Zaki’s Land of Ideas

IMG_0032.JPG

Originally uploaded by davidirby

Amazing Photo of the Hearst Building

The Event Horizon problem in Debt Syndication

One of the challenges in creating a unified data architecture in debt securitization is the asset backed debt travels through several event horizons where all the information about the underlying asset is lost and replaced with a rating.  What is a needed is a business model that advantages organizations that adopt a database architecture as close to the ground as possible.

Architecture and Leverage.

One thing that is important to appreciate is that there is a high level recursion in debt syndication. The question is how should databases of assets deal with this recursion. Should their be replication of the data at each recursion or should their be network of databases that constant talk to each other?

01/06/2008

Originally uploaded by maniaczmanian

Muji Midtown is the New York Times building. Aesthetic convergence

I’ve recently listened to Andy Kessler talking about how the only way to revive the world of 3 letter acronym finance products(CDO,CLO,CMO,etc) is to introduce a common API across the finance industry. These comments mirror my own thoughts when I was learning about this securities from ex-gf at Goldman. This seemed like the way to overcome the issues in the securitized debt asset class that where being described more than a year ago by Nouriel Roubini and Barry Ritholtz.

Wall Street seems to be unable to execute this potential savior to their business model. Though major financial institutions are no longer sticking their heads in the sand, they are begining the process of the jockeying for the next business model. This is not a collaborative environment. What is needed is for an outsider to publish both an API and an open-source implementation of the API.

Other than the task of coding an implementation, the two major challenges of creating an API are  API use cases and  architecture. I think I’ve made some progress there. The purpose of use cases are to provide clues about what features exactly you would like to provide in an API

Use Cases
• Tranche modeling- Modeling the performance of a tranche in any macroeconomic model. This is the old stuff but you want to move this to client. Currently this capability rests mostly with the entity that structures the product and the rating agency. Also one would like to provide the ability to run more granular models that permit regional or industry specific modeling.
• Geocoding- One would really like to able to allow the owner of a CDO to extract geodata on the assets underlying their securities.
• Semantics- One would like to be able extract the business semantics underlying a structure. This would allow the structure itself to run as query or create newsfeed of business events affecting the product.
• Low probabilty event alterting. Another useful figure would to be able to extract the well characterized macroassumption from a product and test them against realtime data.

Architecture

Not too sure about this. Probably want your implementation to focus on Oracle/PostGreSQl as a database.

I’ve successfully installed KnowledgeTree 3.5.2b Open Source Edition on a hardy machine(amd 64)

The primary motivation for doing this is if you want to share system resources with other LAMP stack software without dealing with virtualization.

I want to provide some the key peices of information I discovered that I don’t see in other places on the web.

•  The source only installation instructions.(http://wiki.knowledgetree.com/KnowledgeTree_3.5.*_-_Performing_a_Source_Only_Install)

  This provides excellent information on dependencies that need be resolved ahead of installation. Most can resolved by searching synaptic. One also needs to use pecl to install fileinfo.

  Also keep in mind that knowledgetree prefers to be installed in the web root rather than a virtual server, thought it should be possible to figure out virtual server config settings that would work.

• Enable .htaccess
  The default Apache2 configuration settings for the webroot do not allow .htaccess files to override the default acpache2 settings.
  https://help.ubuntu.com/community/EnablingUseOfApacheHtaccessFiles
• Enable apache modules.
  Knowledgetree uses several nonstandard apache modules.
  Enable them with sudo a2enmod.
  You need two modules.
  rewrite and one other which I forget but can be determined from the apache2 error logs.
• Finally patch the file path ultility for linux.
  http://forums.knowledgetree.com/viewtopic.php?t=4068

Little setup guide - Developer at Openbravo

Notes on OpenBravo install on Xubuntu feisty

• The Binary Installer is the best I’ve seen for linux. Great install and uninstall
• You need the installation guide and the developer faq and the little setup guide to make much progress
• There a quite a few non obvious interface quirks.
• Users have multiple roles but can not assume more than one role simultanouesly
• After installing, you need to switch the Openbravo user from its Big Bazaar admin role to the System Administration role via the Person icon in the upper left.
• The updates openbravo just published to their Technical FAQ don’t seem to work.  I’ll put the directions below

Edit the following XML nodes in /webapps/openbravo/WEB-INF/web.xml

<context-param>

<param-name>ReplaceWith</param-name>

<param-value>http://servername:port/
openbravo/web</param-value>

</context-param>

Also, edit the following properties in /AppsOpenbravo/build.xml

<property name=”web.url” value=”http://servername:port/
openbravo/web“/>

<property name=”context.url” value=”http://servername:port/
openbravo“/>

Powered by ScribeFire.