Archive for the ‘SOA’ Category

The Power of Default in the Land of Endless Possibilites (A Catch-22 Stewpod)

February 27, 2006

itcLogo.gif

This one is a “Catch-22 Stewpod”. Stewpod – because I gathered here, under a single theme, two unrelated podcasts; Catch-22 – well, you will have to keep on reading.

If you take any subject in life you would normally find yourself with many possibilities to pick from. Here are some technological examples: RSS feeds? – Feedburner‘s official numbers are ~250K; Blogs? Technorati‘s official is 27M; Linux Distros? Distrowatch‘s is 370~ ; Thinking about implementing SOA? Well, think again – there are so many different interpretations to what’s SOA; Forced to implement SOA with IBM? Hmm, there are 13 different products under the IBM’ SOA umbrella; which one do you need, or maybe do you need them all?

The point is clear – there are always too many possibilities, options and choices. But unlike the conceptual illusion that the more options one has, the more “free” (as in freedom), rich and significant his/her life would be, the lecture of Prof. Barry Schwartz – “Less is More”, demonstrates that the opposite is actually true: endless possibilities mean endless angst. Too many possibilities could well lead to paralyzation, stress and despair. Constraining the number of alternatives is, therefore, an existential necessity. Barry Schwartz suggests having a predefined default whenever possible.

Whoever used any RSS Reader knows they are coming with a predefined subscription on some selected, highly popular feeds. That’s a default. In the Enterprise world, the default’s equivalent might be a Technical Reference Model (or Manual) listing all the vendors, products and technologies authorized for use; or, as it’s usually the case, IBM is the de-facto default for practically anything.

The “Default” as a mean to regain happiness in a re-constrained world raises some evident moral questions (i.e. socio-economic and political :)), such as who’s responsible for setting the default value, for the “power of default” and its economic gains are invaluable.

Have a look, for instance, at the following Sys-Con SOA Awards. There are 21 categories related to SOA, and IBM, though never a winner, is a runner-up in 20 out of 21 categories! I would argue that this is an amazing demonstration of the power of default.

Another well known demonstration of the Default’s Power is Microsoft, with IE, Media Player, shipped as part of their operating system.

So who’s setting the default? In progressive Enterprises I assume that the Enterprise Architecture Office is in charge of the defaults. But in most Enterprises, it’s undoubtedly fear that determines which vendor gets this invaluable label. And here IBM plays, again, a major role, for “No one gets fired for buying IBM”. This axiom is a late adaptation of Keynes‘ famous remark: “It is better to fail conventionally, than to succeed unconventionally” (as quoted by James Surowieki, the other half of this stewpod). Yet, conventional failures might lead to disastrous outcomes (and see Tristan Yates: How IBM Conned Our IT Execs Out Of Millions, for an example).

Unlike the Enterprise, in the Web2.0 sphere there’s no fear, and out there the default is determined by the people’s attention. Technorati Rank, del.icio.us’ popular, digg etc. are all manifestations of the public attention. The public elects its defaults.

This leads us to the second lecture by James Surowieki – “Independent Individuals and Wise Crowds”. Humans are not ants, maintain Surowieki, and “the wisdom of crowds” is nothing but a myth. Too often the crowd is wrong and less intelligent than a single individual. Surowieki gives many amusing and interesting examples, as well as one practical advice on how to increase one’s independence in the face of the Web2.0 herd: keep your (social, intellectual, commercial, computational) ties loose. In other words – increase your options, possibilities and choices! “Inject randomness”, concludes Surowieki (Oy! Do I hear Grazing Lists?).

References:

On the Power of Default

Linked: The New Science of Networks, Albert-Laszlo Barabasi

Web2.0 Herd Criticism

Andrew Keen: Web 2.0 Is Reminiscent Of Marx

Enterprise Defaults and Pressure

Muli Koppel: Pressure (The first article in the Pressure trilogy)

schwartz-barry.jpg

Technopod Rating: 9
podcastIcon1.gif Barry Schwartz – Less is More

Original Podlink: Barry Schwartz – Less is More

James Surowieki.jpg

Technopod Rating: 9
podcastIcon1.gif James Surowieki – Independent Individuals and Wise Crowds

Original Podlink: James Surowieki – Independent Individuals and Wise Crowds

technorati1.jpgTechnorati Wisdom of Crowds, Technorati Less is More Podlink ,Technorati the Wisdom Podlink

Scripting and the Real-Time Enterprise – Scott Mace and Peter Yared

February 23, 2006

itcLogo2.gif

The social network of “scripting” (Ousterhout, Dave Winer, ScriptingNews, Open Source, Wikipedia, Tagging, Ruby, LAMP…) has a compelling, seductive power in an era where there’s no Time, as it offers means for rapid creation and assimilation of ideas (through the Blogosphere), standards (micro-formats), tools (open source) and products (mash-ups). And while not everyone of us is a code scripter, many of us are content-taggers, which is an equivalent to “Ontology-scripting”. So we’re scripters.

Yet, when it comes to the Real-Time Enterprise, a conflict emerges between the Quicky Culture of coding and the Managerial Culture of operating the code (as manifested in two similar design-principles: “Design for Operations” and “Design for Failures”). Scripts and Enterprise-Grade Products & Services are, therefore, considered antipodes.

ActiveGrid, an Enterprise-Grade scripting IDE and run-time environment, aspires at mitigating the gap between these two cultures. Without judging the quality of ActiveGrid development and runtime environment, nor its positioning compared to roaring ROR, I think the behind-the-scenes philosophy of this environment, as presented by ActiveGrid’s CEO, Peter Yared, is fascinating.

Scott Mace, Opening Move’s host, leads brilliantly Yared in the garden of today’s memes, J2EE and .NET vs. scripting, ActiveGrid vs. Vista, scale-out through utility computing; dynamic, ad-hoc, distributed, transaction-level policy, logging & auditing; the reality of 2PC (two-phase commit) in J2EE AS; freedom of languages’ choice on the LAMP stack vs. Java’s write one, run anywhere; and much, much more!

Enjoy.
yared.514.jpgmace-200x250.jpg

Technopod Rating: 9
podcastIcon1.gif Scripting and the Real-Time Enterprise 59:07, 27.1 mb, recorded Apr 20th, 2005

Original Podlink: Opening Move with Peter Yared

References:

John K. Ousterhout – Scripting: Higher Level Programming for the 21st Century

activegrid_home_logo.gif ActiveGrid

technorati1.jpg Technorati ActiveGrid, Technorati the Podlink

Food for thought from Amazon CTO Werner Vogels

February 20, 2006

Another Ganger, Mike Vizard, recently joined the podosphere under the auspice of ACM Queue. In this ACM Queuecast, Vizard talks to Werner Vogels, Amazon CTO, who really surprised me (I am talking as an ex Data Center manager in a mobile operator).

When it comes to Information Technologies, every service at Amazon is a micro-cosmos. A dedicated group of application developers receives total, end-to-end, life-cycle responsibility over the service. This Micro-Cosmos is autonomous, or as Vogels put it: they can develop the service with whatever software development methodology and with whatever technology and tools they see fit – as long as they deliver on time and make sure, constantly, that the service meets its SLA.

This total responsibility leads to an extreme version of the “Not Invented Here Syndrom”. Amazon is not relying on any 3rd party vendor to guarantee its availability, data consistency etc. If a service is designed for no-data-loss, then the application is programmed to guarantee that. If you thought there would be a delegation here to some Enterprise Storage functionality, like Raid, BCV, Snapshots – well, you were wrong.

Some more goodies:

– The design is always paranoiac and pessimistic, i.e. design for failures.

– Resource Sharing? Not in Amazon. Any Service is human, software and hardware self-contained; a total encapsulation.

What about Web Services, Grid Computing, Virtual Machines? Well, I guess you should pick it yourself by listening to the podcast 🙂

vogels.werner.632.jpg

Technopod Rating: 9

podcastIcon.gif Werner Vogels, Amazon CTO, ACMQueuecasts 22:40, 20.7 mb, Jan 23rd, 2006

Original Podlink: Discipline and Focus

technorati.jpg Werner Vogels, Technorati the Podlink

SOA Platform Evolves to Require Policy-based Governance

February 19, 2006

briefingsdirect.JPG
Dana Gardner talks to Miko Matsumura, Infravio, about SOA, SOA Governance, End-to-End Life-Cycle Service Management and more.

Miko Matsumura is an interesting guy, besides his marketing faculties :). Whenever I stumble upon one of his works I am not disappointed. Matsumura is tightly coupled to the OASIS SOA Blueprints and Reference Architecture, two technical committees that were established five years too late. See Reference links.

In the last years Matsumura has been working in two SOA Governance companies: Systinet (which was recently acquired by Mercury) and Infravio, a fact that gives him an extra score, imho, as finally Enterprise SOA Management is getting the thought and products it requires.

miko.jpg

Technopod Rating: 8
podcastIcon.gif SOA Platform Evolves to Require Policy-based Governance

Original Podlink: SOA Platform Evolves to Require Policy-based Governance

References:

SOA Blueprints Technical Committee – Wiki

Miko Matsumura’s Intentional SOA White Paper

Muli Koppel’s Blog: The Toolsmiths, The Manager, His Repository and Its UDDI Lover

technorati.jpg Technorati Miko Matsumura, Technorati the Podlink