The Future of the Mac: Networking
The next two years promise much faster short-range and medium-range wired and wireless networking, and zippier mobile connections.
What’s the point of these increasing speeds, when today’s networking gear already feels reasonably fast? Video, primarily. The name of the game is streaming or downloading high-definition video—across a room or from one end of the house to the other—while still leaving enough room on the network for other activities such as file downloads, Facebook, and email.
Back when Thunderbolt technology was still under development, Intel (its main mover) said that optical cabling would enable multiple-gigabits-per-second connections to displays, peripherals, and networks, and would permit cords to run as far as 30 meters. Things didn’t quite pan out that way.
Apparently optical cabling was hard to produce and expensive, so Apple and Intel switched to copper wiring instead. That wiring still supported the Thunderbolt spec’s two channels of simultaneous bidirectional (“full duplex”) 10-gbps data. But it didn’t allow for those long cables; instead, connections could be no more than about 10 feet (3 meters). Using wire also necessitated the addition of chips to the cables, to handle signaling and to ensure backward compatibility with DisplayPort. The one upside: Thunderbolt cords can pass up to 10 watts per device, more than double USB 3.0’s capacity.
In the future, when compatibility with older connection standards becomes less important, Thunderbolt will likely return to optical cables. Those cables will allow for 30-meter runs. And the intelligence could move from the cables into computers and mobile devices, making Thunderbolt cables cheaper. On the flip side, the cables will likely be able to deliver just 4.5 watts or so per device. My guess is that an optical update to Thunderbolt could arrive in Apple hardware within two years—but rather than replacing current Thunderbolt ports, it would have to come in the form of a new port on Apple’s pro models.
In the meantime, USB 3.0 has found its way into all new Mac models (except the Mac Pro). At 5 gbps, USB 3.0 is not exactly slow, and compatible hardware is widely available. It meets many of the needs that Thunderbolt fills, except for standard support of external displays. I predict that the scarcity and high costs of Thunderbolt-compatible hardware will keep USB 3.0 as the preferred choice for people who don’t need the highest possible performance.
While gigabit ethernet is available in all Mac gear with ethernet ports (except the AirPort Express and Apple TV, which top out at 100 mbps), the 10-gbps flavor of ethernet seems unlikely to come to Apple hardware anytime soon. That’s largely due to the cost of adapters and switches. Thunderbolt may be the solution. Apple already offers a Thunderbolt-to-gigabit-ethernet adapter, and the company could conceivably sell a 10-gbps adapter as that market matures. But outside of server rooms and data centers, gigabit ethernet will probably remain the default choice.
Through the ether
The 802.11n flavor of Wi-Fi has become the default in all Apple devices that have wireless capabilities. Newer 802.11n devices (including the iPhone 5) boost wireless speeds by supporting both the 2.4GHz and 5GHz frequency bands. But wireless networking will soon get even more of a boost thanks to the advent of two new wireless technologies: 802.11ac (which is an update to 802.11n) and 802.11ad (for in-room superhigh-speed streaming).
The 802.11ac update, already shipping in some equipment from several vendors even though the standard is not yet finalized, can boost wireless networking speed to a raw rate of over 1 gbps, but only in particular cases. While 802.11n tops out at a raw data rate of 450 mbps in Apple equipment and similar networking gear, comparable 802.11ac base stations will have a minimum top rate of 867 mbps.
Since 802.11ac works only in the 5GHz band, 802.11n will remain the standard for communicating in the crowded 2.4GHz band. In addition, much of 802.11ac’s performance improvements will be realized only in particular circumstances or when you’re using advanced hardware; for that reason, the greatest boosts in speeds won’t be apparent in consumer networking, but rather in enterprises, on academic campuses, and perhaps at large-scale hotspots such as airports and convention centers.
Despite those limitations, Apple could add a preliminary version of 802.11ac to its base stations as soon as the next major refresh of the product line, and that typically happens in February or October. Adding 802.11ac to mobile devices might not enhance their speed much, but it would improve efficiency: A base station with that standard built in can simultaneously and separately communicate with multiple simpler 802.11ac devices (those that can’t send multiple data streams at once) instead of interacting round-robin among them.
Although 802.11ac will eventually become part of the certified Wi-Fi spec, the 802.11ad standard is something else altogether. It offers four channels, over each of which data can race at rates up to 7 gbps. But such speeds are available only over distances of no more than about 30 feet. That’s because 802.11ad uses the 60GHz band to transmit data, and signals at those frequencies can’t penetrate objects well. It’s certain to be restricted to applications in which the networked devices are in line of sight (or reflection) of each other.
For that reason, 802.11ad’s primary uses will be for rapid transfer of large files—such as sending a movie file from a computer to a set-top box—or for streaming uncompressed high-definition video. Normally, high-definition video is stored in compressed form on hard drives, DVDs, or Blu-ray; when you stream video over the Internet, it travels in compressed form, too. Once your device receives the compressed data, it decompresses the data for playback. Repeated compression and decompression can compromise video quality. If you can send video uncompressed—which 802.11ad enables—you can watch it at the highest possible quality.
The ideal convergence will be base stations and adapters that incorporate 802.11n for 2.4GHz and 5GHz, 802.11ac for 5GHz, and 802.11ad for 60GHz, switching as necessary to the best medium for the task or the reception quality. Some chipmakers have already announced plans to make sets of chips for just that purpose—but you shouldn’t expect to see 802.11ad in shipping hardware any earlier than 2014.
LTE (Long Term Evolution), a mobile broadband standard that cellular operators are rapidly deploying around the world, is all the rage right now. AT&T, Sprint Nextel, and Verizon Wireless have LTE installed in a reasonable percentage of the United States already, and by 2013 all three companies expect that most or all of the areas where they currently offer cellular data will also have LTE.
But the kind of LTE that carriers have put in place so far is more of a low-speed version of the spec. When LTE was first being devised as a thorough overhaul to the evolutionary approach of GSM-derived 3G and 4G networks a few years ago, the developers knew that an even faster version would be possible. That version, now called LTE Advanced, calls for speeds up to 3 gbps with fixed devices and hundreds of megabits per second for in-motion receivers traveling rapidly, as in cars and trains.
Like LTE, LTE Advanced can make use of frequency channels of many different widths, in contrast to the hard limits of 3G and 4G networking technology. LTE Advanced goes even further, and can allow aggregating channels that are spread out, as opposed to continuous frequencies, making it possible for a carrier to assemble bandwidth at lower cost.
LTE and LTE Advanced have essentially won the mobile-standards battle worldwide for cellular networks, too, so any advances we see will come from this technology.
Apple buyers shouldn’t hold their collective breath for LTE Advanced, however. Carriers have to acquire more bandwidth to make it truly useful, and in many cases are already paying billions to install the existing LTE technology flavor. Reports indicate a target of 2014 and beyond for any substantial upgrades and availability in handsets and other receivers.