The Blue Packet

Once in a while, I get e-mail asking me how the telco industry works. My usual reply is "it is far, far more complex than anything else you could ever imagine", and the comeback (if any) is usually around the lines of "but isn't it just a bunch of servers like the Internet"?

Enough deeply clueless people keep doing this that I thought it worth while to try and describe how a new mobile service is designed, implemented and launched end-to-end. All five (yes, five) years' worth of it.

What follows is entirely fictional. None of it actually happened, although in a Dilbertesque sort of way there is nothing to prove it might not actually happen, or that similar events (taken one by one) might not have a chance of occuring occasionally in the industry. It is also one of the few things on this site that I will definetly not allow to be published anywhere without my consent (cc), since it was so much fun to write that I am thinking of using it as a blueprint for a book.

Again, according to my Disclaimer, it has absolutely nothing to do with what I'm involved in. It might, however, give you an idea of how mindboggingly complex the industry actually is. Real service deployments are often much simpler than this, but require a lot more care and effort than people usually give their operators' credit for, so please take this for what it is - good-natured satire, and a glimpse at the enormous amount of very real work that takes place behind the scenes.

The Concept

The "blue packet" is a revolutionary new mobile service that consists of the following: you press a button on your phone, and the screen turns blue. Like all other mobile services, this is done by issuing a "blue packet" out of a central server and sending it through the network to your phone.

Why blue? Well, because I happen to like blue. It's a "packet" because it has to have some new technical bent, and the blue screen bit... Well, you can figure that one out yourself. This is a parody, so I might as well have some fun too.

Year Zero

  • Strategy, ever in the lookout for ways to maximise revenue (whether it be ARPU or lowered OPEX/CAPEX), identifies a broad area of new services called "tinted" packets. These are, essentially, your vanilla "grey" packets, but with some added value (i.e. color) added by the mobile operator.
  • They submit a series of presentations on the topic to Management, who ignores them until a generic revenue forecast is added. Management then sends down a memo asking for "more information on this color thing", and Strategy gets together with Marketing to allocate budget for an initial focus group consisting of a "sounding panel" of industry luminaries (analysts, consultants and a couple of random bloggers), who meet at an undisclosed (yet unbearably sunny) location and explain their vision on colored packets.
  • After recovering from the meeting, someone at Marketing culls the list of suggestions (removing non-colors like "champagne", "margherita", "bloody mary" and "bikini") and ends up with a shortlist of product concepts, which is headed by "white", "sand" and "blue".
  • A short presentation is delivered to both Management and Branding (which has become a separate entity from Marketing these days), and both "white" and "sand" are discarded due to lack of coherence with brand values and the company's mission statement. The concept is considered "good to go", and Marketing is cleared to reserve a significant portion of next years' Product Development budget for it.

Year One

  • Marketing asks Strategy for a tighter projection on the possibilities of blue packets and calls in someone from Engineering to have a look at the concept. Laughter ensues, so Marketing allocates budget to a set of consumer focus groups to try and validate its point.
  • Engineering goes off to a conference with Strategy and realises there might actually be something in this, so they spend the whole trip trying to find an actual use case scenario for blue packets. Engineering's phone crashes, and a service concept is born.
  • Meanwhile, the people in the focus groups are asked what they think of blue packets. A little old lady replies "Why, I guess they would be all right, I suppose...", and that pushes the score squarely into "Positive Feedback From Prospective Consumers".
  • Marketing presents the focus group results up to Management, and Engineering adds a draft service specification. Strategy signs off on both documents as "doable with industry support", and Management forwards the documentation to the company's representatives in standards bodies and starts talking cryptically about "painted packets" during business lunches with vendors.
  • During a routine 3GPP plenary session, a new workgroup is formed for "orthogonal hue modulation of data packets". The group's charter can be summarised as "to seek out new ways to selectively incorporate discretionary variations in bit-level representations of hue data in the mobile space". Everyone in the industry joins the workgroup to see what they're driving at.

Year Two

  • Vendors start incorporating buzzwords like "chromatographically targeted packets" into their product roadmaps and hinting that the feature might be developed sooner if operators show an interest in it. Engineering, who has to sit through all of these presentations, occasionally forwards them to Strategy and Marketing to keep them updated.
  • Marketing allocates a dedicated product manager and classifies the service as "secret". Copious amounts of documentation on service concepts, use cases and requirements start being drafted and sent to vendors "under confidence".
  • Engineering and Strategy begin taking an active role in the 3GPP discussions, and jointly define a comprehensive service architecture consisting of the CPSN (Color Packet Service Node), the CPGN (Color Packet Gateway Node), and the CPBPDN (Color Packet Button Push Detection Node). Thirty other accessory nodes and inter-node interfaces are also defined to glue these three main nodes to the existing 3GPP reference architecture, now at release 42.
  • Special consideration is given to both emergency color packets (defined as "presenting a high percentage of the primary RGB component") and usage of color packets whilst roaming. Billing interface specifications make up for over 90% of the documentation, since at this point there is no idea of how the service will actually be tariffed and vendors want to cover all the bases.
  • Someone mistakenly opens a browser instead of Word during one of the workgroup sessions, which leads to the discovery that the IETF already has a "rainbow packet" RFC. This effectively stalls the entire working group for three months. Nevertheless, Strategy manages to push for and get approved an optional "third RGB component" packet format, which Engineering then includes as a mandatory requirement in an RFI sent to all the vendors.
  • Marketing starts discussing a "special feature button" with handset manufacturers. It is generally agreed that it should be placed on the outside, and that it should emit a strong blue light at all times to encourage consumers to press it.

Year Three

  • 3GPP finally decides to one-up IETF and expand the working group's charter to cover "arbitrary length representations of color components laid out in orthogonal grids", until someone points out they're trying to standartise bitmaps.
  • Engineering and Marketing start having joint meetings with vendors eager to make available trial implementations of "green packets". Engineering presses for GA (general availability) dates, complains about the ludicrous prices quoted for trials, and points out that there is no valid business case yet, which throws Marketing into denial and forces it to churn out a product description with blue packets defined to ten decimal places in the Pantone scale.
  • Engineering simply points out that Pantone is for print media only, and forwards the spec to IT, in care of Billing.
  • Inside IT, Billing notices that there is no user profile field for blue packets and asks Systems and Provisioning if they know anything about "non-gray packets". Provisioning starts asking questions about the lifecycle of "colorful customers", and Systems (which actually reads RFCs) adds a couple of zeros to its budget to upgrade all its servers "for rainbow packet support".
  • Management notices this and asks Engineering if it's really necessary. Engineering says IT does not know the first thing about pushing packets through the network, be they colored, striped or with polka dots, and blames the whole thing on Marketing. Strategy tiptoes out of the boardroom, heading for a conference on "fluffy packets" (the next big thing).
  • Some Linux guy in either Engineering or IT sets up a server that generates perfect blue packets with 3 lines of Perl code, but nobody notices him.
  • Engineering gets Purchasing to issue an RFQ for "not green and not red packet services" and gets Radio Planning and Transmission onboard to validate the service rollout requirements. Radio Planning immediatly starts looking at special blue antennae with a higher blue-to-gray dispersion rate, while Transmission wants to swap out all the optical gear and replace them with blue lasers to minimize signal loss.
  • "Not quite blue" specialist tech consultancies appear and start trying to schedule meetings with Engineering and Marketing to hawk their "deep knowledge of the telecommunications industry" and share their views "on how we can jointly move forward into the multicolored packet marketplace".
  • Marketing is dazzled by a vendor that is pushing an integrated Push-To-Blue solution, except it only works if the handset is used upside down during a full Moon in Hawaii.
  • Marketing takes Management and Purchasing to a live on-site demo, where Purchasing finds that the whole thing costs slightly less than the Space Shuttle. Management, pressured by Marketing, signs off on a trial nonetheless.

Year Four

  • Cisco adds a footnote to its latest IOS release (just below an evil bit bugfix) that says "implemented preliminary support for IETF rainbow packets".
  • Engineering asks Legal to have a look at the trial support contract, which specifies that an expert colormetrist has to be flown in from Nepal four times a year (all expenses paid) to compare the color output with measurements taken in Mount Everest during the Maltese bumblebee's mating season. Legal expresses concern that that clause may collide with insect import laws, and launches an intellectual property investigation to find out if anyone has patented the color blue.
  • Operations is asked to allocate floorspace for the trial system, and immediately asks for training on it. Engineering pushes the thirty extra people IT crammed into the available training slots, and spends two weeks with Operations and the vendor trying to get the whole thing to work in time for a Marketing demo.
  • During the demo, Management pushes a button and the phone screen turns pink (Marketing asked Design to come up with "something new"). Management decides to expand the demo to an in-house trial.
  • A handset vendor ships in a set of phone prototypes with a large, extremely bright blue button on the outside. Marketing covers the button with silver duct tape so that people don't get blinded by the blue light and hands them out to Terminal Testing and a few "normal" users.
  • During the trial, some guy in Operations gets called in one night to figure out why the service isn't working. He tries the vendor hotline, which is in the process of being re-outsourced from India to the US and has an incomprehensible answering message in Gujarati that translates to "Your message is important to us. We are moving offices and apologise for the inconvenience, please call back next week". He eventually opens the cabinet to find an Apache server driving a Korean clone of an HP printing head attached to the network card. He jiggles the printing head a little and it starts working again. He tells no-one.
  • Terminal Testing files enough bug reports to justify failing the "duct tape packet" handsets three times over. They also point out you can't actually place a call with them. The vendor says that will be fixed in the next release.
  • Rumors come in that the competition is preparing to launch a "green packet" service and corner the market, so Engineering and Marketing get together with Management and force vendors to "show their committment" to ensure the company gets there first. "Committment" turns out to be French for a rather delightful dessert on a seaside hotel.
  • Customer Services gets wind of all this and sends out a memo that expresses concern about customers calling in and asking for a "puce packet" service, and if we're developing one why weren't they involved in the first place? Things are settled in a joint meeting with Marketing in which it is shown that we are actually going to launch a blue packet service, that Customer Services would be brought on board six months from now (way in advance of launch) and that nobody wanted to distract them from their main goal, which is to keep customers happy.
  • Marketing takes advantage of that meeting to draft a series of slides listing the advantages of blue packets over any other color packets, complete with a SWOT analysis. It also files a new set of requirements for a "blue packet service portal" focused on "blue packet content" and containing a "wide, customer-driven personalized choice of exciting blue items", that mercifully gets lost on the photocopier.
  • The Linux guy adds a couple of more lines to his script and starts generating blue packets with yellow polka dots. Ten more lines and he gets it working with one of the trial phones. He eventually rewrites the phone firmware entirely and sends it back to the vendor with support for rainbow packets, a 1000-line patch fixing all the previous bugs and a game of Pong played with a blue pixel as the ball.

Year Five

  • Marketing starts negotiating handset volumes and pricing, and asks Advertising to start buying slots for the upcoming campaign. Billing eventually corners it and they enter a series of long, protracted meetings in which it is discussed if what is billed is the button press, the transmission of the blue packet from the CPSN, the transit of the blue packet via the CPGN or the actual "screen turning blueish" bit. Marketing shows their SWOT analysis, but nobody cares by now.
  • Engineering has to sit in on all of these and take the flak for not pushing 3GPP in precisely the right way IT feels the service should work, but it gets even by mentioning Strategy's upcoming "soft packet" concept as a good reason for implementing blue packet support the right way.
  • Engineering also points out it has its hands full rolling out special antennae and blue lasers all over the place, so that they can be tested months in advance of any live customers entering the scene, and that it will get their bit done months before IT goes into beta-testing.
  • IT is spited by this and commits to a new, aggressive platform deployment strategy that will replace the entire company's systems with the ones Engineering got Management to cut from the budget last year - all to ensure that the "navy packet" service is ready on time.
  • Legal and Purchasing finally agree on the vendor contract, which mandates per-server, per-user, per-connection, per-event and per-phase-of-the-moon licensing fees. The fine print reads that if the company exceeds its alloted licenses, the servers will stop working until another Nepalese guy comes over and twiddles a little knob marked "number of licenses" on the underside of the servers - for a fixed fee, of course.
  • The trial service is replaced by the production service, which happens to be twice as big and thirty times as expensive. The Operations guy sneaks in one night, opens the box and sees two Korean print heads inside. He decides to leave the company and start his own telco equipment outfit.
  • Advertising gets back to Marketing with a list of advertising slots and a bunch of ad concepts. Since it is impossible to show blue packets on a TV screen, they come up with "The Blue Packet Mascot", which looks like someone dressed up like a bar of blue soap. Casting for the role begins.
  • Engineering begins field trials and stress testing of the service, which mostly consist of getting everyone with a test handset to press the blue button at the same time. Special "stress testing consultants" pop up at the critical moment, and are chased away with blue lasers and rather spiky-looking antennae.
  • Management calls for a press conference on Launch Day, and gets photographed alongside "The Blue Packet Mascot" - an out-of-work actor who is unrecognizable inside a costume made entirely out of blue soap.
  • Enthusiastic reporting hails the "blue packet" as the next best thing since sliced bread in mobile services, and customers buy "blue screen" handsets in droves - to use the blue button as a flashlight.
  • Strategy comes back with a study showing that mobile internet access (i.e., a straight-up Internet connection) may be the next big thing. Since they forgot the revenue projection slide again, everyone ignores it.
  • The competition launches an "orange blinking packet" phone the week after, but the phones end up being banned because they look like emergency vehicle lights and confuse drivers.
  • The USPTO allows someone to patent "the use of the color blue in data networks".
  • The vendors manage to sell Marketing on to the idea that playing Pong with a blue ball is an original idea, and it asks Advertising to base the entire Xmas campaign on "Blue Pong".
  • The Linux guy eventually leaves the company to start his own Internet venture and ends up buying Google five years later.