So, as you might imagine, I was pretty busy during the Denver conference. I will not try to detail all of that here; however, I have built a more lengthily report on Wikispaces at SHARE in Denver 2009 That report has some of my notes from sessions I got to attend - mostly the keynotes, and information that I could share about my activities on the board. The next SHARE conference will be in Seattle March 14-18, 2010.
I attended the Open Repositories 2009 Conference in Atlanta, GA this week. It certainly has been an interesting experience. So, what is Open Repositories? That's a bit hard to answer in one sentence. Even their website http://openrepositories.org/ does not explain other than to say they have a conference each year. It appears to be a group of people who have several strong beliefs about how to build repositories for universities, libraries and research institutions. They believe in open source, open access and sharing of ideas about how to make all that happen. The people who attend are a mix of Librarians (information science), Researchers, Computer Science people who are interested in repositories, Application Developers who create these repositories and the various companies that supply repository services. There are three main repository players here: Fedora Commons, DSpace and EPrints. This year Microsoft Research joined in with a free offering called Zentity that extends Microsoft Office 2008 to build, access and modify a hosted repository.
One of the big news items at this conference was the merger of Fedora Commons and DSpace into a single not for profit company called DuraSpace. The sessions range from vendor presentations, to presentations about repository design, and includes presentations by developers about how they build some applications. There is a group of the applications developers who attend (or not) who actively build and contribute code for the the three main players.
I took a lot of notes during the conference, and started to post them on the DLT Department wiki, but as I gathered more information, I decided to put the notes in my personal space because I think the information may be of use to more than just the DLT department. They notes are at Open Repositories 2009

Last night (about 8:15 pm) my daughter Katie (now Verbano) delivered a perfect little boy. Gavin Christopher Verbano is 20 inches long and weighs in at 7 lb 11 oz. The picture above is father Chad with Grandma Grace in the background, and of course, the star Gavin. You can see more pictures at
http://tinyurl.com/GavinV/
I'll post more pictures later today. I'm heading back to the hospital to visit.
Al W.

I put them on picasaweb at http://picasaweb.google.com/alw023/RobinRetirement#
I'm a tiny bit late announcing this, but you'll see why in my next blog.
Al W.
This year I tried to avoid using the flash as much as possible. The results are a bit mixed. No red eye to fix (Yippee!). It certainly was less intrusive for taking pictures; however, the lighting was pretty dim in some instances. I set the camera to Manual mode, bumped the ASA up to 800, set the speed down to 60 and the aperture to 3.5. That worked pretty well in most cases. In the dimly lit rooms, the pictures required a lot of post processing. The result is grainy with drab colors. The other challenge was to catch perky speakers and conveners when they were standing still at speed 60. I'm sure there is a lot I need to learn about low light photography, and I welcome any suggestions.
Al W.
- TrueCrypt is open source and free
- It works for Mac, Linux and Windows
- It looks like it does a good jot encrypting - though I'm not an expert
Al W.
- Elastic Fox is a Firefox plugin that helps track your Amazon Machine Images in EC2
- There is a Service Health Dashboard to track the status of all your Amazon Web Services
- Here is an example of Dynamic Scaling if your web service is wildly successful
- Assay Depot is another success story. This blog includes a discussion of how they do backups on AWS.
Al W.
Most vendors want you to believe it is a product they sell. Wrong! It's supposed to be about our business, not theirs. From Irving Wladawsky-Berger's blog SOA, Services and Business Architecture, he says "SOA has been gaining ground as an effective mechanism for defining business services, the software that implements such services, and the software-based tools that enable people to effectively take advantage of them." and "The hope is that with SOA and the many different tools developed around it we will be able to design, simulate and test business services in business terms - prior to their implementation in IT systems."
SOA is a business process - Service Oriented - which depends on repeatable business processes being identified, automated and (hopefully) shared. The A part of SOA gets us into trouble and implies that SOA is an Architecture when in fact it can be implemented and supported by many architectures. That is why Brenda Michelson refers to SOA implementation as Service Orienteering. We in IT love our technologies, and hence we tend to focus on the Architecture part of SOA - perhaps because that is most familiar to us. Bad TLA!
Al W.

I think it is a lot more than just an IBM "On Demand" services approach, and I believe it is the next step beyond Grid computing. Google is actually one of the leaders Cloud Computing. Cloud computing is how Google supports the Google Aps, Search, Sites, Pages, etc. Google is currently working with Universities to train students in the way that Google does its computing. IBM is buying in and partnering with Google. TJ Watson and three other IBM research centers are players. Google has also developed a Google File System (GFS) to deal with the massive amounts of data involved. A driving force here is that a lot of research institutions, like us, are flat running out of room, cooling and power to house massive HPC installations to meet research computing needs. I think one of the places the Google and IBM partnership may be going is to form a research computing cloud where each contributes some resource that can be shared by all. That is beginning to sound like a Stone Soup approach. In the early days of HPC we used that approach, and we called it condominium computing - each department bought some part of the massively parallel networked computer resource, and then got some time to use it all.
The intended end state of cloud computing is that computing resources become a commodity like electricity. Google's CEO, Eric Schmidt believes that 90% of all computing will reside in a cloud, with things like high-end graphics processing still residing in the other 10%. IBM started two notable research projects in cloud computing this year: Critical Enterprise Cloud Computing Services with Georgia Tech and Ohio state, and Resources and Services Virtualization without Barriers (RESERVOIR) in the EU. Amazon is already offering it in their EC2 and S3 services. Some think that the Amazon stuff is too expensive at about $1 per hour for compute time, but I'm not sure I agree. I also expect these vendors will negotiate price. I'm sure IBM intends to move to where they can sell this to their commercial computing customers at some point, too.
Interestingly, these cloud computing services are based on several open source projects including Linux, Apache Derby, Hadoop, Lucene, and Nutch. Yet another example of how open source applications can be combined into products that vendors can sell.
One of my tests for reality is to see who is talking about this topic. Irving Wladawsky-Berger is now talking about it in his blog Reflections on Cloud Computing. When people like Irving focuses on something like this, I begin to think it is real.
Al W.
The Thunder Storm picture is from Flickr by Vermin Inc
The SOA program was particularly relevant to PSU. It delved into a lot of the technologies, but even more interesting, it provided different perspectives into what it means to implement SOA provided by three different well-know speakers.
The general session keynote speaker was Jeanette Horan, IBM VP of Business Process and Architectural Integration. She presented IBM's Internal SOA Journey, which tells how IBM is deploying SOA to go from 155 Data Centers with 192 CIOs to 6 Data Centers with one CIO. This is a well funded, very structured, top-down driven approach.
The SOA Project keynote speaker was Brenda Michelson, an independent consultant and Contributing Architect for Patricia Seybold Group, presented a different approach. In her presentation The SOA Journey, Brenda compared implementing SOA to hiking the Appalachian Trail - not something you usually do all in one hiking trip. She refers to developing SOA as Service Orienteering where the focus is on specific business benefits.
Another keynote presenter was Paul Giangarra, IBM Distinguished Engineer in the Office of the CTO, IBM Federal. His presentation, SOA Start to Finish, focuses on the issues of information sharing within the enterprise with real-life examples from SOA projects in the Federal Government.
The Virtualization topic included a lot of IBM z/OS and z/VM zLinux presentation as expected, but also had a good variety of non-mainframe sessions. For example, VMware presented Virtualization in the Next Generation Data Center, and IBM presented x86 Virtualization Technologies. One of the most valuable to me was Desktop Virtualization, the Return of the Thin Client presented by Paul Seay, a Chief Architect for Northrop Grumman and a fellow SHARE board member.
SHARE is also delving into emerging technologies, especially through the Integrating Technologies Project. One of the most interesting examples is a session called Web 2.0 Goes to Work presented by David Barnes from IBM. The videos that he uses within his presentation are Available on YouTube, and he tagged all his web site references on Delicious.
Here is a general wrap-up article for SHARE in Orlando. In two weeks I'll be heading for Chicago with our SHARE Program Managers to pull together the session content for SHARE in San Jose.
Al W.
