LOS ALAMITOS, Calif., 08 November, 2017 – In December 2016, the IEEE Computer Society (IEEE-CS) released its technology predictions for 2017, with a focus on industrial IoT, artificial intelligence, machine learning, cognitive computing, 5G, accelerators, and others on the front line of computing breakthroughs. Today the IEEE-CS graded its 2017 predictions while providing key insights on the direction of each trend.
“The past year has seen dramatic adoption of deep learning and rapid introduction of cryptocurrencies and blockchain technologies,” said Dejan Milojicic, IEEE-CS past president. “We have not seen such innovation dynamics in a long time, all setting the stage for an even more promising 2018.”
Following are the IEEE-CS predictions, grades, and analysis:
Industrial IoT: B+
Of all IoT efforts, industrial IoT gained the most credible advances over the past year. The reason for the B+ grade is that it has not yet reached our expectations for broad adoption.
Self-driving cars: B-
Self-driving cars continued to improve, but wide adoption is still hampered by legal, ethical, and, to the least extent, technological advances. This resulted in a lot of negative press. It did, however, help with assisted driving.
Artificial intelligence (AI), machine learning (ML), cognitive computing: A+
Deep learning in many technology areas contributed most to adoption, hence the A+ grade. It is starting to be utilized in many other technologies, including at the edge of the network and in data centers, and is paving the way to broader adoption of AI and ML.
5G continued advancing, but is still in initial deployments. Broad adoption is still a question given the advances of the existing standards.
Accelerators have been the foundation of many of the advances of deep learning. They have been deployed in the form of ASICs, FPGAs, and GPUs in data centers (Google, Microsoft, and Amazon) as well as at the edge of the network.
Disaggregated memory/fabric-attached nonvolatile memory (NVM): C+
The opportunities of NVM continue to promise innovative disruption up and down the software stack. However, the properties observed in early devices have not shown a compelling economic value, and have caused delays in delivering the core technology by all the major vendors. This resulted in a low score of our prediction. We still expect that NVMs will be on the upswing once hardware products become available.
Sensors everywhere and edge computing: A-
Sensors continue to gain wide adoption at home, in industry, in transportation, in smart tools, etc. Edge computing enabled by accelerators is gaining momentum across industries. There is, however, still more talk than actual deployment this year.
Blockchain (beyond Bitcoin): A It appears that blockchain is gaining momentum, both in industry (startups and mature) and academia (see IEEE Spectrum October issue.) We will see if this momentum continues next year and gains more adoption.
Hyper-converged systems: B
These had a solid year, and the reason for not giving our prediction an even higher score is that they still have not delivered the “software-defined everything” vision.
What was missed last year and how well the technology has fared:
Microservices are getting wide adoption in services computing as a basis for design and implementation. Being lightweight, they are very effective and complementary to containers.
Containers became a de facto environment for development. Lightweight, they enable much more scalable and quicker tools. Despite some security issues, they gained dramatic adoption in the last year.
Ethereum, other digital currency: A+
Digital currency is on the upswing in industry and being treated as equivalent currency. Investors in these currencies gain a substantial return on their investment. They are also becoming adopted for payment.
Overall Score: A-
Looking back, we correctly predicted most while only missing a few.