaS migrations.)

Sun Microsystems co-founder Andy Bechtolsheim lists several key benefits for  application vendors to embrace cloud computing:

• Easier to reach new customers
• Lowest cost way of delivering and supporting applications
• Ability to use commodity server and storage hardware
• Ability to drive down data center operational cots

Peter Cohen, et. al, indicate additional cloud / SaaS benefits:

• Greater flexibility to meet fluctuating demand (particularly for applications that have high peaks in usage followed by relative quiet)
• Better access to the application for remote workers
• Lower risk of a failed or delayed deployment
• Instant access to the latest product enhancements, and assurance that all users are on the same version
• Faster rollout

Of course, cloud computing is also a hot topic on college campuses, ranging from Stanford to MIT to small colleges like Marion College.

At the 2009 Cloud Camp Boston un-conference, some attendees suggested that backup and data storage were the bulk of today’s cloud services spending. However,  IDC projected that applications will be 52% of cloud spending by 2012.  IDC estimates that worldwide spending on cloud services was $16B in 2008, surging to $42B in 2012; growing at six times the rate of the “traditional” IT market.  Gartner pegs worldwide cloud services revenue at $46B in 2008, $56B in 2009, and $150B in 2013.

Cloud computing will continue to grow, benefitting users, IT organizations and application vendors.

Additional contact info:
TKG consultants: www.TKGweb.com/bios Twitter: @tonyparham