While the specifics of Dr. Schwartz' contract efforts are confidential, what follows are "sanitized" descriptions of some of these projects. In each case, Schwartz delivered a combination of software, architectural documents, detail-oriented documents, and strategic recommendations.
Prior to joining Local Matters, Schwartz consulted for the company, to help them launch their newly forming Internet/Yellow Pages search services business. For this work Schwartz inteviewed company principals, developed business and technology requirements, investigated a wide range of open source and commercial technologies, and provided recommendations to the company. (After this work concluded Schwartz joined the company.)
Another company required a scalable data replication service to work with a suite
of XML-based business-to-business integration tools they had developed.
Schwartz began by meeting with technical and product personnel at the
company, after which he wrote a high-level requirements document. At
this point he performed an in-depth assesment of available data
replication technologies and approaches (spanning both commercial
products and research "conceptware"), with the goal of designing and
implementing a service that met the company's particular needs and that
incorporated lessons learned over the past 20 years in the
mostly-disjoint worlds of relational databases and IETF protocols. The
outcome of this effort was a 25,000 line Java prototype implementation
that he stress tested for several months, plus Internet Drafts submitted to the IETF
specifying the architecture and replication protocol.
Another company needed help with several problems related to a new WAN
packet switching product they were bringing to market. Very early in
this company's life Schwartz performed a strategic assessment of the
marketplace, to help them sort through a dozen possible product spaces
they could go after with their product. After the company had become
more established, Schwartz helped with a variety of technical problems.
First, he assembled a test environment that allowed them to emulate
packet loss rates and transmission latencies typical of WANs, and
measured the performance of their product against replayed packet traces
gathered from one of the large Internet portal sites. Schwartz also
used this testbed to help the company assess a commercial WAN emulation
system they were considering buying, and along the way uncovered a
packet reordering problem that degraded performance in the emulator.
This discovery was sufficiently useful to the emulation company that
they discounted the purchase price for Schwartz' client company (and
fixed the uncovered problem). Next, Schwartz did work integrating and
testing the company's product with RFC 1323 support for large
bandwidth-delay product networks. The outcome of this work was
incorporated into the company's product, enhancing performance. At
this point Schwartz perfomed detailed integration testing of the
company's product with an IPSEC stack, to provide early guidance for how
the company should defend against a number of security attacks. Schwartz
also tested several dozen network attack tools, installed the
most relevant tools on the company's network, and trained
engineers in their use, to help the company assess several other
product security issues. Schwartz also spent time investigating details of
the company's chosen embedded operating system vendor's technology.
Finally, Schwartz produced a document providing background about the
security issues the company faced, recommending and prioritizing
actions the company should take to address these problems, and providing
dozens of concrete details about issues uncovered by the tests.
Another company hired Schwartz to help them architect and implement a scalable,
high performance, database-driven Internet service. Schwartz began by mapping
out a technology strategy for the role of XML, object oriented data, and relational
databases. He then develped DTDs and a relational data model (along with an approach
for mapping between these representations), along with sample queries and use
case scenarios. He also produced a data architecture document explaining the data
representation and enumerating a set of architectural principles for the company.
He then architected and prototyped a
publish-subscribe service to support
near-realtime peer-to-peer data sharing, and
submitted an Internet
Draft to the IETF specifying the
protocol. Schwartz has since been
prototyping several parts of the data and
network architecture.
