5G wireless networks will handle much more data than current 4G wireless communications, which are powered by emote data centers that power . These networks will require computing power to be concentrated in servers mounted on the cellular base stations close to the smartphones and other Internet of Things devices they support. That will also spur changes to the computer networks connecting all that hardware.
That could open the door to silicon photonics transceivers that take advantage of light to transfer data faster and with lower power. On Monday, Intel announced that it had expanded its 100-gigabit-per-second (100G) silicon photonics products from data centers to the 5G network edge. The new transceivers can withstand the harsh conditions cellular base stations are subject to. They can transmit information to radio transmitters over six miles away.
“Our hyperscale cloud customers are currently using Intel’s [silicon photonics] to deliver high-performance data center infrastructure at scale, said Hong Hou, vice president of the silicon photonics division, in a statement. "By extending this technology outside the data center and into 5G infrastructure at the edge of the network, we can provide the same benefits to communications service providers while supporting 5G fronthaul bandwidth needs.”
Silicon photonics is currently used inside data centers to move information between servers. While a number of switch manufacturers are shipping discrete silicon photonics, the integration of optical components into the same package as electronics has been constrained by the high costs of packaging and production. These issues have been worsened because of the lack of software tools that ease the development of integrated silicon photonics.
The Santa Clara, California-based introduced a silicon photonics transceiver that supports 100G communications in 2016. The company has now ships a million units of the product per year into data centers. Intel’s 400G products are expected to enter volume production in the second half of 2019. Intel expects its connectivity business, which includes silicon photonics, to address an $11 billion market by 2022. Currently, the market represents about $4 billion.
The new silicon photonics product could eventually present another way for the troubled microprocessor manufacturer to accelerate computing from a system standpoint. The company, which bagged around $30 billion in data center revenue last year, has lost its process technology lead in recent years. Server chips based on Intel’s 10-nanometer node are still two years from being released, giving competitors a golden opportunity to court major customers.