Many of the ICT technologies that are now taken for granted came about through lucky accidents or unintended consequences as the equipment vendors attempted to 'cross fertilise' different disparate types of product line – a tradition that continues to this day.
The mother of invention may be necessity, but good old-fashioned trial and error is an equally viable approach to technological innovation. For as long as IT and telecommunications companies have needed to make a profit, they have tinkered with their product formats, often creating hybrid technologies that splice elements of one thing with those of another in a bid to come up with a magical, money-making formula. Sometimes it works; sometimes it doesn't – but the results often contribute to the over direction of ICT technological development.
One such innovation in the data networking industry happened largely by accident after the initially intended product didn't work properly. In 1994 enterprise switching and routing specialist Xylan called its multi-protocol product the PizzaSwitch. The technology continues to underpin part of the hybrid Ethernet/ATM/FDDI product portfolio still offered by Alcatel-Lucent, which bought-out Xylan for $2bn in 1999.
However, the PizzaSwitch was never really meant to be, having been assembled at the last minute largely by chance. Xylan was trying to build a product to serve a $16m contract win from the Japanese government, which funded a project to connect the country's schools after finding that half were using the Token Ring LAN protocol (802.5, developed by IBM) and the other half the rival Ethernet protocol (802.3), with the only high-speed protocol able to interconnect them being FDDI (fibre distributed data interface).
"In the first four months Xylan had no product because they could not get the ASICs to work, so they brought in ATM (asynchronous transfer mode) switching and put translation from Ethernet to ATM chipset in it, and also Token Ring to ATM, and FDDI to ATM," says David Palmer-Stevens, now systems integration manager at Panduit, but once a senior executive at networking vendors Xylan and Cabletron. "So Xylan launched the PizzaSwitch, which connected them all together by accident because it had to support everything they had."
The product's versatility was highly innovative at the time, but also something of an early 'disruptive' technology because the Ethernet and Token Ring markets were separate domains and disinclined toward any kind of convergence.
Pizza the action?
Another 'pizza'-based product, albeit one which bombed in comparative terms, came from Nine Tiles Networks, a small Cambridgeshire-based company that previously produced operating software for a range of personal computers, including the Sinclair ZX, before branching out into do-it-yourself dual-ring network CoRoNet (Contra-Rotating Network) in 1991.
"This [CoRoNet] was a token-based network technology that was incompatible with standard Token Ring," says Steve Broadhead, director at Broadband-Testing, an independent labs facility which has been conducting product analysis for more than 20 years. "It arrived in a giant pizza box, complete with all the hardware for six users and a toolkit to manually build the whole network, including crimping the cables. It was not a major success."
Squeezing as much technology into a product as possible in the hope of getting a customer to bite continues apace. The practice is perhaps more often driven by commercial rather than technological factors as companies look to combine their strengths in one market with their capabilities in another to maximise cross-selling opportunities.
A notable recent example comes from telecom, network and cloud infrastructure platform vendor Kontron, which in January introduced the SYMKLOUD, essentially a data centre application server aimed at mobile operators who may see fit to offer cloud-based machine-to-machine (M2M) communications services to their enterprise customers. SYMKLOUD mixes copper and fibre-based network switching capabilities into a 2U rack unit alongside load-balancing and server-based processing capabilities. Its ambitions are predicated on the expectation that operators will see it as a platform to host mobile and video transcoding, unified communications, and video surveillance, as a service applications.
The hardware is in no way application specific or agnostic, and nor does it use specialised components. Kontron appears to be looking for a platform to sell which can exploit its sales and marketing expertise and customer base in embedded computing, of which M2M is one component, with a networking platform that can tap into surging demand for cloud computing services and the infrastructure which providers need to host and sell them. There is an argument to say that where ICT software and services converge the supporting hardware will follow, and it is one that Kontron appears to believe in; whether its customers will agree remains to be seen.
At least Kontron has a bit of space to play with and a relatively abundant supply of electricity to power the components it packs into its equipment.
The same is not true in the mobile industry, where handset manufacturers face a never-ending struggle to shoehorn as much technology and features as possible into a small form factor, but where the resultant effect on battery life can be severe: all of which makes the small number of manufacturers that continue to support dual SIM cards – handsets that allow a subscriber to access two mobile phone accounts from a single device – perhaps that more surprising.
The same is true of push-to-talk functions, an early hybrid approach to mobile voice first launched in the mid-1990s that pushed 'one to one' and 'one to many' group calling sessions over cellular rather than private mobile radio networks, allowing workers to access walkie-talkie type functions from standard mobile handsets.
Advancing the industry
Though still supported by a small number of handsets, push to talk largely failed in the face of fierce opposition from mobile network operators concerned it would eat into their core cellular voice revenues, a fight since lost to the emergence of over-the-top mobile voice applications from Skype, Apple FaceTime and others, which effectively stole the push to talk market as well.
"Things like dual SIM devices actually made a lot of sense, but were far more popular in Asia where the networks span different countries," says Steve Haworth, chief executive at telecommunications service provider Teleware, and a mobile industry veteran. "Similarly, push to talk ' where you get a group of five people and if you press the button on the phone, it was like a one-way conversation ' generated a lot of interest, but failed massively. Part of the reason was that voice was sent over the data channel and the operators did not like that ' but look at Skype and others now."
The hybrid concept was embraced for a while by PC makers who were faced with a fragmenting market where laptops, netbooks and tablet devices were all competing for market prominence.
In 2010 Lenovo launched what it described as the industry's first hybrid PC, the IdeaPad U1 hybrid notebook, featuring a detachable screen. The design was engineered to provide two PCs in one device – each with its own processor and operating system ' that work either a 'clamshell' laptop or a multitouch 'slate' format tablet device.
The two CPUs can synchronise to work as one with the ability to share battery power, 3G, data and documents. In this way, the base laptop system can serve as a hub and docking station and the slate tablet as a mobile device. The two PCs have been engineered to work together and independently through Lenovo's Hybrid Switch technology that enables seamless toggling between the two processors. For instance, IdeaPad users can surf the Web in laptop mode and then continue from the same point without interruption if they detach to tablet mode.
Successful hybrid technology
Perhaps the greatest missed opportunity within the networking industry of all time came from network equipment vendor SynOptics Communications. The now defunct company began experimenting with fibre-based networks in the 1980s before coming up with one of the first implementations of a 10Mbit/s Ethernet LAN hub to use unshielded twisted pair cabling, an initially proprietary approach to which hybrid 10Base-T connectivity options were added at a later stage of development.
The branded LattisNet hubs were configured in a star topology as opposed to the linear bus topology that saw individual 10Base2 nodes connected directly to one long coaxial cable using a series of BNC barrel connectors, T connectors, and terminators. It was a viable competitor to IBM's Token Ring networking solution mentioned above, which at the time threatened to crowd out the nascent competitor Ethernet completely.
The LattisNet products were successful, later being rebranded and sold by bigger player Western Digital; but SynOptics was unable to capitalise on its patent, choosing to work alongside other vendors and standards association the IEEE to develop what, in 1990, became the dominant 10Base-T Ethernet standard open to all comers.
"SynOptics [has] sort of a dilemma," Tom Bredt, managing director of SynOptics' financial backer Menlo Ventures wrote in 1995. "If we retain the patent rights and defend our exclusivity, we basically have to give up the idea of becoming the industry standard. On the other hand, if we want to have our technology and architecture adopted as the industry standard, we basically have to throw our patent on the pile and offer a free license to anyone.
"So we decided to go for the standard and to give free license to the patent. By becoming the standard, the market acceptance of this technology would be dramatically increased. In fact, had we kept the proprietary standard, I believe someone else would have gone to the IEEE and a different approach would have been adopted ' and SynOptics would have been left in the dust."
Its early innovation drove SynOptics sales revenue from an estimated $1.8m in 1986 to over $40m in 1988, with further innovation around intelligent hubs and a multi-protocol network management system seeing those figures rise to $77m in 1989 and $176m in 1990.
Despite introducing hybrid LattisNet products that added 10Base-T connectivity modules, by 1991 the growing number and strength of larger rivals selling IEEE-compliant 10BaseT equipment (US rival Cabletron in particular) began to take its toll, leading to a decline in share values and ultimate sale to Bay Networks in 1994.
Ahead of the curve pitfalls
Alongside standardisation barriers, there were the innovative multi-function products which were, arguably, way ahead of their time, but also scuppered by questionable marketing decisions, an uninspired or frankly misleading product name, and/or a failure to explain their significance to potential customers.
Cabletron's SecureFast switching software architecture for ATM and Ethernet networks was arguably the precursor to the network virtualisation and software defined networking (SDN) platforms now being pushed heavily by almost every vendor in the data networking and storage industry. Back in the mid-1990s, however, it never got far from the starting blocks.
"If you want one flat, open network, which is completely transparent, how do you control the broadcast storm?" asks David Palmer-Stevens, a senior Cabletron manager at the time. "Cabletron came out with technology that does that ' layer two traffic control, workgroups, and on-demand broadcast – and called it SecureFast."
He adds: "We spent more time explaining what SecureFast meant than explaining the technology – and that was the beginning of the end for Cabletron. It was thought that virtualisation could not work across any router, and Cisco had to work with VMware to come out with the Access 1000 (which later morphed into Cisco's Nexus 1000 unified architecture platform), but if Cabletron approached VMware then, that discussion would never have happened."
OS security: finding the right fit
Elsewhere, US software firm Banyan designed VINES (Virtual Integrated NEtwork Service), a network operating system broadly based on Unix, to compete with market dominators Novell Netware and Microsoft LAN Manager. VINES sported two innovations: dual CPU support on the server – long before any applications could make use of the technology – and hardware-based security support using a dongle, which attached to a serial port on the back of the server itself.
"The problem was that each hardware dongle plugged into the back of the previous one," recalls Broadband-Testing's Steve Broadhead. "So, for example, if you had more than three additional services, you had about 10 inches of unreliable dongles sticking precariously out of the back of the server ' and if those failed, so did the NOS.
"What is even worse, they did not even learn their lesson from Novell who had the same problem on an early version of NetWare when using a dodgy dongle to secure the license for each server."
Possibly – probably – the quirkiest hybrid data communications device of them all to ship in the UK was Amstrad's E-mailer (or 'E-m@iler', as it was branded). A mixture of analogue telephone and fax machine twinned with a tiny Internet-enabled PC sporting a small 480 x 320 pixel LCD screen offering dial-up web access and email capabilities for people who (for whatever reason) did not want to buy a conventional PC to get online.
Launched in 2000, the £79.99 device was revised in 2002 and 2004, adding some limited video features along with the ability to play ZX Spectrum video games.
Although promoted heavily by Amstrad chairman Lord Alan Sugar (it even appeared as an office adornment in early series of the BBC's TV show 'The Apprentice', as fronted by Sugar), the E-mailer did not really find appeal with the mass market as a would-be object-of-desire (despite having a sort of retro-chic about its design). Nor did it catch on with niche consumer segments such as senior citizens who might want to exchange emails with family and friends, but were daunted by the cost and complexity of standard PCs.
In actual fact, the Amstrad E-mailer was arguably not a less expensive alternative to conventional personal communications. Internet, faxes and email were charged on a pay-as-you-use basis via dial-up to a premium number, costing at least 12p a time, with SMS messages and downloads charged per transmission ' a market model which proved pricey for customers at a time when voice and data services were being provided for next-to-nothing on the regular Web.
Amserve, the Amstrad subsidiary set-up to market and sell the E-mailer, posted a series of multi-million pound losses following the product's launch after disappointing (and reportedly heavily subsidised) sales, which impacted its parent company's own profits.
The Amserve E-Mailer service was transferred to BskyB in 2010 which closed it down a year later, though E-mailer devices could still be used as ordinary telephones, but only as long as the owner plugged it into a BT socket at specific dates to facilitate a software download and reconfiguration.