Innovation by Accident

Not every product we used today was planned out perfectly from the start. It is often the case that products are discovered accidently, or the way they are used commercially have nothing to do with the original intent of the products creator. Famous examples of this are things like the slinky toy, x-rays, and the microwave. Each of these inventions were created during research and engineering efforts to solve other problems. The slinky came into existence when an engineer dropped a large industrial spring, x-ray imaging and the microwave were both invented while researching radio waves. There are other instances where a byproduct or quick fix action ends up becoming a prominent feature of a system. Two technologies that fit this description are Network Address Translation (NAT) and Short Message Service (SMS) text messaging.

   

Network Address Translation

To understand the accidental impact of NAT, we must first briefly review the history of the Internet. The Internet that we use today was first developed as a communications network for the United States military. The Advanced Research Projects Agency (ARPA) built ARPAnet in 1969 as a way to connect military mainframe computers together. The original addressing scheme for ARPAnet was 8 bit and allowed for 256 different host addresses. The original ARPAnet started with 4 hosts and quickly grew to 213 hosts by 1981 (Bort, 1995). Realizing the limitations of the ARPAnet addressing scheme, Robert Khan and Vinton Cerf started working on a new addressing scheme. The fourth iteration of their work produced IPv4 addressing. IP stands for Internet Protocol and v4 is the fourth version they created. IPv4 is a series of 4 8 bit addresses that can support 4,294,967,296 unique addresses (Bort, 1995). This version was a tremendous upgrade from the original 256 possible supported hosts, but due to the rapid expansion of the Internet in the 1990’s, the mismanagement of address allocation, and other technical issues with routing and traffic management, the Internet was facing a real problem or running out of space that needed to be addressed. IPv4 address space was very large and there needed to be a way to route traffic to the appropriate networks around the Internet. The original solution was to build classes into the address space. IPv4 classes are simply a way to identify the size of a network based on the first 8-bit value in the address. The classes were broken up into 4 classes:

·         Class A –

o   First bit value of 0 – 127

o   126 networks with 16,777,214 hosts each

·         Class B –

o   First bit value of 128 – 191

o   16,384 networks with 65,534 hosts each

·         Class C –

o   First bit value of 192 – 223

o   2,097,152 networks with 254 hosts each

·         Class D & E –

o   First bit value of 224 – 255

o   Used for multicasting and R&D

This wasn’t a perfect solution because the amount of addresses per class were not scalable. For example, there are only 126 class A networks that could be given out, and each one had almost 17 million usable hosts. There are over 2 million class C networks that can be given out, but each one only has 254 available hosts. During the 1990’s, many large companies were given class A networks and didn’t use anywhere near the amount of hosts they had available, so those addresses were essentially lost. Meanwhile, smaller companies grew and required more and more class C networks as their demand for hosts increased. The class C networks weren’t given out in sequence, so companies had network addresses that weren’t mathematically close to each other, which increased the difficulty of routing Internet traffic. The problem was quickly becoming unmanageable.  

IPv6 was developed to expand the total address space for the Internet. It could support 340,282,366,920,938,463,463,374,607,431,768,211,456 total addresses (Loshin, 2001)! The problem was that IPv6 wasn’t drafted until 1998 and it was taking much too long to become a standard. A temporary solution needed to be found. NAT was developed to temporarily solve the issue with lack of address space but ended up solving many other issues that it essentially delayed the IPv6 rollout of IPv6 for almost 14 years! NAT is a protocol that runs on a router that borders an internal network and the Internet (Trowbridge, 1997). What NAT does is simply translate IP addresses on the internal network with IP addresses being used on the Internet. The feature that makes NAT so useful is that this translation can be one to many. This means that an organization can host multiple systems internally while only using one address to access the Internet. NAT adds information into the header information of network traffic that is used to assign that traffic an internal and external IP address to use. This way one external IP address can be used to provide Internet connectivity to multiple hosts. NAT inadvertently solved many of the problems with IPv4. Since IPv4 addresses could be reused internally, corporations only needed a few valid IP addresses to provide Internet connectivity to all their hosts. NAT averted the risk of running out of addresses so successfully that IPv6 could be delayed for years with almost no repercussions. This allowed IPv6 development to continue and provided a very robust addressing solution that should allow for sustainable address space for years to come. There were several other factors that dealt with IP addressing that contributed and augmented NAT, such as Classless inter-domain routing, that can be explored to provide a more detailed picture as to how NAT helped change the way the internet worked.

Text Messages

Text messaging has quickly become the standard way of communicating with mobile phones over the last 15 years, and it wasn’t a feature that was planned to be used for commercial use at all. Telephone lines have historically been analog systems, which means that they use waveforms to transmit voice and data instead of digital data such as bits. Early telephone systems handled all aspects of the call using signals that could be sent through the same wires that were used to send the voice waveforms. For example, phones ring by having a telephone switch send a higher than normal amount of electricity to the phone, which used to activate an electro-magnetic bell in the phone and made the phone ring. As telephones and telephony systems became more complex, the signals passed on the wire did as well. The signaling data eventually had to move off the voice transmission lines, and onto separate lines for management, which is called out-of-band signaling. Large phone switches would communicate things like timing, line availability, and other management information over out-of-band signaling. In 1984 Friedhelm Hillerbrand and Bernard Ghillebaert realized that this signaling traffic wasn’t always being used (Kuerbis, van Stolk-Cooke, & Muench, 2017). They developed a way to send ASCI characters along the signaling lines when they were not being used, which let them send text messages from phone switches to end users. The signaling formats that would send the messages could only support messages of 128 characters at a time, and both Hillerbrand and Ghillebaert believed that end users would only be able to acknowledge the message. Global System for Mobile communications (GSM) met in 1985 and started the process of developing the concepts behind Short Message Service (SMS), which is the standard used to send text messages. Since all phone traffic requires signaling data, providers could give text message access to their customers while incurring almost no cost developing or implementing the service. They were simply using a resource they already had in a different way.

SMS allowed for broadcasts to phone, like Hillerbrand and Ghillebaert first envisioned, but also point to point messaging between phones. SMS messaging was first commercialized by Nokia in 1994 and gained popularity with the advent of smart phones like the iPhone. In 1999 text messaging between networks became possible and SMS messaging dramatically increased. The average mobile phone user sent about 35 text messages a month in 2000, by 2010 200,000 text messages were being sent every minute, and over 6.1 trillion texts were sent that year(Steeh, Buskirk, & Callegaro, 2007)! Text messaging packages with cellphones started as an expensive perk and are now a necessity for any phone plan. Text messaging remains one of the largest examples of companies charging customers premium prices for a service that cost them almost nothing to implement. It has also become the de facto way to communicate today and it was never intended for that use!

References:

Bort, J. (1995). The address mess. InfoWorld, 17(46), 75.

Kuerbis, A., van Stolk-Cooke, K., & Muench, F. (2017). An exploratory study of mobile messaging preferences by age: Middle-aged and older adults compared to younger adults. Journal of Rehabilitation and Assistive Technologies Engineering, 4, 2055668317733257. doi:10.1177/2055668317733257

Loshin, P. (2001). Network address translation. Computerworld, 35(8), 60.

Steeh, C., Buskirk, T. D., & Callegaro, M. (2007). Using Text Messages in U.S. Mobile Phone Surveys. Field Methods, 19(1), 59-75. doi:10.1177/1525822x06292852

Trowbridge, D. (1997). A natty solution to a knotty problem. Computer Technology Review, 17(2), 1-1,6+.