Lecturer in Solaris 8 and Sybase to admin level.
A short intensive course for administrators to cover Solaris 8, T-SQL and Sybase on Solaris.

Course
Risk management and decision making under uncertainty.

Setting up Emerald Lab
Relocating my office and lab. Study of the source code of SunOS 5.8 (Solaris 8) Sparc and x86

Principle Consultant IP Billing Strategic Study (Product Selection Process) Wholesale Portal Provider
Design of a IP Billing architecture and, conduct an IP Billing product selection process on behalf of the client, a major Portal Wholesale Provider. This included a feasibility and requirements study for billing current and emergent technologies on a pre and post pay basis. For mobile and fixed, any-time/any-where use of streaming video, GPRS, SMS, WAP, and telephony, Evaluation of current and new billing products. Preliminary architecture development And in association the chosen product vendor selected a system integrator.

UNIX and CISCO SME/Guru
Computer Based Training Company
Acting as the resident Technical Guru, providing technical input for a series of system administration, and tuning courses across all the main flavors of UNIX (AT&T/NCR, Solaris, HP-UX, , AIX, Digital, and SCO) , CISCO to CCNE level and LANs plus WAN design, operation and management. Final QA/verification of the technical content of the finished products. Used Skillbuilder.

System, and Performance Technical Management Consultant
Young European Telecommunications Provider (Telephony, Broadband and ISP).
As a managing engineer involved in the re-engineering, and architectural shifts of a young European Telco. Extensive business process evaluation took place as part of investigating the performance problems experienced with the Customer Care and Billing system, for yet another new and fast growing Telco. The problems were caused by the system integrator having mis-configured and under sized the system, as well as by better than forecast traffic and new connections, (again in part down to the integrator's "experts"). This resulted in an inability to bill customers. There were also problems with the accuracy of the bills due to a lack of QC of the raw billing data.
Initially the billing operation, based on Kenan Arbor Wireline package, with a Sybase 11.0.3 database, was running on a HP K clustered hardware. There was a lack of management infrastructure, including written procedures, with the production database unprotected, no dedicated DBA, no separation of development, testing and production environments. And Y2K looming in the distance.
During the last 12 months, Billing has been moved to HP N class machines and 15TB EMC raid system. A Storagetech, robotic Tape Library was also commissioned. The EMC produced more than a few teething problems including Sybase error 605, the loss of the main database page chain integrity at one point. As this database is now 200GB, and growing at around 20GB/month.
Extensive monitoring and tuning of the whole billing system (Arbor/BP), including a number of reviews of the problems and possible solutions and the impacts of those implemented were undertaken. The target being to produce "good" bills and catch-up with around 6 months backlog .
While the general technical performance improvements were successful The lack of O&M procedures, change control procedures, and good management infrastructure. Slowed down and in some cases frustrated the implementation of other mainly non-technical changes, that would have improved performance, measured in terms of bills produced.
The lack of infrastructure, also impacted the rate of database growth,due to the inability to guide and rate many CDRs initially. Much of this was the legacy of bad decisions by the system integrator as well as a quagmire of internal politics. The HP tools Glance+ and PerfMon as well as a whole range of special monitors (designed by me) were used to collect and analyze the data.
Throughout the later part of 1999, the Y2K problem, and its solution, influenced much of this work. In the middle of 1999, once the billing "catch-up" was running. A 6-week study was undertaken, in parallel with the work on billing throughput. This was to investigate performance problems reported by users of the "new customer entry system"(provisioning). The principle tools apart from the special monitors developed for the billing analysis was HP Advisors with WAN and LAN interfaces. The network switches were Cisco 7000 series, with 100Mb/s drops, the WAN was T1, there was interconnection to the ISP provisioning through two firewalls (running Firewall/1 on a couple of E1000s, proxy server and various Cisco routers. The report produced, stated the network was "clean" although some connection reject storms were observed from the database server. The problems were in the customer entry system design and implementation. It needed to be scrapped and rebuilt as it was a totally un-professional job and any attempt to fix it would be a waste of effort and money. Despite a bit of a political storm, my views prevailed, I was thanked for my efforts, and the rebuild project is underway. We selected Powerbuilder 7 and Jaguar CTS which together with a major upgrade to Sybase 12, will provide a propely managed, maintainable application and provide addaquate performance. .
There was also some consultancy in helping to manage the testing of a TNM system, developed by the Swiss company FROX, running on Sun E6000 and Solaris 2.6 This suffered from response times well outside the spec., and other major design faults for which it was reject.
The switches were a mixture of Siemans, Ericsson and Nokia running ATM/FameRelay over spare fiber capacity from the national power company's network. The different ways in which CDRs are packaged by the different makes of switches, and the differences in their response to a request for lost or corrupted CDR files were some of the causes of the non-technical associated with bill production such as lost or duplicate CDRs in the billing database.
Additional consultancy interleaved with managing the billing production, was provided to a team developing a data warehouse system (DSS), apart from system performance aspects there was support for data migration performance. Provide additional constancy for the configuration and deployment of BMC Patrol- Best/1. and for the evaluation, selection and deployment of Rational Rose and Clearcase across the whole organization.
Postscript: The case of each of the above pieces of constancy , they were team efforts were I either managed existing teams or created a new task specific team, although in each case I took the overall technical and managerial responsibility to complete the work, in budget and as quickly as feasible . The teams ranged from 4 people (TNM and provisioning test/evaluation) up to the billing team of 20+ people.

Y2K and Performance Technical Management Consultant
Internal Infrastructure Major International Telecommunication equipment Manufacturer.
Provided advice and assistance during the investigation of the impacts of the Y2K design fault on the Clients systems. Investigation of the effects of the projected future loads created by the growth in the number of users and newer faster versions of application software across all machines, PCs, Workstations, servers, web intranet servers and the growth in http related traffic, high-availability clusters and worldwide networks. The internal networks provide connectivity on a worldwide for a very large user community and range from TP Ethernet mainly based on Cisco kit to TI/EI links between major sites. An additional requirement is the creation of an integrated, highly performant and cost effective solution to existing, and projected performance problems. The direction of the process of translating these recommendations into a full design leading onward to the bid process and through to selection of a vendor. This includes the detailed sizing of the servers, CPU, Disk and I/O requirements based on the final: Oracle and Sybase Database designs, http projections, networks and user machines be they PC or (Sun) workstation. The current setup includes: Sun SPARC and Ultra WorkstationsU1, 2, 5, 10, 20 with many IPXs used as print-servers, running Solaris 2.4, 2.5.2/3 and 2.6; AUSPEX servers HP9000s running HP-UX 10.10, 10.20 and 11 The HP servers are used mainly for accounting and commercial applications such as ledgers and production control running a mixture of Oracle 7 and 8 as well as Sybase 11. There were around 200 major databases with an average size of 75 Gbytes. There were in excess of 3000 printers (mainly HP Laser Printers but some large A0 plotters, and high-speed band printers, mostly) primarily with network cards but come of IPXs and net-extend boxes. Currently the intranet servers are two E2000s.

Technical Management Consultant Telecommunication Service Provider.
Investigation of performance problems experienced with the Customer Care and Billing system by a new and fast growing Telco. They were running a combination of Kenan and Vantive using a pair of Sun UE6000s in high availability configuration for billing. In addition, a pair of HA configured UE4000s for customer care. Both configurations were running Sybase 11 with a main table of approximately 140 Gbytes, Solaris 2.5.1 and a variety of other supporting Sun products such as Netbackup. Due to the speed of installation (12 weeks) and better than forecasted rate of new connections the whole setup was found lacked adequate resources in terms of CPU power, disk i/o or capacity. A total redesign from the top down was recommended. In addition, there were major divergences from the design methodologies recommended by Kenan, Vantive, and Sybase. Part of the Upgrade plan was to move to UE10000s.

Technical Integration Manager Telecommunication Service Provider.
To manage the managers of a multi-project revamp of a major GMS system in the area of back office Customer Care, systems. Covering the assurance of the technical integration aspects across the whole program (approx. $300 Million). Work including various OFTEL and BATA requirements with respect to Number of Digits per mobile phone number (mobile number portability), format of the call records, portability of numbers between service providers as well as Year 2000 compliance. This involved liaison with the development teams. Systems were a very large IBM MVS/ESA system, ABS 21.0 with mods, VAX-VMS, DEC VAX DEC Alpha, Kilostream, Frame Relay, Sun Ultra 3000s, Solaris 2.6, CISCO routers, Firewalls, an number of Oracle DBs around 200 Gb each, IBM Mainframe, SNA, DECnet, TCP, IP, Telnet, DFT, DEC Mail, Network security including Firewalls, Scanners and Document Image storage. The GMS system has a rated volume of 2.6 Million calls per day, based on Ericsson AXE10 transit switches and Nokia kit. A mix of Microsoft Project, PMW, PRINCE and ISO9001 methods.

Initial Phase of a project to build a Language Translator Module
This involved canvassing financial support from a European Government and the Minor Languages section of the EU. A preliminary market review, feasibility study, business plan and project plan was creation. This is now in a state of suspended animation awaiting a decision. Based on current experience it may not reach an outcome until late 1998 early 1999.