Monday, January 16, 2012

Many Core processors: Everything You Know (about Parallel Programming) Is Wrong!

David Ungar is "an out-of-the-box thinker who enjoys the challenge of building computer software systems that work like magic and fit a user's mind like a glove.". this is a summary from  SPLASH 2011 in November 2011
In the end of the first decade of the new century, chips such as Tilera’s can give us a glimpse of a future in which manycore microprocessors will become commonplace: every (non-hand-held) computer’s CPU chip will contain 1,000 fairly homogeneous cores. Such a system will not be programmed like the cloud, or even a cluster because communication will be much faster relative to computation. Nor will it be programmed like today’s multicore processors because the illusion of instant memory coherency will have been dispelled by both the physical limitations imposed by the 1,000-way fan-in to the memory system, and the comparatively long physical lengths of the inter- vs. intra-core connections. In the 1980’s we changed our model of computation from static to dynamic, and when this future arrives we will have to change our model of computation yet again.
If we cannot skirt Amdahl’s Law, the last 900 cores will do us no good whatsoever. What does this mean? We cannot afford even tiny amounts of serialization. Locks?! Even lock-free algorithms will not be parallel enough. They rely on instructions that require communication and synchronization between cores’ caches. Just as we learned to embrace languages without static type checking, and with the ability to shoot ourselves in the foot, we will need to embrace a style of programming without any synchronization whatsoever.
In our Renaissance project at IBM, Brussels, and Portland State,   we are investigating what we call “anti-lock,” “race-and-repair,” or “end-to-end nondeterministic” computing. As part of this effort, we have build a Smalltalk system that runs on the 64-core Tilera chip, and have experimented with dynamic languages atop this system. When we give up synchronization, we of necessity give up determinism. There seems to be a fundamental tradeoff between determinism and performance, just as there once seemed to be a tradeoff between static checking and performance.
The obstacle we shall have to overcome, if we are to successfully program manycore systems, is our cherished assumption that we write programs that always get the exactly right answers. This assumption is deeply embedded in how we think about programming. The folks who build web search engines already understand, but for the rest of us, to quote Firesign Theatre: Everything You Know Is Wrong!
 This video, is an interview with David Ungar from IBM Renaissance Project on programming many core computers and non-determinism, is mind boggling and expands beyond the text above.


David Ungar likes anything creative, from poetry to starting a business. The moment we built a 1,000 core process is like discovering a new space .  He also says , Small Talk - originally developed i

Note:  I added on February 7, 2014  the  original talk at Splash-2011  Everything You Know (About Parallel Programming) Is Wrong!: A Wild Screed About the Future 

Thursday, January 12, 2012

Why Include the Network Virtualization to the Cloud

Following the publication yesterday of the Network Virtualization (NV), The New El Dorado  many people asked more about M2Mi  (Machine to Machine Intelligence).

In an article originally published in the Intel publication Journey to Cloud Magazine: Volume 1, Issue 2 , I read one of the best rationale for Network Virtualization (NV) as organic part of the Cloud.

Here is a summary:

Network virtualization is a technology that allows layers of security and connectivity services to be deployed over heterogeneous, multi-vendor devices. The term “virtualization” applies since the technology abstracts applications away from propriety holes in and between incompatible devices. Incompatibilities among devices make networks complex and difficult to maintain. Human intervention often leads to mistakes, This is why more than 75 percent of traditional IT budgets are spent purely on networking

Why include the network in the cloud? Because:
  • ALL END USERS access the cloud application through the network.
  • THE NETWORK IS HOME to most security tools and devices.
  • The NETWORK IS COMPLEX and labor-intensive and is thought to require near "black magic skills" to maintain .
IT experts often refer to “the wilds of the data center.” The network is a big reason for the tangled jungle metaphor is the cloud security battlefield—the first and final layer of defense where most attacks start, from distributed denial of service to virus intrusion.

M2Mi’s Network Virtualization coordinates, security and access technologies like the Intel® SOA Expressway and Intel® Expressway Cloud Access 360 turn into sequential layers of dynamic protection, customizing applications specific security settings and augmenting them with automated response capabilities.

From a security point of view, network virtualization must assume all underlying network devices. Not all traffic can be trusted without extensive verification and validation

M2Mi uses Intel® Trusted Execution Technology (Intel® TXT) in the cloud to enable encrypted communication between the CPU and the local operating system. Intel TXT and M2Mi are used to deliver a trusted boot-up, securing the cloud from start to finish. incoming traffic is white-listed (i.e., registered traffic sources are identified and then allowed to move through a reduced security line; average traffic sources are scrutinized and X-rayed; and sources labeled suspect are put through additional measures before being allowed to board the cloud).

Metaphorically, the white-listing of the Cloud traffic sources resembles the security checks at the airports, where all passengers traffic is allowed to pass after a standard security checking, while some passengers require additional security measures before boarding the airplane.

Using this metaphor, we want the airport check point to look like  this:

and not like this:

How security checks should not look like
These features offer a nice surprise to the cloud operators. Network and server virtualization are billed differently in the cloud. And the network usage creates substantial incremental revenues potential.

The network device manufacturers have filled your network with wonderful features that network virtualization can expose as billable, premium services. These features include (1) secure socket layer (SSL), (2) packet inspection,  (3) quality of service (QoS), traffic prioritization,and data encryption

This enables premium cloud add-on services to be delivered based on value-add instead of purely usage rate pricing models.  Customer pays for premium billable services to increase security. This makes sense. There is no such thing as a  security policy policy that fits all

The challenge for M2Mi solutions in the Journey to cloud is how to securely manage across millions of remote mobile devices, through Telco networks, across enterprise environments, and safely through the cloud

Tuesday, January 10, 2012

Network Virtualization (NV), The New El Dorado

NV: What it is and what is the market size?

On Open Flow discussion list, someone from Stanford University asks on October 5, 2011:
I am trying to get a feel of how big the network virtualization industry will become. Do you have some sources on how much the industry is or will be worth? I have seen some figures on the compute and storage virtualization industry (such as from VMware), but I have not found such figures, or even discussion on network virtualization.
 The reply comes from another contributor at Computer Science Stanford University
If you're a computer scientist, this is entirely the wrong way of thinking about it. Instead I would ask: what does network virtualization mean? how does it affect the networks, systems, and applications you are interested in? What transformative change can it bring about? What awesomely cool stuff can you do with it? How will it affect the world? Understanding the technology and what you can do with it is the important part.
If you're a VC, well then, it's going to be HUGE - bigger than Cisco, Juniper, Comcast and AT&T combined!*
Nobody formulated  yet a widely accepted definition of what  NV is, but everybody knows this technology has a HUGE market size. Google trends shows a sudden jump in the last18 months

In absence of a substantial definition or market data, the interest is triggered by news like Big Switch Bets On OpenFlow, Network Virtualization  (ONF) where we read that
"...some ONF members describe OpenFlow as a virtualization technology. "We think we can be the VMware of networking,"
Or even news that Dario Calia (from Cisco) Joins Vello to Lead Network Virtualization Development, another ONF based start up or  ConteXtream Launches Software-Defined Network Virtualization Solution for Cloud and Hosting Providers. The latter is a an event from December 2011, barely 1 month ago

Google trends is a mechanical tool, not an intelligent one, as it extracts the NV trend from content titles it believes contain the words "NV" Thus, it misses significant players, for example the Cisco Acquisition of Linesider, completed in December 2010

 It misses companies like Nicira , Embrane and Machine to Machine Intelligence (M2Mi). In this blog I would look into two key  companies: Nicira and M2Mi


Nicira is a the darling of Silicon Valley, while no one really knows what they do. Their web site is vague,yet  we see an impressive list of management people and investors. Anderseen Horowitz , top progressive VC on the Silicon Valley, Diane Greene, ex CEO VMware and  Andy Rachleff, Benchmark Capital and Stanford professor. Nicira does not list any customers, and it has $40M in funding

In the 2009 article Open vSwitch Project Launches, Reuven  Cohen writes
Before I go into the details of the project I will say I find it very interesting the complete lack of any insight into who is behind the project. The website doesn't include company or contact information.... According to the whois lookup, the website is registered to MartinCasado a student at Stanford with no affiliation with Citrix. Some further digging through the mailing archive I discovered the only real activity seems to be coming from a stealth company called Nicira.

Now we know Open vSwitch comes from Nicira, which means they are not fans of ONF. To find out what is going on, we must a New York Times blog from October 2011 What Nicira is up to . This New York Times piece soared the curve trend of Nicira as the the real embodiment of NV

Steven Mullaney, the CEO says vaguely
. “We are the most non-disruptive disruptive company out there,” he says. “All the networking companies will be around.” What he may be ripping out, however, is the most valuable part of those companies’ business: the brains that govern what they do. He would not say when this will start, but indications are that it is a matter of months.
A matter of months?
The company originated with Martin Casado, now its chief technology officer. According to Mr. Mullaney, Mr. Casado was working on network security for Lawrence Livermore National Laboratory some years ago, when he was asked by United States intelligence agencies to figure out how to run a global network that could continually change levels of security and authorization.
“They needed something that would be flexible all the time,” Mr. Mullaney said. “He couldn’t do it, so he went to Stanford to solve the problem.” While at Stanford, Mr. Casado began what would eventually be Nicira with Nick McKeown, a professor there, and Scott Shenker, a professor at the University of California, Berkeley.
Here we have to admit the big truth . There is not Network Virtualization without proper security. And this security - not the virtualization per-se - is the biggest stumbling block. Here is  an example of use of the of theNicira technology at Japan's N.T.T (not clear if it is customer or a beta tester)

 There is no way to perform a live migration of data from Mushashino to Atsugi, without proper security. Nicira may or may not reach the security standard they set themselves as a goal (continually change levels of security and authorization). If they don't,  the security feature may have to be set down a more pragmatic achievable level, The story on  how Stanford and Berkeley universities became involved shows the technology is not yet proven. One day it will, yet the first Nicira adopters will take a higher risk

M2Mi (Machine to Machine Intelligence)

M2Mi is a  start up  based on NASA Ames Research Park at Moffett  Field, California. They started in 2006 and will reach six years as a profitable business in 2012. Their Active ConīŦguration platform provides orchestrated services through the combination of an Orchestration engine and Network API libraries. Orchestration allows you to “transactionally” sync interactions across different devices all in one process.

According to the NV entry from Wikipedia
Machine-To-Machine Intelligence (M2MI) technology covers both Internal, External and Multi-vendor software and hardware based technologies. M2MI is unique in its approach of applying "whitelist" blocking across all multi-vendor network elements, this approach ensures that Virtual Machines can not be "ARP spoofed", a technique used to compromise Virtual Machines at the network level.
Security of  M2Mi is proven.  Their patented GlobalockBox  can be used with AWS S3 and,has a cloud front end and middle ware that works with plug-in libraries for encryption. For virtualized clients, it uses a commercial library for encrypting business applications; for government projects, it uses the NSA Suite B Cryptography for IPSec (a library that incorporates elliptic curve cipher technology)."

M2Mi's presence in NASA Ames Research Rank hints they have  government customers. But their list of business  partners includes some of the top commercial companies in Silicon Valley: Intel, Oracle, Ericsson, Joyent and Cisco Systems.

Like the other players in NV, M2Mi are not disclosing much beyond their proven and patented technology.

Geoff  Brown, M2Mi CEO

Last year George Brown wrote on  Cloud Computing Google group:
Quite simply put, we address a significant market pull arising from the requirement to ... provide layered robust security services of which network provisioning is one element. I think it is fair to say that this area is complex, while customers want a simple to use, "cookie-cutter" solution.
Geoff sums up his company as follows:
From my perspective we established and now lead the growing field of Network Virtualization
An now what?

I agree that  (1) NV  is not a "cookie-cutter" solution,  (2) that  the solution must blend NV with the Cloud Data Center Virtualization and  (3)  the security must be near military grade. The market is HUGE and we can not wait to see a proof  from analysts.

By the time the analysts (Gartner, IDC, Forrester and the others) will create the taxonomy of the NV market and estimate its size, it will be too late. The treasure will be shared by early players. Like in  any poker game, some players bluff, and all keep their cards away from prying eyes.

M2Mi appears to have a proven, albeit hard-to-use complex solution.  Nicira has the top names backing and  promoting it, but we are promised results "in the next few months", a  time frame continuously sliding so far.

 Cisco,  VMware and HP appear as leaders in the IT community from a year ago. Brocade, Emulex, Force 10, Huawei, IBM, Juniper, Oracle/Sun are also listed.

Cisco, VMware will the prime candidates to acquire any of the startups described in this blog. They have the power to collect the revenues from he market defined as "bigger than Cisco, Juniper, Comcast and ATT combined!"

Tuesday, January 03, 2012

Will Betterness Work?

Umair Haque Beterness Manifesto is one of the most influential document published in 2011. Below are three essential quotes. They are designed as a remedy to the malaise we observe mainly in US economy and psyche. Summarizing the quote, betterness implies:
  • We are the freest people in history. It time to use our freedom
  • We are worthy and we have talent. Start a new company or find a better one, rather than wasting your life in "organizations that misallocate it, underutilize it"
  •  Over the next decade, the businesses that can't do better, the ones you're giving your talent away to, are to go extinct anyway. Cut the cord now, before the axe falls and cuts it for you.
  • Stop living somewhere where you're forced to, like it's groundhog day
This is what will bring us betterness. But how is this applicable to people living in North Korea? They will read the four bullets of advise above, as if applicable to people living on the moons.

Perhaps everyone agree that this is a problem in North Korea, but what about most of Europe? Even United Kingdom, the civilization that created the democratic institutions as we know them today, people are happy to be underutilized as long as they have a job security and small town values preserved. And regarding point #4, no one really wants to live anywhere else.

Why Umair Haque  - himself  a member to the world top 50 thinkers award recipients in 2011 - met much more adoption in US than England? The two countries have a long time peculiar relationship. The British supply original ideas and individual brains of historic significance (e.g. Darwin, Lord Keynes,  or the Rolling Stones) but US is the place where these ideas not only become reality, but blossom.

Here is the paradox: betterness, designed to  revitalize talents and prosperity everywhere, will distance US from the rest of the world and maintain the advantage US economy has when producing iconic brands like Apple, Google, Facebook and tomorrow many more. One can not create iconic brands without tolerance to risk and freedom.

Therefore the betterness will create even a wider gap between North America and the rest of the world, with exception of a few countries (Israel is one of them). Betterness will create inequality.

Inside US itself, the Silicon Valley mentality dominates only in isolated spots. And nearly 90% of American employees simply lack the desire, never mind the guts, to leave a 20th century corporation that assures them a salary month to month while the lifetime employment is nothing but an illusion that can be shattered at any given moment.

For those talented and extremely gifted,  there is always a the option to open offices in US  and operate from Silicon Valley. You can not wait for ever for the political system to change in your country. You can not wait for ever to see deeply ingrained suppressing mentalities (like the caste system in India) to disappear over night.

This where I try to help.

Real change doesn’t begin with governments, presidents, or prime ministers. It begins with each of us. In the 20th century, never-ending mass-marketing, monopoly, and mega-politics came together to convince us, each and every one, that we’re not really free: just free enough to choose between different flavors of the same old toxic junk. It was a trick, a ploy, a television hallucination. We’re the freest people in history. It’s time to use it like we meant it."
"You're worth something. Stop giving your talent away to organizations that misallocate it, underutilize it, and possibly even abuse it. If you're doing something meaningless, quit. Betterness can't happen if you're spending your life churning out toxic junk. It can only happen when more meaningful work is done. Find a company that's better. Better yet, start one. No, it's not easy. But here's the thing: over the next decade, the businesses that can't do better, the ones you're giving your talent away to, are to go extinct anyway. Cut the cord now, before the axe falls and cuts it for you."
" If you're living somewhere meaningless, move. Exurban sprawl, mega-highways, big-box stores: that was the American Dream in the 20th century. In the 21st, it's closer to the awesome Richard Florida's dream of thriving, tightly-connected communities, that make up vibrant cities. Living somewhere where you're forced to, like it's groundhog day, hit the same old big, lame, toxic businesses, over and over again? Those places and spaces were built to support an industrial economy. Today, they're a barrier to letting it crumble and fall. Move somewhere where there's a local community made up of passionate, talented people, a community you can nurture and that nurtures you. It just might be good for your soul."

The Betterness Manifesto by Umair Haque

Sunday, January 01, 2012

Galati Flower Mill, aka Moara Bros. Aronovici

Moara Galatilor in 1948. Note the name Bros D & A Aronovici
and the sign IFA  (Intreprinderea Fratii Aronovici) The Aronovici Bros Enterprise

Moara Galatilor in 2011 April. Bros Aronovici name was removed

April 2011: A dog takes care of the the run down flour mill

2012: Tell the status quo to go to hell

Umair Haque'- one of the top 50 thinkers of 2011 in Mastering the Art of Living Well advises younger people what to do to contribute to the "betterness".
  • Do not study so you can be a faithful, loyal, unquestioning "employee" with the commoditized, routinized analytical skills to get the (yawn, shrug, eye-roll) neo-Fordist job done
  • Be a wholer person.  What society has a shortage of living, breathing well-rounded humans; with a moral compass, an ethical core, a cosmopolitan sensibility, and a long view born of historicism
  • Without a refined, honed, expansive sense of what great accomplishment is, you stand little to no chance of ever pushing past its boundaries yourself.
  • Your youth should be spent pursuing your passion — not just slightly, tremulously, haltingly, but unrelentingly, with a vengeance, to the max and then beyond.
  • Dream laughably big
  • If your quest is mediocrity, then sure, master the skills of shuffling Powerpoint decks, but if your quest, on the other hand, is something resembling excellence, then the meta-skills of toppling the status quo — ambition, intention, rebellion, perseverance, humanity, empathy — are going to count for more, and the sooner you get started, the better off you'll be
  • When you fail, and fail big — forgive
  • Mistakes aren't the end of the world, but the beginning of wisdom — and firmly step forward into possibility.
  • The status quo is no way near prosperity:   dying metropolises, battered exurbs, mass unemployment, nail-biting fear of the future, plutocracy and protest, the crumbling ruins of empire
  • So map the horizons of your own journey, and, when the status-quo tells you it can't be done, tell the status-quo to go to hell

Blog Archive

About Me

My photo

AI and ML for Conversational Economy