Transforma logo
Subscribe Login

What Intel announcements say about the need for an interdisciplinary approach to looking at new technology

MAR 24, 2020 | Matt Hatton
region: ALL sector: ALL Internet of ThingsHyperconnectivityArtificial IntelligenceEdge Computing

In the absence of Mobile World Congress there were nevertheless ample chances to catch up with some of the key vendors to discuss technology. I took some time to chat with Intel. It made a series of eye-catching announcements as well as having a lot to say about some other interesting topics that float our boat.

Ostensibly our conversation focused on a set of announcements:

  • Expansion of the second generation Xeon Scalable Processor, a AI-embedded CPU/accelerator, delivering 36% more performance and 42% more performance per $.
  • First Intel Atom system-on-chip (SOC) for mobile basestations in the form of the P5900, which will bring the Intel architecture to the BTS and ship with various vendors. Expectations are high.
  • The “Diamond Mesa” eASIC - A structured ASIC to sit somewhere between FPGA general processing and ASICs. Not as flexible or quick to market as standard FPGA but more powerful. Intel is hoping there is a gap in the market there.
  • Ethernet 700 Series network adaptor aimed at solving synchronisation issues.

Discussion also turned to deeper discussions around AI and edge computing and Intel’s portfolio here including OpenVino, OpenNESS and Visual Cloud.

All of this points to Intel pushing at some open doors. Our thoughts on AI are well recorded. The impact is going to be massive. And most of it will be built on specialist hardware in the form of AI accelerators. Edge, and particularly the management of the relationship between edge and cloud, will also be key trend in the next few years. The marriage of 5G and MEC is clear, and for me the big value of 5G in the immediate future is around MEC. So Intel is pushing a lot of Transforma Insights’ hot buttons.

The interesting thing for me is the extent to which, to put these announcements into context, one needs to understand a very wide array of new technology areas. To understand IoT deployment parameters requires an in-depth knowledge of the changing nature of telecoms networks, and the evolution path of both of those is intimately tied up with artificial intelligence. To give an example, I wrote recently about 5G network architecture and the implications for where processing might be done in an IoT application. Read about it here . Without a relatively deep understanding of the capabilities of 5G it’s impossible to understand the likely impact on IoT. This is the background to the multi-disciplinary approach we’re taking at Transforma Insights.

Over the last few years, while our analysts have largely been nominally focused on ‘IoT’ the reality is that a large amount of the interesting conversations on that topic have revolved around the relationship with adjacent technologies such as AI, data exchange, edge computing, product lifecycle management and blockchain. Some technology areas are inextricably linked with IoT, while for others the relationship is more arms-length. But to do justice to the study of any of them requires a working understanding of the others. Furthermore, enterprises aren’t narrowly focused on AI, IoT or any other particular technology area. They want to understand the broad sweep of the disruptive technology landscape, the relative impacts and opportunities of each of them, and the use cases that they jointly and separately enable. Only a holistic approach to technology horizon scanning can achieve that. This is why at Transforma Insights we have chosen to cover the broad spread of disruptive technologies within our DX12 set of technology families (as illustrated in the graphic below) allowing us a complete perspective on the Use Cases which are the real consideration in terms of impact on an organisation’s operations.




Next Post MAR 31, 2020| Matt Hatton
AI gets a mixed scorecard in helping tackle Covid-19
All Blog Posts