The ZAPMAIL PROJECT
I first heard about ZapMail from John Schwarzmann somewhere in the 1982 timeframe. He had been travelling around Europe with Vince Fagin, looking at the opportunities to expand a planned facsimile network there.
Zapmail(Code named GEMINI) was to be an electronic delivery service using high quality fax machines. The Overnite Letter launch had been successful, but FWS envisioned that this product and all business mail might one day be impacted by electronic mail. Although networking costs were then high, a new generation of satellites were being developed that would enable small dishes at business offices. Thus, the cost of transmission at some point would be very economical in the future.
Marketing forecasts showed that customers were willing to pay $35 for a 10 page document which could be transmitted in a couple hours. The document would be so close in quality to the original, that legally, would be a legal document.
When FedEx was first launched, there were too few cities served, which limited what packages could be shipped.
It was decided that FedEx would launch this service full blown, and first build the market, and in parallel work on getting a low cost satellite solution to solve the communications costs issue.
Direction
| ZapMail I |
Pick up customer document, take it station & transmit to remote station, deliver it to Customer (1200 Fax machines required); 2 hours total time |
| ZapMail II |
Sell/lease lower cost zapmail machines directly to customers; transmission time 1 hour or less |
| ZapMail Satellite |
Replace existing telephone line technology with Satellite communications |
| |
|
Management
Chuck Winston was brought in as SVP Telecom to manage the project. He brought with him Chris Pinkney who became the project manager for the overall project, and Ace McInturff who would manage developing the Zapmail facsimile machines.
NEC (Nippon Electric Company) was chosen to develop both the Zapmail I and Zapmail II machines.
The plan was to launch a full service. A full sales force, central and remote maintenance service, and central customer service department. This eventually grew to over 3,000 people.
Network( Voice Swith or Packet Switch)
Before my involvement, there were discussions on how best to provide communications for Zapmail.
One method was to just let the Zapmail machines dial each other, especially for low volume customers. The voice engineering group had decided that they couldn't provide a solution for cross country telecommunications. The existing pbx and acd voice switches couldn't handle 56 kilobits at that time, and there wasn't a voice system that could multiplex multiple transmissions to a station with high volumes.
Early on it was decided that a X.25 Packet Swith processor was needed. These could handle multiple transmissions. A company called AMNET out of Boston was chosen and 3 systems (nodes) ordered. Amnet had installed a Network switch for the Swedish Air Force.
Requirements were for each system(called a node) to each handle hundreds Zapmail machines at speeds of 12kbps dial,, 14.4kbps & 56kpbs leased, and 56-224kbps trunk lines(trunks are circuits that just connect the nodes together),
New Department(Packet Switching)
To test and deploy these packet switching nodes, and develop interfaces to the Zapmail machines, a new department was formed.
John Schwarzmann was made Managing Director, and I was transferred to John's department which consisted of only one other person, Gary Ragsdale. Gary had previously been the Manager of Voice Engineering. John reported to Chuck Winston. Chuck also hired Jim Colson as an Managing Director to set up an operations and field service department.
John, later in the year, hired Kurt Kunzel to build an MIS group, and from Bell Labs he brought in Don Wallace and Brad Simmons, and later Gary Holmes from DADS development.
Satellite Systems
Parallel to Zapmail development and deployment, there was a group called Advanced Projects. Alan McArtor was made VP, Satellite Systems. His charter was to develop short and long range satellite systems that would eventually provide lower costs to connect the Packet Switch nodes, and eventually dishes that could be put on the top of customers buildings. These systems were to follow the initial Zapmail rollout by a few years.
Alan hired Gil Mook for Engineering and Jim McKinney to build an Operations group. Gil hired Jim Whitworth to manage engineering and work with satellite vendors. Don Tackett was hired for Project Management.
My Kingdom for a Packet Switch
Soon after coming aboard, I took a trip to Tampa to do a project review with AMNET. Paradyne Corporation was based there, and Paradyne was OEMing the equipment; providing sales and service.
I became suspicious after my review. Going thru the plans, looking at the switch in operation, and going thru the circuit diagrams, I could not figure out how this hardware could meet our needs.
The system consisted of 3 different types of electronic boards. None of the boards had interrupt logic on the communications chips. What this meant that there was no way this switch could handle speeds over 12 kilobits per second, much less 224 kilobits per second, that were required.
Chief Engineer
Upon returning home, I called the Chief Engineer of AMNET with a list of questions, just to find out how they could be doing high speed communications for the Swedish Air Force.
The chief engineer was very honest, and the answer was, 'they hadn't done it for the Air Force, but they had plans to do it'. The Swedish Air Force system was in underground bunkers and was for emergencies, and they only required 1200 bits per second.
Solemn Day
I relayed this info to John and Gary. This system wouldn't work. John set up a emergency review with Amnet, in Tampa; and I was told, I might need to move to Tampa for several months to salvage the project.
Tampa
Long story, short, in a high level Tampa meeting, Amnet went over their plans to redevelop the system with us. Their engineers had never designed systems with interrupts or DMA; which was needed to handle high speed communications. They presented their plans to redesign all 3 boards...but again without the level of chip design required. Paradyne asked John if they could go ahead and ship the 3 systems,,,even though they wouldn't work, so that Paradyne could make a sales announcement....John refused this request.
No Switch/Enter Tandem
We had no switch and needed one. John started a review of the known manufacturers and we met with them all. AT&T(USA), NEC(Japan), Siemens(Germany) and others. All could do low speed, but not high speed and none could handle the high volumes of data, that these facsimile machines would generate.
Eventually, Tandem Corporation proposed a new plan. Tandem built business computers, not packet switches. Their computers were known as NONSTOP computers, because the configuration consisteed of 2-16 processor boards, dual disks, and dual memory. As long as the computers had power, the application would simply switch to another processor or disk in case of failures.
Tandem had initially been eliminated from competion before Amnet was chosen, because of cost. Tandem did have X.25 protocols, and they could handle transmission speeds up to 224 kilobits per second, enough to interface with satellites eventually.
Their proposal was simple. Make the length of the packets larger which would greatly reduce the costs. Initially the packet size(how many characters are transmitted at once) was 256. This small size helped with phone lines which had high error rates, but it also required many more processor boards to handle the switching. If the packet size could go to 1024 bytes, this would work for FedEx. Making the packet of buffer size larger greatly reduced the number of CPU's in each system, and thus the cost.
This was tested successfully. A contract specification was quickly written that required Tandem to develop an operations interace, and a Broadcast application, and an admin system (that other packet switches provided)....12 nodes were eventually ordered and a test system.
Tandem provided a core development team of Daryl Boggs, Dan Lomar, and Ken Smith.
One Year Before Launch: Chuck Winston was SVP over Telcom and Zapmail.
Reporting to him for the Zapmail Project were:
| |
Chris Pinkney managed the overall project plan |
| |
John Schwarzmann developing Packet Switch Network |
| |
Ace McInturff developing Zapmail I and II |
| |
Jim Colson developing Ops Center and managing rollout and 600 person tech force. |
In parallel,
Marketing under Carl Williams was developing forecast and pricing plans.
Bill Razzouk was developing a 600 person salesforce.
Kathy Crockett was developing a mini call center (ZapCentral).
| |
Memories from Sherry Taylor:
I began working in 1983 in Charlie Geronamo's organization during college and then moved over to work for Vernita Rodgers during the start up of Zapmail. My husband, who is still with Fedex, and I were discussing the "old days" just last week. I was explaining to him the energy and the teamwork at ZapCentral. At the time, because Fedex was my first job and I was 21 years old, I don't think I really looked at Zapmail as any thing other than a fun job. It is fun though to reflect back on that era and realize what we were a part of. I was just telling my husband about how we practically lived at 2955 Democrat Rd and they bribed us with food all the time. How funny to read the recollections on Fedex Legends about bringing meals in everyday to keep people working long hours. I was so excited to stumble on this website and take a trip down memory lane. I worked with or was acquainted with so many people that are mentioned on this site. I left Democrat to work in Razzouks shop as a sales rep for the dedicated Zapmail machines in 1985 or 86. Once that was discontinued we sold the Powership. In 1988 I moved to NYC as an account executive. I retired in 1991 after my son was born. My husband is currently working in sales with the new division of Tech Connect. I had an opportunity to tour the facility recently and was blown away how far we have progressed. The Tech Connect division is a reminder of Zapmail in so many ways as they are just trying to get it off the ground. I forwarded the site to him as I know several of the tech guys he works with have been around since the early 80's. |
Winn Stephenson & Jim Moore, were rapidly expanding DADS, so that couriers could be dispatch to pick up and deliver documents.
Colorado Springs Development was developing overall dispatching, tracking and admin systems.
The Memphis Billing group was developing a system to bill ZAPMAIL customers.
Several Months before Launch
Brad Simmons and I were on the Corporate jet with Chuck Winston who was over the overall project. He asked us where we were. I told him, I thought we needed more time to launch. The Zapmail machines had dozens of bugs, as did the Packet Network. Everyone was working on stabilizing the system, but it would better to have more time and make it rock solid. Chuck paused looked at us and said something like(paraphrased)......"Good, then I'm going to accelerate the rollout ,if we can and speed up the expansion......". We were shocked more than a bit. Instead of slowing down, he wanted to speed up.
He said, 'you are engineers and it's your objective to find problems and fix them, and everyone is making progress....my objective is to build this system, spend the money and have it available, so that no other company will dare try to copy our business model. we are going to build a business before any competion can even start to compete with us...".
Rollout
About three months before launch, only 8 of the 1200 Zapmail machines had been activated. Jim Colson then made some management changes and put everyone in his department on a 7 day work week. He brought in breakfast,lunch and dinner(This was only done in Memphis at 2955 Republican). Three months later, almost all the Zapmailers were installed and working.
(Note * Trying to verify if 8 is the correct number above. Some on the field installation team remember the installations being much higher that this. If you were there, send in what you remember)
The Network had been rolled out consisting of 12 major packet switch nodes.
Launch
On the day of the launch, Jim Barksdale came to ZapCentral to send the first Zapmail document. It was to Fred Smith who was, I think, in another city. It contained a memo congratulations to FWS on the service launch.
Only 50 documents were transmitted that day. Over a thousand was forecast. Even though volumes eventually reached 55,000 a day, the forecasts were never accurate.
Another innovative document was transmitted by a recent lady graduate. She transmitted a Zapmail document to Fred Smith, with a congratulations memo, and her resume. This was proactive and innovative.
Zapmail II
The second phase of Zapmail required lower cost Zapmail machines that customers would have on their premises. They could either purchase of lease these. It eliminated the requirement to call for a courier to pickup a document, or to deliver it if someone else was sending you one. These machines connected to the packet network by either a dialup 12kbps phone line, or a dedicated 14.4kpbs dedicated phone line or a 56kbps dedicated circuit.
Many were sold with dial lines which took a couple minutes per page for transmission. If you had a 20 page document to send, or receive, it could take up to an hour depending on the number of bits on the page. From a queuing and modelling point of view, the dial up line wasn't effective for customers that received or transmitted 4-5 large documents a day.
Zapmail Expansion
At it's peak, there were about 1200+ zapmail machines at FedEx stations, plus probably hundreds used internally in FedEx offices. There were about 12,000 customer machines sold(but not all of them had been installed before the end).
Jim Vaughn was brought in as a VP for Operations. Mary Alice Taylor was made a VP of Logistics for Zapmail and the rest of FedEx. Six hundred salespeople, six hundred remote technical service people, 100 customer service reps, and probably 100 for Operations.
TRANSFER
At this point in history, I had been working on Zapmail for about 1.5 years; the future satellite system was becoming more important. John Schwarzmann, Brad Simmons and myself were transferred to the Satellite Systems group to work on the next generation systems. We were initially located at Ridgelake. John transferred back to Zapmail shortly after this first transfer and continued to work with Bob May, who was now the SVP of Zapmail. Chuck Winston had moved out of the company. Bob had the distinction of having 'ZAPCZAR' as his front license plate.
ISTN
The satellite group worked with Tandem and M/A-Com to provide a satellite for communications to the major Packet Swith nodes. This was called the ISTN system.
SATURN
As the number of customer premise zapmail systems grew a project called SATURN. Up to 100 smaller Tandem systems would be located closer to the customers to reduce communications costs. The Zapmail II machines would communicate locally to the Tandem. At this point in history, COSMOS IIb was being planned which would scan packages with the SuperTracker hand held device. These Tandem boxes would also be used to transfer these scans to COSMOS.
Another Satellite system was purchased which would provide satellite communications to these 100 notes. Each dish at a site would provide two 56kbps circuits.
Future Plans
I worked for Jim Whitworth, and continued engineering and planning for the satellite system that FWS first envisioned that would make Zapmil economical.
This system would provide a small satellite dish on a building. To reduce costs, up to 4 customers or 4 machines could be connected to each dish. Target price was around $12,000 (or $3,000 per customer). A central location would house a new packet switch system which was based on high speed LAN connectivity. It would provide a central switch of all Zapmail documents.
A contract was signed with Hughes to begin work on the large satellite which was to be the size of a bus. A launch slot was planned on the Space Shuttle even though the satellite wouldn't be ready for years.
The Ultimate End
January 28, 1986 the Space Shuttle Challenger exploded. All shuttles were grounded. It was uncertain when or if they would ever launch again.
By this time, customers were now able to purchase Group III fax machines, which hadn't been available or affordable at the start of Zapmail. Now for low volumes, these machines were much cheaper to transmit directly to other customers directly. You couldn't transmit or receive documents simultaneously, but this wasn't a requirement for most users.
FedEx looked at many alternatives. A central queuing system that would allow customers to electronically make an appointment to connect to another machine.
Another option was to add in a Group III board, let customers just dial each other and other off the shelf facsimile machines.
Also considered was a Business only network, where maybe 10-20 thousand business customers would connect into the packet network, only by dedicated circuits. They would pay for the circuit and 25 cents a page. The systems would be closed between those companies, and costs kept to a minimum until a Satellite solution could be developed. Before the Challenger disaster, only about 2000 of the Zapmailer II machines had been sold or leased. An interim decision was made that FedEx would provide a business only network. There would be no dial lines supported. This meant that the price of the Zapmailer II machines and use would go up. Bill Razzouk was able to quickly sell 8000 more systems in the next few weeks, which were dial only machines. An upgrade price would then be offered to connect them directly to the network.
In the end, with hundreds of millions sunk into the project and no clear plan for the future, it was decided to shut down Zapmail in October 1986. Wikipedia reports that $320M was written off.
Shutdown
Dennis Jones was chosen to do a financial analysis of the shutdown and came to initially talk to me about costs. I relayed to him the people was the big item to address. Around $10M had been spend on packet switch hardware, and hardware could be sold. Another $10M on communications yearly costs, which could simply be cancelled. But the real costs are people, which numbered over 3000.
All FedEx departments were told to do five year headcount plans. Based on these numbers, zapmail employees were interviewed and placed into other departments. The six hundred zapmail salespeople became package sales employees. The field service technicians were kept to support the growing number of devices in the field including terminals, phones, radios, SuperTrackers and the ever expanding Customer PowerShip computers.
FedEx took care of the employees, and almost every person was placed in a new department.
People
No one likes to work on projects that fail, but the employees who worked on the project were fully dedicated.
And while all these employees were trying to make the Zapmail system work; other FedEx divisions continued keeping FedEx running, expanding and keeping the main business going. They were also dedicated to the company.
FXTV
When major announcements are made at FedEX, FWS likes to tell employees first, whenever possible. The company had gotten so large, that it could take a week just to get around to the 40+ buildings, just in Memphis.
While announcing the Zapmail shutdown, FWS said something like, you know, we really need to have some type broadcast capability so I can talk to all the employees at once,,,and maybe we can use it for training.
Thus began FXTV. With some of the transponder space already leased, FXTV was started and Memphis executives could talk with groups of employees around the country. It expanded to remote training programs, and even an update news type program. Management could now schedule broadcasts to every location and communicate regularly with staff.
New Focus
With the Zapmail shutdown and write off, the company turned to International, Flying Tigers, Cosmos Squared and other endeavors. Some of the Zapmailers were kept for a few years as fax machines. The other 13,000 machines were sold to scrappers. The printers were all standard, and put into Laser Printers. And there was a DRAM shortage so all these memory chips were recycled into other devices.
Half Dozen of the Tandem processors were used for International and Powership programs. The next generation satellite system was redesigned on paper to be a common network for voice and data for the company, but technology had advanced to the point where it wasn't a viable option.
FedEx Networks under Winn Stephenson began a new strategy of buying T-1 circuits and multiplexors such as N.E.T and Timplex to build a new backbone system using the major call centers as nodes. This provided great cost savings providing call overflow between call centers as well transmission of all COSMOS, DADS and scan data.
Tidbits
In the end, the max daily volume was 55,000 zapmails. A small percentage of these were internal comat documents.
Around 5,000 of these documents originated in Washington DC and sent to another office in Washington DC.
About 12,000 of these documents were broadcasts. Broadcasts were when large customers sent the same document to 1-100 people at a time.
From Clayton Roderick:
Zapmail2 this was the Tandem control center 2955 republican – after Zapmail died.. We were all looking for jobs.
Zapmail-day ---… Working at Zap Central ------ before moving into the Zap control center
Remember the power cord incident in 2828 – That manager “Jim Somebody” was trying to make a big deal about it.. :)
I got in a little trouble one time when I was working in the Zap control room, one of the Zap Central girls called in and stated that a customer was on line and complaining about the Zap Machine transmitting too slow – I told the agent jokingly to tell the customer that there were too many birds on the line and have the customer to go out and shew them away – well she told the customer that and I got in a little trouble…. I didn’t think she would relay that to the customer…..
|
