Monday, February 02, 2009

Azure and Cloud Computing Posts for 1/29/2009+

Windows Azure, Azure Data Services, SQL Data Services and related cloud computing topics now appear in this weekly series.

Note: This post is updated daily or more frequently, depending on the availability of new articles.

• Update 2/1/2009 10:00 AM PST: Additions

Azure Blob, Table and Queue Services

My “Retire Your Data Center” Article about Azure’s Table Storage is VSM’s February 2009 Cover Story of 1/31/2009 includes a copy of the deck:

Visual Studio 2008, ASP.NET, and the Azure Services Platform combine to simplify local development of data-intensive Web apps and automate their deployment in Microsoft data centers. The result: You get maximized availability and reliability with almost limitless on-demand scalability, while you pay only for resources consumed.

My “Visual Studio Magazine” Merges Print Operations with “Redmond Developer News” article reports that RDM is merging with VSM; Michael Desmond will replace Patrick Meader as VSM’s editor in chief.

Abel Avram’s Presentation: REST: A Pragmatic Introduction to the Web's Architecture analyzes Stefan Tilkov’s A Pragmatic Introduction to REST session at QCon London. According to Abel:

[Tilkov] thinks that REST is not an alternative to SOA but it can serve SOA to reach its goals. Stefan also covers other related topics: HTTP, WS-*, SOAP, CORBA, RPC, enterprise, in an attempt to make the listeners understand what REST is and what is not and how it helps.

Mike Amundsen recounts in his byte-order marks are evil, dangit! post of 1/29/2009 how he was bitten by recently-added byte-order marks in XML blobs. Azure Blob Services now “support multiple encoding types for returned Entity XML,” which depends on byte-order marks (BOM) at the beginning of XML documents. (Not all XML documents are encoded as UTF-8.) Fortunately, the work-around was easy, as Mike demonstrates in his sample code.

The RESOLVED: invalid characters returned from SDS queries tonight thread in the SQL Data Services (SDS) - Getting Started forum has more details on Mike’s BOM issue.

Alex Chitu’s First Official Description of GDrive post of 1/30/2009 describes the product category of GDrive as ‘Online file backup and storage’ and descriptions: "GDrive provides reliable storage for all of your files, including photos, music and documents;” and “GDrive allows you to access your files from anywhere, anytime, and from any device – be it from your desktop; web browser or cellular phone;”. The quoted text is from a JavaScript file included with with Google Pack, but doesn’t appear to be truly “official” to me.

Update 2/2/2009 10:00 AM PST: Om Malik chimes in with a Why Google Needs the GDrive to Fight Microsoft post of the same date that warns against a “Google monoculture”:

I believe Google is looking to build something unique, a service that it would position as a direct competitor to not only Microsoft’s SkyDrive and Live Mesh services, but to the software giant’s SharePoint services. My guess would be that they would marry GDrive storage with Google Apps and other applications, such as Google Talk. In doing so they’d create a virtual “computing environment” in the cloud. …

Paint me cynical –- though I like to think of myself as realistic –- but I don’t think Google’s backing of President Obama and his campaign was done with purely altruistic intentions. Given how close the company’s management is with the government officials, I worry that Google will one day go too far — and get away with it.

SQL Data Services (SDS)

Jason Massie’s The Death of JOIN post of 1/31/2009 explains (with the help of CarpDeus’ The End of of JOINs? and The End of JOINs? Part 2, as well as AdaptiveBlue’s How and Why Glue is Using Amazon SimpleDB instead of a Relational Database) why relational databases don’t scale and how JOINs differ for SQL relations and the entity-attribute-value tables used by SDS and SimpleDB.

Jamie Thomson recounts creating an SDS demo app with Ryan Dunn’s SSDS REST Library (a .Net library intended to make it easy to interact with SDS-hosted data) in his Open Space Coding – Using the SSDS REST Library

• Jason Lee’s SQL for Cloud Computing: Microsoft SQL Data Services MSDN Spotlight article on carries this deck:

Developers face a new set of challenges and opportunities managing data for cloud computing. This architectural overview of Microsoft's SQL Data Services (SDS) highlights the benefits for the enterprise and shows how you can use SDS to augment your existing on-premises data infrastructure.

Mike Amundsen recounts in his byte-order marks are evil, dangit! post of 1/29/2009 how he was bitten by recently-added byte-order marks in XML blobs. Azure Blob Services now “support multiple encoding types for returned Entity XML,” which depends on byte-order marks (BOM) at the beginning of XML documents. (Not all XML documents are encoded as UTF-8.) Fortunately, the work-around was easy, as Mike demonstrates in his sample code.

[Copied from the Azure Blob, Table and Queue Services topic.]

.NET Services: Access Control, Service Bus and Workflow

Stephen Ashley’s The digital cloud needs trust post of 1/29/2009 discusses the need for cloud computing purveyors to gain the trust of potential users. Stephen writes:

Trust has to be earned before people can get over the psychological barrier of letting go of their own data. I have seen this with both large companies and individuals. People argue about security etc. even when it has been quite easy to demonstrate that their data is likely to be safer in the cloud.  Trust will only be built as people begin to understand the concepts around cloud computing, the technologies involved and the benefits they bring. I see that for most organisations it will be a progression, with plenty of lessons learnt along the way.

Elinor MillsCloud computing security forecast: Clear skies ZDNet blog post of 1/27/2009 quotes Yankee Group senior analyst Phil Hochmuth, a senior analyst at Yankee Group, on the topic of whether “the data and collocation centers that corporations contract with to host their data?”

It does come down to vetting the practices of the provider and making sure they meet the standards you want for your business.

on the evening before the opening of Cloud Computing Innovation Day in Santa Clara, Calif. Mills quotes other IT executives on the relative security of corporate data in the cloud versus on-premises hardware and goes on to write:

The security concern with cloud computing is a cultural issue, said Rebecca Wettemann, a vice president at Nucleus Research.

"The question is would I rather be at a huge data center where a vendor is contractually required to keep my data secure or would I rather rely on my staff to do it properly?" Wettemann said. "You need to trust that your vendor will manage your data."

So far, there haven't been any significant security breaches with an on-demand services vendor, she said. And people are getting used to the idea of being able to access their data anytime and from anywhere because it is out on the Internet, she added.

Live Windows Azure Apps, Tools and Test Harnesses

Shan McArthur’s CMS hosting in Azure thread in the Windows Azure forum describes how ADXStudio’s product team ported the firm’s Web CMS product to run on the Azure Services Platform. Shan concludes:

It was a lot of work, but I am confident that it was worth it.  Going through this effort has made our product better by supporting more deployment environments, running in higher security environments, and being more extensible for the future by using providers.  Having the first-to-market advantage of being the first commercial CMS to run in Azure is probably going to be beneficial as well. …

We are preparing for an open beta program now. Our plan is to launch a new community site prior to opening the beta. If you are interested in getting in on the beta, please fill in the contact us form on the ADXStudio website.

Markus KlemsCloud Mashups post of 1/30/2009 postulates that Platform Mashups are a

[N]ew type of mashup that we can see today[, which] combines Cloud Computing services and integrates them into a single service or application. Amazon’s GrepTheWeb is a good example for Cloud Computing service compositions within the domain of a single provider [5]. However, the recent announcement of Appirio’s ReferMyFriends App shows that also cross-Cloud mashups are viable [6]. Other examples for cross-Cloud mashups are Facebook + EC2 back-end [7] and + AppEngine back-end [8] (although it is probably only a matter of time until [8] will become one of the single-domain examples).

His early Classification of Cloud Computing Stakeholders post and its updated diagrams provide a useful ontology for cloud computing.

Scott Hanselman recommends RTFLF - Read the Expletive Log File when debugging an ADO.NET Data Service running on a heavy-duty staging server in this 1/29/2009 blog post. The same advice applies to Azure-based Web projects running in the Azure Fabric, because log files provide the only debugging capability.

This means, of course, that the Azure team needs to make access to log files a quicker and easier process.

J Sawyer’s Presentation content posted! of 1/28/2009 announces uploading of his Building your First Cloud Service presentation that he delivered to Texas’s Aggieland .NET User Group and the North Houston .NET User Group. (J’s a Microsoft developer evangelist.)

The presentation’s 20-MB *.zip file includes the PowerPoint slides and full source code for the simple "Hello, World” introductory demo and the full-featured “Blog” demo.

Azure Infrastructure

Hoi Vo’s Design Principles Behind The Windows Azure Hypervisor post of 1/29/2009 discusses the differences between the Windows Azure Hypervisor and that for Windows Server 2008. Hoi, who’s a director on the Azure team, concludes:

Much of the development for the Windows Azure Hypervisor would only work in our environment, taking advantage of our specific homogenous data center environment. Some of the innovations would be useful to customers with a different data center design and will be incorporated in future releases of Hyper-V (e.g. Second-Level Address Translation will be available in Hyper-V v2.0).

This isn’t what prospective Windows Azure users wanted to hear.

Steven Martin analyzes the the last few weeks and the road ahead for the 6th annual SOA and BP conference in his Microsoft's Annual SOA and BP conference - SOA meets Cloud post of 1/29/2009. Steve writes:

As we talk to customers around the world, it’s clear that the message on SOA being a “how” rather than a “what” is really sinking in. Now we spend more of our time talking about extending Service Orientation than we do defining it. I took a very non-Microsoft approach – instead of blathering on about technology, I instead spent time debunking the myths (which still pervade the space) and connecting customers with their peers to share what they have learned and best practices.

One thing that was clear from talking to customers this week is that the SOA landscape continues to evolve. As I wrote months ago – the convergence of key technologies like Virtualization, Cloud, Modeling, SOA and SaaS continues at a rapid pace. It’s clear that cloud computing will have a significant impact on SOA moving forward. So much so that one might argue that sooner rather than later, the app pattern which is Service Orientation will be a given - much the same way that no one refers to Object Orientation as a "feature" of an app anymore. Service Orientation will be the flour of the composite app cake.

Tom Steinert-Threlkeld carries Steve’s point forward in his Tipping Point Reached For Cloud Computing? ZDNet post of 1/30/2009. Tom’s post starts with a chart from Saugatuck Technology that estimates worldwide SaaS solution demand for 2008 and 2010 as the percentage of companies adopting SaaS:

Tom goes on to ask: Has the tipping point come for software to be widely adopted as an online service? and then answers his rhetorical question:

William S. McNee, president and CEO of Saugatuck Technology in Westport, CT, believes it has.

He says that mainstream adoption within small and medium-sized businesses is “accelerating” – and that 20 percent of enterprise IT workloads will be run “in the cloud” by 2013.

This will lower operating costs, reduce IT staffs and cut down on carbon footprints, he says.

Speaking recently at a forum sponsored by a software-as-a-service purveyor, NetSuite, held at the New York Stock Exchange, McNee said, “On-premise solutions are going to drop off the cliff.”

Basing his comments on a survey by Saugatuck of 150 chief financial officers with budget authority, he said the movement toward what increasingly is being called “cloud computing” will put not just payroll and other administrative processes on remote servers and software, but “mission-critical computing,” as well.

John Foley’s The Oregon Trail Of Cloud Computing post of 1/29/2009 for InformationWeek contends that:

Cloud computing is like the Wild West, where the players are rough around the edges, the borders are undefined, and the homesteaders are subject to unforeseen risks. In this environment, IT governance is nearly impossible -- but an absolute requirement.

and emphasizes the need for improved governance:

InformationWeek's Mike Fratto provides an eye-opening look into the state of governance in his article "Cloud Control," which appeared in the Jan. 26 issue of InformationWeek and is posted on Fratto talked to a handful of IT pros who are working through the issues of cloud computing governance. His sobering conclusion: "The courts and industry groups will eventually help develop guidelines, but for now, we're on our own."

Foley concludes:

Amazon remains unacceptably opaque. The company refuses to reveal the locations of its data centers. "You can't audit what you can't see," writes Fratto. "This is a deal killer in many regulated industries."

Neil McAllister argues The case against Web apps in his 1/29/2009 InfoWorld post subtitled “Five reasons why Web-based development might not be the best choice for your enterprise.” However, Neil compares the benefits of conventional GUIs with the limited UI repertoire of browser-based projects.

Neil says “Web UIs are a mess.” I agree, but ordinary users had better become accustomed to clumsy HTML data entry controls in the world of “Everything as a Service.”

Other Cloud Computing Platforms and Services

The Information Technology Laboratory of the U.S. National Institute of Standards and Technology (NIST) is contemplating the identification of minimal standards and architecture to enable federal agencies to create or purchase interoperable cloud computing capabilities. NIST’s Perspectives on Cloud Computing and Standards presentation lists the following characteristics of a future “Federal Cloud Infrastructure:”

  • Agencies would own cloud instances or ‘nodes’
  • Nodes would provide the same software framework for running cloud applications
  • Nodes would participate in the Federal cloud infrastructure
  • Federal infrastructure would promote and adopt cloud architecture standards (non-proprietary)
  • ‘Minimal standards’ refers to the need to ensure node interoperability and application portability without inhibiting innovation and adoption thus limiting the scale of cloud deployments

NIST promises to create the following “Special Publications” in 2009:

  • Overview of cloud computing
  • Cloud computing security issues
  • Securing cloud architectures
  • Securing cloud applications
  • Enabling and performing forensics in the cloud
  • Centralizing security monitoring in a cloud architecture
  • Obtaining security from 3rd party cloud architectures through service level agreements
  • Security compliance frameworks and cloud computing (e.g., HIPAA, FISMA, SOX)

You can expect to find draft versions of these publications from a link to the Special Publications (800 Series) on the Computer Security Division’s Computer Security Resource Center Publications page.

It will be interesting to see NIST’s approach to HIPAA, FISMA and SOX compliance with cloud computing.

• Kas Thomas mulls Thoughts on Google Monoculture and the Cloud arising from Google’s "This site may harm your computer" deluge of 1/31/2009. Kas notes that it took Google an hour to return its search services to normal after detecting the human-induced error and reminds us that “Becoming dependent on a single supplier is bad ...” and “graceful failover is good,” which applies to cloud computing as well as Internet search.

• David Linthicum’s The Harsh Realities of Private Clouds podcast of 1/30/2009 “talks about the lack of good private cloud computing technology, and the forthcoming vendor battle over the emerging private cloud market.”

The podcast relates to Dmitry Sotnikov’s No real private clouds yet? post of 1/16/2009, which posits that private clouds are important because:

  1. Some data and systems belong on premise due to security and legal concerns, or simply companies wanting to stay in control.
  2. The story of not having to provision and manage hardware resources when using public clouds from Amazon and others sounds great but the reality is that enterprises already have hardware of their own. Instead of tearing down datacenters in which they have been investing all these years - why not start to use them more efficiently?

but Dmitry argues “that no real solution is currently obvious on the market.”

Links to all of Dave’s 30+ cloud computing podcasts are here.

• Mr. Blog’s Google App Engine teases, but ultimately doesn’t deliver post of 1/30/2009 rants:

However, Python is not the worst of GAE’s problems, not by a long shot. In the name of executive summary brevity, I’ll jump right in.

  1. “GQL” and the Datastore - I don’t know who Google is trying to kid here by hinting that the “datastore” is somehow like “SQL”. God help you, if you fall into that trap. It must drive any real RDBMS expert mad.
  2. urlfetch - only works on port 80 or 443 and times out after a few seconds.
  3. mail.send() - can’t send to more than a few addresses at a time, whether invoked once for all recipients, or using multiple invo[c]ations.

All the above should be prefaced in the GAE documentation, in huge red letters: DO NOT EXPECT THIS TO WORK because they don’t, at least not in practice to do anything real. [Author’s emphasis.]

Re item #1, see this blog’s SQL Data Services (SDS) section above.

Geva Perry’s Links for 2009-01-28 [] post of 1/29/2009 points to a 1/25/2009 post by John Gerber that quotes Gartner analysts Daryl Plummer and Thomas Bittman at the Gartner Emerging Technologies conference in Las Vegas:

By 2012, 80 percent of Fortune 1000 companies will pay for some cloud computing service, and 30 percent of them will pay for cloud computing infrastructure.” Plummer defines cloud computing as “a style of computing where massively scalable IT-related capabilities are provided ‘as a service’ across the Internet to multiple external customers.”

Gerber goes on to write:

Forrester analyst James Staten interviewed more than 30 companies and concluded that cloud computing has been “wildly popular” with small businesses but large companies have been skeptical. Forrester has posted the report, “Is Cloud Computing Ready For The Enterprise?” Staten blogged about his report in the post “Are Fabrics Web 3.0?.” Larry Dignan sums up some of the notable benefits in his post, “Cloud computing hasn’t gone Fortune 500 yet, but it’s coming” …

Krishnan Subramanian recommends SaaS Risk Reduction – Don’t Keep All Eggs in One Basket – A Case Study in his 1/29/2009 post in the sixth member of his SaaS Risk Reduction series. Krishnan recounts the domain-related problems “a frustrated SaaS user described his/her ordeal with Google Apps for Domains Free edition.” Krishnan writes:

Let us now see how this user could have avoided this scenario if he/she had followed one of my first suggestions in this series. Don’t keep all eggs in one basket. In that post, I had argued against using a single vendor for all your SaaS needs. Some users didn’t like the idea because they thought keeping email with one vendor and calendar with another is too much of a hassle. I do agree with them about the hassle but paranoid ones are better served by spreading the apps/data with different vendors.

The question arises: Do different vendors solve the problems that might be expected to require multiple, redundant vendors? I doubt it.

Maureen O’Gara’s Intel Said To Be Throwing SaaS Blanket Across Europe post of 1/29/2009 quotes from the Seeking Alpha blog’s Intel Builds SaaS Cloud across Europe post of the same date by Joe Panettieri who writes:

A small Canadian software company is helping Intel Corp. (INTC) to launch a software as a service (SaaS) platform that will soon blanket Europe, Intel confirmed to me earlier today. The initiative could give Intel much-needed SaaS experience and will likely drive demand for Intel's vPro hardware across Europe.

The European SaaS initiative involves an emerging Intel platform called Multi-Site Director. The SaaS platform is built upon so-called managed services software from N-able, a privately held but fast-growing software company that I track on