Early Computing Environment in 1974 (Interview with Chuck Graham)
In 1974 the head of the IT department was Leon Tyree. Chuck was the 7th programmer hired by Leon Tyree and Wes Terry;
The first programmers I worked with were, Larry Brock, Jere Bledsoe, Bill Stuart, Colen Kelly Steward, Don Rose and Jimmy Fuller. Larry only worked for 2 months and Jimmy for 3 months after I was hired. We were back to five programmers for a few months before other programmers were hired.
Chuck arrived in March of 1974 when the package count was 6000 per night(on a good night), with a goal of 10,000 by December of that year.
Chuck: I started working at Quince in the old RCA Building. We then moved to Southaven, Mississippi for 4-5 months, then to the Airport hangar, then to Nonconnah.
We were using time rented on a Burrroughs 3500 , then went to B6700. We eventually had three Burrough's 6700 Mainframes.
In the very early days, Ron Ponder & Charlie Brandon would work on a computer model of routing/hubs and Chuck remembers the Burroughs 3500 would slow almost to a halt when their application ran. (Ron was a PHd from Univ of Memphis and later became CIO of FedEx).
The future didn't look bright in 1974. We lived paycheck to paycheck. In 1977-78 we got our first Boeing 727 with deregulation.
Who made a real difference in the early years?
Pete Wilmott. He came in 1974 and the company was in financial trouble. He negotiated a $50M loan with his connections. He also brought financial accountability. Art Bass was an original consultant and was instrumental in getting compliance up. It started at 30% and went to 99%.
What made FedEx successful? We worked hard and did what was required. People and their drive, we bled purple, we had will and desire to make the company successful; and Fred was the leader in Innovation.
What else was innovative and made FedEx successful? The airbill was innovative. It was attached to the invoice and sent to customer service for billing and onto the customer. We called this "Country Club Billing".
What are you most proud of? Working on the Billing Systems, origin-destination and back to origin billing; and my role in Imaging and DCI at FedEx. Chuck wrote the first pilots payroll program with David Kasper(a pilot); parts of that 1976 program is still running in the company;
It began with 80 column card input....
The Revenue department would key info from the airbill to punch cards for batch billing processing
Leon was a MD and Wes was our manager. Together they hired us worker bees. Stew was between jobs when he hired on and actually started a couple of weeks before anyone else. During that time he worked from home and drove to Little Rock a few times to meet with Al Garcia to begin defining what became known as the Component Tracking System. It wasn't fully implement until later but I'd have to say that was the first system to begin development and Stew wrote the first line of code for it. In Cobol to run on the Burroughs 2700/3500. That and all systems to follow were done in Memphis.
The second system to begin development was of course, the billing system. We did have hopes we'd eventually have customers to bill. At that time we didn't have fancy things like desks, chairs and offices so for several days in February, '73, we all met in Carl William's den and began defining/designing the first billing system. It would be inappropriate and impractical to say who created that system. It was purely a team effort. Between Larry Brock, Bill Stuart, Don Rose and myself, we each developed our slice of the pie. Then the whole pie was put together, tested and implemented as one system. Forgive the pie analogy. For me, I wrote the programs that did the pricing. Bill, Larry, and Don did the rest along with Chuck Graham who came aboard soon thereafter. When my part was done I was sent to work with Mike Staunton to begin developing the Flight Operation System (FLOPS, tongue in cheek). That was the great granddaddy of the FORTE system and is another whole story.
All of the core systems that supported the growth of FedEx the first 4-5 years were built by a hand full of IT people. I don't think we had over 15 programmers during that entire period. All systems were built in-house except the payroll system was purchased and modified by internal staff. Each system has a story of its own. Those were busy and rewarding times. Don, I think that's all right. But as you well know, that was a several moons ago. So, fill in whatever blanks I left out.
Jere. When you were talking about no desks, chairs, etc I thought you were going to tell about us stacking up computer paper boxes 2 high for chairs and 3 high for desks. That is how we coded in the early days. Before we got our first computer, all of our car trunks were like traveling computer rooms with trays of punch cards, boxes of paper, and mag tapes everywhere. Those were the days for sure.
Yeah Don, that paints another picture of the dawn of ITD. As I recall, we had to go to Carl's house as I mentioned because we didn't have heat in Hanger 7. In February it was pretty cold in those WWII hangers.
Yes it was cold. I remember when they finally got the boiler fired up and we had heat (it was steam heat for those that weren't there). It was great to be able to go to the restroom without having to break ice in the toilet.
Transition to IBM, and the Cook Story
In the middle of the 70's and with deregulation, FedEx was growing exponentially.
Most systems were batch oriented with the initial developers concentrating on billing a customer, payroll and pilot flight rules to start with.
Message switch was the system used to essentially 'text' or email stations to find out about package delivery or billing info.
Customers called the local station for pickups.
FedEx needed larger capacities for processing and to convert to real time applications.
Different Options for Mainframes
One concept was to string 10+ Digital Equipment Systems together on lans and continue to grow with smaller systems.
Some wanted the largest Burroughs system called the B7700; Data Communications in Memphis handled processing for over 500 TV and radio stations around the country. DCC, where I had worked, had 2 B6700's and a Honeywell Multics system. DCC had been able to get a B7700 in only 3 months; and FedEx wanted to pressure Burroughs to get them one quickly, to cover growth.
Another group wanted IBM.
Ancel Hanks told the story of Charlie Brandon who was essentially the CIO of FedEx going to visit the CEO of Burroughs. FedEx package volume and thus processing needs were growing and larger systems were needed. Charlie asked the CEO to deliver a Burroughs 7600 in the next 3-6 months. This was the largest system Burroughs made and was in use by many banks around the world. DCC in Memphis had just taken delivery of a B7600 and got it quickly due to a cancellation of another order. Lead times on these systems were a year or more.
The Burroughs CEO did not exactly show any respect to Charlie or FedEx and offered no hope of getting a system.
Charlie came back to Memphis and told Kathy Crockett then over the Call Center to plan on getting off the Burroughs dispatch program in 6 months because they were changing over. He left it to her to find something to move to.
How long did Charlie give Kathy to get off the Burroughs?
Ancel:I think it was maybe 6 months. There was a contract with a minicomputer development company. Don't remember the name. Eventually the contract was cancelled due to non-performance. She and Tollefson reached an agreement on a skinnied down system to sustain them that could be up in 3 months. Don't remember the name. Eventually the contract was cancelled.
To compound the issue, the management team was trying to look for alternative uses for the aircraft. Aircraft sat parked on the ground all day long in remote cities. One use would be to fly passengers during the day.To do this it would need new applications that could handle booking passenger reservations.
(In one story, an older lady drove to work past the Airport at Newark each day. Each day she saw the FedEx aircraft parked as she went to and from work. She thought it was broke & she finally called FedEx and wanted to know..."When are you all going to fix that plane!! It is an eyesore " )
Freight By Night/Passengers by Day
One idea was to fly passengers on direct routes during the day, and then fly packages at night. The company ordered five 737's with seats on rollers that could be quickly removed and convert the aircraft to freight. It also applied for a license to fly passengers out of the Midway airport next to Chicago.
To fly passengers, the company would need a passenger reservation system. Airlines used the ACP(Airline Control Program) operating system which usually ran on IBMs. To hire programmers to work on an ACP system, you had to be in Denver where most airline computing was based.
The story goes, with the need of more processing power, FedEX told Burroughs 'get us a B7700 in 6 months or we are going to IBM by the end of the year'. This was the newest and largest mainframe that Burroughs built. The problem was that there was an 18 month wait on these systems. DCC in Memphis had only been able to get one in 3 months, because a bank in France had cancelled their order, and DCC was able to get their delivery slot…..so this was not to be
New Opportunity: Cook Industries & the Russian Wheat Harvest
Cook was a large grain commodity company based in Memphis with offices around the country. It had been a very successful company buying and selling commodities, with over a dozen offices around the country. At its peak, Cook Industries was the 3rd largest grain commodity company in the world.
Cook had a computer center on the 17th floor of Clark Tower in Memphis. Jim Barksdale was the head of the IT department which also did some data processing for other clients.
The USSR had massive crop failures. It needed to purchase large quantities of grain to feed its population. The US allowed large quantities of grain to be sold to Russia and even loaned money for them to purchase grains, with the condition that the grain be shipped on U.S. ships. Cook was able to get it's share of these commodity deals.
Major profits were made with the failure of the Russian grain harvests thru the middle 1970's.
The story goes, that Cook management bet on the Russian grain harvests continuing to fail, and they hedged purchasing. Russian harvests eventually improved and Cook found itself in financial difficulty. Also, the Federal government was always medling in Russian purchases trying to get better prices for oil and several times stopping approved purchases from Cook and other companies.
There was also a grain scandal involving all the grain companies where grain inspectors were paid to allow lower grade grains, and lower amounts into ships to increase profits. In one week, there were almost 300 indictments across all the commodity companies. Eventually Cook Industries was fined as were the other companies, but Ned Cook was never charged with any knowledge of what had been done. It was reported that some companies sent ships over with grain mixed with rocks, and in another shipment an old car chassis was found mixed in with the grain.
Facing a $64M yearly loss, Mr. Cook & Jim Barksdale approached Fred Smith with an idea that FedEx take over the Cook IT staff. This was positive in several ways: Cook had an operational Data Center, trained IBM personnel, and a future slot for an even larger IBM mainframe.
(from Ken Pasley:The IT department was actually a separate services company called ISD, a wholly owned subsidiary of Cook Systems. Jim Barksdale was the President.It provided batch processing for several other companies. These services were discontinued shortly after Cook personnel moved to FedEx)
Who was involved in the FedEx buyout of Cook Data Systems? Some feel it was all done by Mr Cook working with FWS. I thought JLB was the person that orchestrated the deal from Cook and FWS was the FedEx person?
It was actually orchestrated by Pete Willmott and Mr. B. They went to church together and evidently discussed it during church because that Sunday he called me and asked if I’d meet him at the office…..and that began the sale. The main lawyer who worked so hard on the deal was named John Coralew (sp). He was really, really young looking but brilliant! He died less than a year later after the sale of pancreatic cancer. He died way too young!
My favorite memory of the whole “sale” ordeal, was the FedEx and Cook people sitting in a room trying to decide over the number of office space that FedEx would assume…and the cost per foot. Mr. B was there, along with Howie Bedford, the lawyers. Anyway, everyone was trying to use their calculators to come up with the right numbers and it was Beford, using pencil and paper that calculated the whole thing before anyone could finish using their little machines. It blew me away!
At the time, Cook had just created a spinoff called Cook Data Systems, similar to FedEx Services. Mr Barksdale was the president. There were right at 100 employees and we had very good delivery positions on the most powerful IBM mainframes.
Federal Express wanted both the big iron (delivery of new mainframes were over a year at that time) and a built in organization that could hit the ground running. Many of the original Cook staff ended up in key leadership positions at FedEx, including Jim Barksdale as COO.
Employees who worked at Cook Industries & Federal Express:
Winn provided the following list of employees of Cook who either came over with the purchase of Cook IT or anytime later.
Floyd Goodman Delk
John Etta Parker
Linda Kay Wright
Ann Mason Holman
New Data Center, New Staff, New CIO
FWS immediately saw the strategic value of doing this, and Cook's proposal was accepted. Cook IT moved to Federal Express and IT named Federal Express Data Systems. This transaction gave FedEx a new operations Data Center, IBM personnel, an IBM system, and a future larger IBM computer slot. About 40 people moved to FedEx, and shortly after this Jim Barksdale moved to take over all IT from Charlie Brandon. IT complexity and needs were growing and Charlie had decided he did not want to manage a large IT shop.
Cook then sold its grain merchandising business to Pillsbury in 1976. (FedEx did do processing for Cook for over a year after this move-Jimmy Sowell)
(Jim Barksdale mentioned this was a great deal for FedEx. They not only got seasoned IBM personnel, a great slot for a future large IBM system, but they also received payments for doing processing for some of Cook's clients. FedEx's $1M or 2M purchase price was quickly recouped by FedEx)
Jim was initially SVP of Data Systems, He was promoted COO of all FedEx around 1982-83 approximately. (Barksdale Quotes)
Reporting directly to Jim were:
VP Systems Development Memphis
Director of Telecommunications
Interview with Jim Barksdale on FedEx IT History: June 5, 2011
Part I -Interview with Jim Barksdale
Part II - Interview with Jim Barksdale
Interview w/ Jimmy Sowell
who transferred with the first wave of Cook personnel to FedEx
Jimmy Sowell(from video) on Jim Barksdale coming to FedEx:
"....Then Barksdale came over, things started getting more focus".
" That is absolutely the truth. He got things going, was very organized, and we were off..."
An Interview with
Hired: 1977 by Frank Smith
New COS Development Center and using ACP
Before the Cook computing move, Tucker Taylor walked into Charlie Brandon's office. He was fascinated by the new Wizard of Avis system, where you could make one phone call and they could reserve exactly the car you wanted, when you wanted it. Avis customer agents had all the information needed at their computer terminal. And when you called back, no matter who you talked to, they had your information. Tucker asked Charlie, "why can't we have an Avis system for packages?"
Taylor and others had envisioned a system that would track pickup reservations to Customers and eventually track packages thru the system to the destination. It would be something like a reservation system for packages. This had never been done before, but the reservation systems software that airlines used was similar. A prototype was developed called Project Sydney(Sydney was Tucker Taylor's first name) on the Burroughs B6700 which was successful. Instead of calling the local stations, the customer calls were routed to a mini call center which took customer information and a test system routed a Dispatch message to the local station. This was first tested in Newark, with a rollout to 11 other sites. To accomplish this ACD's were used (Automatic Call Distributors) and Rockwell Collins was the vendor which was selected. The test software ran on the Burroughs 6700.
(Tom Bullion: I was working on a tracking system (became COSMOS but at the time we called it DIC-AIDS (digital information center and automated information and dispatch system – Smith loved it and wrote COSMOS on a napkin – I ask him what that stood for and he said to make something up – Customer Oriented Service and Management Operating System (Marketing changed it to something in the ‘80s)
Charlie Brandon approached Howard Bedford who had developed airline reservation systems for many of the Airlines with the idea. The story goes, that Howard actually developed the Wizard of Avis. Howard was excited about the possibilities but would not move to Memphis. It was decided to start a new facility in Colorado Springs and Howard's startup team was also hired. The thought was that Colorado Springs was an excellent place to live, and hundreds of airline programmers lived not too far away in Denver(about 60 miles). This would be a good incentive in the hiring of developers from Denver, to live and work in Colorado Springs.
With the hiring of staff in Colorado, their first application was to put the COSMOS test application on an ACP system on an IBM.
I started in Colorado Springs the last week of September, 1978. The facility was first opened in the summer of 1978.
Our first project was the creation of a customer dispatching and package tracking system built on top of ACP (the Airline Control Program). The ACP based system was called COSMOS, pretty much from the beginning. We did not use any passenger reservation system (PARS) code. ACP is really just a relatively simple special purpose operating system, which for the time, enabled very high transaction processing rates. All of the code was written in IBM 370 assembly language.
Of course, the first package tracking application was pre-tracker. The agents just keyed tracking numbers from the multi-part airbills on IBM 3270 terminals.
There was a Burroughs based dispatch system that existed before we started COSMOS.
The Colorado Springs Facility would concentrate on developing a Customer Service and Dispatching system, and a system for tracking packages. This System was named COSMOS (Customer, Operations, ServiceMasterOn-lineSystem )
Memphis Development would continue to expand and enhance multiple support systems: Billing, Payroll, Accounting, Hub, Pilot & Flight Systems.
The facility opened around the 1st of August, 1978. It was called the Advanced Systems Division. When I started August 28, 1978, there were about 18 employees. Howie Bedford was the VP and was domiciled in COS, and Charles Brandon was our SVP. Hank Howell was a MD and key Managers at that time were Tom Zywicki, John Toscano, Alfred Weiniger, and Jose Armendariz. Dave Dalton, Randy Johnson, and Joe Williams came to work as Managers the first part of 1979.
Howie Bedford got fired about 4 months after the facility opened and Hank Howell became the VP. Jim Tollefson and Ron Kaufman were hired as MD's in mid-1979. Jim Barksdale became our SVP about 3 years after the facility opened. All of the employees flew to Memphis in two different groups and we all worked at the hub for a night. The first group of about 25 went one week and the other 25 a couple of weeks later.....and we all flew commercial! Before we moved to the Rockrimmon facility in April of 1982, we had expanded to 3 different buildings on Janitell Road.
Jim Barksdale's first visit to the Colorado Springs Data Center
from (Skype)Interview with Bob Higgins
Pre- COSMOS Operations
Field Operations was mostly de-centralized before COSMOS. Customers would call a local phone number. They would get a local employee at the station, who would book a dispatch(to pick up their packages) or handle any other type calls.
The information would be given to a Dispatcher at the station. The couriers would either call the station to find out where to go next, or if they had a radio, would contact the dispatcher on the radio. The dispatcher would keep up with who he gave which pickups to. There was a pigeon-hole type box where he could keep track of pickups. There was a slot for each route or driver. (See Radio/Dads for more detailed information)
In larger cities, they couldn't keep up with these calls. Sometimes there would be 10 people answering the phones, and two dispatchers handling the radio. Only about 10 percent of the dispatches were given to the courier over the radio. There wasn't enough capacity. The rest were given to the couriers whenever they could find a customer phone or pay phone to use.
Enter COSMOS & Centralized Call Centers
COSMOS coupled with a centralized Call Center and FedEx's growing networks would solve most of this problem. Later DADS would solve the congestion radio problem, and Cosmos IIa and IIb would add in the package tracking component.
First, Call Centers were rolled out to the top major cities. Voice engineering connected the call centers to local cities via FX circuits. Customers would still call a local phone number, but dedicated phone circuits would carry those phone calls to a major call center versus someone at the local station.
An agent at the call center would ask for your account number and type it in; this would fill the screen with information about that customer; the agent would ask for 1) number of packages 2)time it was ready and 3) if any supplies were needed or special pickup instructions.
By entering this info the customer was given a COSMOS #; if there number was 0341, that meant they were the 341st customer pickup call for that stations.
Automatically a message was sent to the specific station that serviced that Customer. Stations were numbered as airport identifiers. Memphis might be MEMA or MEMB, Boston would be BOSA etc.
Within seconds at the destination station, a 20 line by 80 column print out would be received with Dispatch info. These would be torn on perforated edges and given to the Dispatcher. He would put them in slots associated with that route or driver.
Interview with Bob Higgins
Benefits of COSMOS & Centralized Call Centers
This was a great advantage for the larger stations. They now had a system to keep track of dispatches and automatically route them to the stations. Customers didn't have to call the local stations, so many agents in one location could service many customers across the country. Although this solved the inbound customers calling the local station, it didn't resolved the radio congestion issue. The station still needs 10+ personnel answering the phones from couriers trying to get information on their pickup locations. (See Radio Dads)
COSMOS provided Customer Service support and a dispatch system. With the keying of destination airbill data it also provided the agents information about package delivery confirmation, times etc; Later it would add location and service information on FedEx products.
With this system in place around 1979; the COS development team continued on with the next phases of the package tracking vision. Develop a system that would keep track of packages, and monitor the flow thru the system. Initial phase was labeled COMOS IIa, then evolved to COSMOS IIb (see Scan History)
In between Cosmos IIa (scanning packages in the stations and SuperHub) and IIb (Scanning packages at the origin and destination customer locations) a system supporting Zapmail was added to COSMOS.
COSMOS & Zapmail
Zapmail was a FedEx service which would deliver an exact copy letter between customers electronically. Example costs were a 5 page letter for $35.
There were two phases of the project. To the the service started, FedEx would place 1100 Group IV Fascimile machines at station locations. The customer would call FedEx who would enter a pick up request in COSMOS. A dispatch would go to the station and be sent to the Dads termina. in the courier van. The courier would go the customer's office, pick up a Zapmail envelope, drop it off at the station. The station would scan and transmit the letter to the destination location over a packet network. COSMOS would also book the delivery courier to pick up the zapmail envelope and take to the destination customer. Total time from customer call to when it the Customer received the letter was 2 hours.
COSMOS would track this entire process and also interface to the Network, administrative messages out of the Zapmail machines and billing.
COSMOS conversion to IMS FastPath
The AirlineControlProgram Operating system was very very fast. It would handle a large quantity of small transaction. If there was a computer failure, the system could boot up quickly and start back processing transactions faster than other operating systems available.
With all the benefits that came with ACP, there were issues that grew paramount and FedEx processing needs grew:
-Applications were written in IBM assembler, which was tedious, and limited hiring programmers
-data record sizes were one size, there wasn't a flexible database
-ACP came up fast but with limited database data integrity
-it was taking too long to make changes because of ACP complexity; 6-12 months for changes
-it was uncertain that IBM would continue to support the ACP operating system or go to a new OS
At some point in the early 80's a decision was made to go to a standard Operating System. Although there were concerns that standard OS's would support our traffic volumes, it was decided that IMS Fastpath , an IBM operating system would be used.
Colorado Springs re-tooled to rewrite the systems in Cobol and a standardized IBM data base.
The first day the system came up, it died within a couple hours as more call centers logged on. Remarkably FedEx volume did not change that day, as our customers just brought us their packages(to the stations or drop boxes) if we couldn't get there before closing.
Jim Barksdale called IBM, and the next day, IBM flew in a lot of memory and added to the System. It was a sluggish start, but now FedEX COSMOS was up andsuccessfully running on IMS.
Early FedEx CIO's
SVP Operations Plng & Info Systems
Initial FedEx Billing, Pilot and Customer Systems;First Ops Research Modelling ;Started COS group which developed COSMOS;Centralized Call Center
Conversion from Burroughs to IBM
Rollout of Cosmos ; Cosmos IIa (first package scanning) & DADS (dispatch to courier terminals in vans)