Just like any commercial technology, the ultimate fate of a telco technology depends on market forces. In the telco industry, the invisible hands of the market have created two opposite pulls. The biggest one -the human needs highspeed and instantaneous data rate; the opposite one -the machines, headed toward low-speed and delay-tolerant IoT use-cases. Such a split has a profound implication on how a telco network should be planned and operated.
IoT is usually characterized by a combination of a handfull of features- the low-cost device with long battery life and the network supporting extreme coverage. The oldest among all telco technologies, the 2G has matured over three decades resulting in inexpensive 2G modems. The low price of 2G modems has enticed many enterprises in the early adoption of machine type communication, even though most of the operators didn’t implement any specific IoT features in the network. Another appeal of the 2G is its universality, almost all operators have a 2G network, and many of them operate in the same 2G spectrum. As a result, IoT roaming is possible without multiband, expensive modems. 3GPP realized this opportunity and specified EC-GSM for the 2G IoT use-case. So, it’s now apparent, shutting down the 2G network will not be possible anytime soon. It’s the IoT market force that will keep the 2G alive.
Meanwhile, the future of 3G looks bleak. The LTE superseded the HSPA in user experience. In the 5G era, operating a 3G network for mobile broadband will be a real dilemma for many telcos. Another surprising fact is that a 3G network cannot serve IoT use-case. Unlike the 2G, IoT will not be the savior of the 3G.
The following explains why it’s not possible to have IoT in a 3G technology, even though the earlier generation has robust IoT capability.
Possibly the most significant difference between a WiFi and a telecom system is about-how a user accesses the network. WiFi is always contention-based access, i.e. when one device transmits other must kept silent. If two device transmits at the same time, a collision occurs corrupting both signals.
On the other hand, except for initial access (aka, during the random access procedure), all telco network access (AS – Access Stratum) is contention-free, i.e. many devices can access the network at the same time without any chance of collision. Such simultaneous access is possible in one of the four ways,
1 -By providing each user with a dedicated portion of the carrier spectrum. This is also known as FDMA. The critical point is that each carrier portions (aka, subcarrier) must be orthogonal (i.e. “independent” in layman’s term). Technically, “orthogonal” means the “amplitude” of a subcarrier is zero at the center of the neighboring subcarriers. If orthogonality is not possible, there are two options -the use of a guard band. i.e. a gap between two subcarriers that no one can use, resulting in a waste of spectrum; or the use of TDMA (next).
2 -In TDMA, only a few selected devices are allowed to transmit at any instance in time; while others must be kept silent. TDMA by no means similar to WiFi access. In TDMA, the base station (BS) is in complete control over all uplink scheduling. Hence, a device must wait for an uplink grant before any transmission. While in WiFi, the access is always random. An access point (AP) in WiFi does not control uplink scheduling of other devices; instead, scheduling is a distributed function and controlled by the devices themselves using the CSMA-CA algorithm.
3 -The spatial multiplexing, a BS can transmit to two or more devices that are in different locations in “space” using the same spectrum and time. Since the devices are distant to each other, the chance for signal collision is low. Multiplexing the “space” is a crucial feature of beam-centric technologies such as in 5G NR and MIMO.
4 -Lastly, by changing the waveform. The shape of an electromagnetic wave can be transformed into two ways-by changing the “amplitude” or the “phase”. An increase in amplitude results in an increase in interference within the network; as a result, transforming the phase is more practical. This is achieved by using CDMA. In a CDMA system, all users can access the network at the same time using the same carrier spectrum. The network can distinguish signals from each user from their unique waveform. A 3G BS (the RNC) applies a mathematical sequence of a specific length (aka, spreading code) to users’ data, generating the so-called chip-stream. The chip-stream then scrambled with BS specific scrambling-code before modulating to the carrier signal. These operations result in the desired waveform transformation. As the number of users increases, so does the length of the spreading-code. It’s essential that the spreading-codes remain orthogonal to each other. To ensure orthogonality, the spreading-codes are derived from the so-called OVSF code tree (OVSF – Orthogonal Variable Spread factor).
In the 2G, the subcarriers are not orthogonal; so it cannot rely on FDMA only. Instead, the 2G uses FTDMA for contention-free access, which uses both FDMA and TDMA.
In LTE all subcarriers are orthogonal. Hence its FDMA is called Orthogonal FDMA (OFDMA for downlink). In the uplink, the LTE uses SC-FDMA (aka, DFT-S-OFDMA). The 5G NR uses OFDMA in both uplink and downlink.
The 3G is a wideband CDMA system; it has no concept of the subcarrier, as a result, all devices need to operate at a full 5MHz wideband spectrum. As explained above, in a CDMA network, many devices can transmit at any instance of time using total system bandwidth. As a result, the overall interference/noise level in a 3G network is high compared to a 2G or a 4G network. No wonder a 3G BS (the RNC) need to instruct each device more than a thousand time a second on -how to control the transmission power.
Back to the IoT, one of the outcomes of Shannon’s channel capacity theorem is that, in extreme coverage situations, data-rate depends on SNR rather than the bandwidth. To increase the SNR, the overall interference in the network must be very low; and for spectral efficiency and for simpler modem design bandwidth allocation must be reduced as well. As we can see, none of these are possible in 3G technology.
As the 3G data becoming more irrelevant due to market drive, many operators may find the 3G necessary for voice-only service; this may change due to 5G SA. The fastest 5G SA deployment needs VoLTE for EPS-Fallback, which may stimulate full-scale VoLTE deployment worldwide. A right kind of global drive for 5G SA might be the final blow to the 3G.