powdermonkey

Real-time embedded open source software OpenCPI website Launched!

Our OSS website is launch for OpenCPI, Sign up!

The Component Portability Infrastructure (CPI) is a real-time embedded (RTE) middleware solution that simplifies programming of heterogeneous processing applications requiring a mix of field-programmable gate arrays (FPGA), general-purpose processors (GPP), digital signal processors (DSP), and high-speed switch fabrics. The “mix” can be over a lifecycle (technology insertion) as well as within a single implementation (to meet SWAP constraints). CPI improves code portability, interoperability, and performance in FPGA and DSP-based environments by providing well-defined waveform component APIs (application programming interfaces) with a set of infrastructure blocks that act as a hardware abstraction layer (HAL). CPI is also appropriate for the incorporation of GPU and multicore technologies. CPI is uniquely positioned to meet the goals of S3 since in some sense component-based systems are computer-science’s answer to dealing with “knowledge capture” and lock-up of intellectual property (IP). CPI is classified at a technology readiness level 6, and is in programs that will establish it for level 7.

Built on the U.S. Government’s Software Communications Architecture (SCA) standard, CPI extends component-based architectures into FPGAs and DSPs to decrease development costs and time to market with code portability, reuse, and ease of integration. All interfaces are openly published and non-proprietary, using an appropriate mix of industry and government specifications.

To overcome the challenges of code portability in FPGA environments in particular, CPI provides a pre-validated set of building blocks to interface the FPGA waveform applications with high-performance switch fabrics, onboard memory, system command and control, and wideband I/O. CPI’s non-proprietary interfaces act as an abstraction layer to increase the portability of FPGA applications. A verification suite is also included to facilitate debugging and reduce development time.

At the highest level, the CPI vision allows users to outsource the technology transition management job to others. Using the CPI interfaces, developers can protect their application development investment by cost-effectively moving their applications to new generations of systems using the latest technologies. CPI is essentially a kit of necessary pieces to create an application platform for component-based applications based on the SCA model extended to a heterogeneous mix of computing, interconnect and I/O resources. When CPI is adapted to, and installed on a platform, we say that the platform is now “waveform-ready”. While the SCA defines the operating environment and APIs for C++ software applications components running in a CORBA and POSIX-compliant environment, CPI extends the SCA environment, according to the Proposal 289 to SCA (FPGA/DSP extension by Mercury funded by program office) to DSP and FPGA technologies. Analyses for suitability for GPU and Multicore technologies have shown promise.

Meeting, May 21 in DC: Open Source, Virtualization & Cloud Computing

Great even to be held in DC, May 21 and its free. Pretty amazing lineup of speakers we've managed to get and its at no cost:

Federal IT on a Budget Forum

Join government IT professionals and industry thought leaders for this FREE full-day event to get smart about implementing Cloud Computing, Open Source, and Virtualization in your agency or organization.

Co-Chairs:
• Pete Tseronis, Deputy Associate CIO, Department of Energy
• Tim Young, Senior Manager, Federal Government Services Group, Deloitte Consulting LLP and Former Deputy Administrator, Office of E-Government and Information Technology, Office of Management and Budget
Morning Keynote:
• Jim Whitehurst, President and CEO, Red Hat ˜ Answering the Call for Transparency in Government: The Open Source Opportunity
Co-located AFFIRM May Luncheon and CIO/CFO Panel:
The New IT Economics:
• Herb Strauss, Chief Strategy and Marketing Officer, Robbins-Gioia [Moderator]
• Robert Carey, CIO, Department of the Navy
• Casey Coleman, CIO, General Services Administration
• Chris Kemp, CIO, National Aeronautics and Space Administration Ames Research Center
• Radha Sekar, Assistant Deputy Under Secretary for Financial Management for the Under Secretary of Defense Comptroller USD(C)

Register now at www.meritalk.com/2009-federal-it-forum.php

OSSI helping DISA convert HR/Admin software into Open Source

great story for OSSI, the Gov and open source software:
DISA to open source administrative software

Other agencies free to use, modify human resource management software: Joab Jackson

The Defense Information Systems Agency (DISA) plans to open source a suite of programs that it developed for administrative tasks. The agency has signed a Cooperative Research and Development Agreement with the Open Source Software Institute (OSSI) to help release the source code of the programs.

The set of 50 programs, collectively called the Corporate Management Information System (CMIS), handles duties such as human resource management, training, security, acquisition and related functions. All the programs were developed by internally by DISA, and are used by more that 16,000 users worldwide.

“Numerous other government agencies have asked if we'd allow them to adopt CMIS for their internal use," Jack Penkoske, DISA's director of manpower, security and personnel, said in a statement. "We believe this will be a win-win for all involved.”

By allowing third-party developers to view, modify, and reuse the software source code, DISA is hoping that others will improve the code when they modify it for their own purposes.

DISA will license the software under version 3 of the Open Software License as well as version 3 of the Academic Free License, according to John Weathersby, OSSI's executive director.

Make Love not CyberWar

Been thinking allot about cyberwar and the orgy of funds raining down upon the tribes of defense contractors and I got to wondering, why fight a war you can't win.

Now try to follow me: the US made it a central part of our national strategy to own the skies, I think the last soldier to die from an air to ground strike was in the Korean conflict. So we spent the funds, developed the tech, trained the pilots and can now pretty much own the sky above any country we want to. We decided that we didn't want any competition in the air and wanted to take away that option from any military.

So cyberwar: we are going to attempt to do the same things, spend the funds, develop the tech, do the training, etc. and we'll own the net. BUT of course the net is this large amorphous blob of foreign governments, public and private entities that don't like to take direction from a central source. In effect we can never own the sky. In this case the lessons learned of 'rule the skys' doesn't work and never will because you can never bound the problem of cyberwar. So what to do?

Don't fight a cyberwar by making sure no one else can either.

Most cyberwar exploits weaknesses in existing technologies (servers not protected, patches not installed, bad security, bad archtecture (internet issues), etc.), so turn that technology that deters cyberwar into a commidity for all to use, make it better and move on.

It's a bit counterintuitive but if you can't bound the problem then you will be constantly in the mode of 'wack a mole' hiting this and that group of nitwits or worse state sponsored folks looking to disrupt things. Or the ole Pearl Harbour attack, etc. This would entail pushing open source software, making it better and publishing it for all (yes including the baddies) to use.

So don't fight this war you can't win or at least make it really hard to see any results fast. Over time the systems should get better so that you can trace back a cyber attack and then drop a bomb on them since we own the sky.

where is Technology Transparency (t2)?

There has been allot of talk about government transparency, which is good, more data to citizens is better than less, etc.

What about technology transparency?

As a taxpayer I would like to be able to use some of the code my government pays for the creation of (unless classified). As a contractor to the military I would like to be able to examine source code for the multiple technologies that are being integrated together to provide the best capability possible for soldiers, sailors, marines and airman in the field and ensure it works, especially if that source code was paid for with taxpayer funding.

The DC CTO and the Sunlight Foundation have recently held contests about creating mashup applications for various data sets, the thing has has made these contests possible is that existing source code was available to change, modify and add to. Otherwise everyone would be creating applications from scratch greatly slowing the speed of innovation. Another great example is Mashable, which has snippets of source code available for use and reuse.

Technology transparency also enables individuals and organizations to learn how the source code functions and maybe ensure that nothing nefarious is going on behind the scenes.

So I hope in the whole transparency discussion going on Transparency of Technology is included in the discussion is included. 

What Open (should) Mean in the Military

Open means having access.

In the Dept. of Defense we need to ensure/create/enable open access to tax-payer funded technology, period.

Tax-payer funded software development needs to be openly available to all (except for really classified stuff). This means that we will need to figure out the acquisitions mechanisms needed to ensure that the intellectual property is made available to the defense industry and ultimately the tax-payer.

The inclusion of open source software into National Security Systems has jumped the chasm and shown the path. Military OSS products such as Delta3D, BRL-CAD, OSSIM, (and many many more) and of course Open Solaris and RedHat have shown the path for deployment of software systems, now we need to make the institutional default.

Prediction: 18 months from now this will be standard.

Let me know if you want to help.

Mobilizing Industry in WW2 & Compare to Today

just finished Mobilizing Industries in WW2, big take away that is pertinent to the new POTUS is clear lines of authority.

Repeatably throughout the book FDR created Offices, Commissions, etc. to run the various aspects of a wartime economy: all clashed with each other, all wasted time (most importantly FDR's), all fought the Army and Navy (while both were trying to take over aspects of economic planning), Congress and special interests were in the mix as well and all made decisions with out working out the impact a decision may have on other aspects of the economy.

Like: defer agricultural workers or people building planes (AG workers had a deferment, so more folks became farmers - manufacturing workers had no deferment, so instead of building a tank, they could be drafted to drive the tank).

There were roughly 125 commissions, etc. created - some subsumed others or were broken off. It didn't get fixed untill May 27, 1943 (Exec Order 27) with the appointment of James F. Byrnes as director of the Office of War Mobilization. FDR called him "the assistant President" and stated that Byrnes decisions were final (freeing up Roosevelt for Grand Stategy efforts). Byrnes portfolio encomapsed everthing expect foreign policy and military grand strategy.

My one concern now with Pres. Obama is that he seems to be building an analagous situation where there are too many chef's or CxO's, TARP managers, Congressmen, Senators, Cabinet Secrataries, etc. with too many overlapping responsibilities reporting or trying to influence him one way or the other. Endless meeting could result...

Applicable lessons from WWII were:
1. Personalities matter and must have the full confidence of the President. Byrnes knew how to get things done in Washington and wasn't too fully versed in econ matters.
2. Domestic and partisan politics will intrude on everything
3. Planning is essential

Standards, Standards, Standards....

It's the new year and so lets kick it off with a same story different day: story from FAS Project on Government Secrecy about Presidential records using non-standard formats

"Volume aside, the White House electronic records were often generated in non-standard formats using proprietary software that somehow must be accommodated by NARA."

Various reports (FCW) have said the program is over budget and behind schedule and the usual suspect, LMCO (lockeheed) is the contractor.

http://www.citizensforethics.org/node/35597

The U.S. Government Accountability Office (GAO) has questioned the ERA's readiness, especially since the project has endured some bumps along the way, including delays and cost overruns estimated at $16.3 million. The life-cycle cost for the complete ERA system, scheduled to be completed in 2011, has been estimated at $453 million, including development contract costs, program management, research and development, and program office support.

As recently as September, after studying the system's progress, the GAO urged NARA to create a mitigation plan in case it could not process the incoming records by Jan. 20, 2009. In a report to the congressional committees (download PDF), the GAO said, "If it cannot ingest the electronic records from the Bush administration in a way that supports the search, processing and retrieval of records immediately after the presidential transition, it will not be able to meet the requirements of the Congress, the former and incumbent presidents, and the courts for information in these records in a timely fashion."

OSS is the US Fed Gov

Interesting panel: Open Source Solutions - 2 Years In Review on Federal News Radio:
Program:

• Status & Progress on Open Source Solution in the Federal Government

• Best Practices from Government using Open Source

• Challenges to still overcome with using Open Source

• A Vision for the Future for Open Source

Panelists:
Patrick Gorman - CIO, ODNI

Chris Fornecker - Program Manager, Integrated Acquisition Environment, GSA

Daniel Risacher - Associate Director for Information Policy and Integration, Office of the Deputy CIO, DoD

Dave Mihelcic - CTO, DISA

Paul Smith - Vice President, Government Sales Operations, Red Hat

Peter Gallagher - Partner, Federal Civilian Agencies, Unisys

Moderator

Jim Flyzik -Flyzik Group

Joint Light Tactical Vehicle award this week: it won't be open I bet

I saw this story:
BMW seeking partners for open-source car software platform and then insidedefense.com had the story: "The Defense Department this week will announce technology development contract awards in the Joint Light Tactical Vehicle [JLTV] competition, according to the Pentagon acquisition chief.

And then talking with Jim Stogdill and reading his excellent post, I was left feeling its too bad that for all the millions that will go into JLTV, its too bad DoD didn't levy some requirement for the operating system to be more open, as in made open source. (Note I am less worried about vehicles, more so about software intesive systems that go inside them)

DoD is now going pretty explicitly to a winner take all model of acquisitons (2 pilot projects and then they pick a winner that goes into full production). Now this makes sense for big iron, physical goods (need to build factories, etc.), but for software its the the wrong, very wrong model.

When DoD picks a winner for a system that is heavily software reliant the Government is placing a bet that the Contractor Team will stay at the leading edge of technology. This never happens, Contractor Teams that win slowly build a Fortress to keep ideas out, while locking the Government into their platform. Then you end where we are now, with weird data standards, legacy systems, etc. Further, 1 team will win, but the loosing Team will have some better ideas or capabilities that the warfighter would love to have on their platform, but won't get because it would mean the Winner would ultimatley get less of the pie.

So, the industrial base shrinks, a division at a company goes out business, etc. OR as an alternative the loosing team could make open source their platform: kind of a 'loose the battle, but win the war" strategy.

Practical guidence for the Gov would be: 1. open the platform up (put an RFP out to make/deliver/build an open source platform or make OSS an existing one). 2. build a community around a capability. 3. let small contracts out for incremental improvements in platforms. 4. simplify the intellectual property rights lanscape for contracting with the government so that a developer can understand them