Keeping an Eye on the Future

The IEEE, the Institute of Electrical and Electronics Engineers, has been issuing a document annually that lays out a roadmap to make sure that the computer chips that drive all of our technologies are ready for the future. The latest such document is the 2019 Heterogeneous Integration Roadmap (HIR). The purpose of the document is to encourage the needed research and planning so that the electronics industry creates interoperable chips that anticipate the coming computer needs while also functioning across multiple industries.

This is particularly relevant today because major technologies are heading in different directions. Fields like 5G, quantum computing, AI, IoT, gene splicing, and self-driving vehicles are all pursuing different technology solutions that could easily result in specialized one-function chips. That’s not necessarily bad, but the IEEE believes that all technologies will benefit if chip research and manufacturing processes are done in such a way as to accommodate a wide range of industries and solutions.

IEEE uses the label of ‘heterogeneous integration’ to describe the process of creating a long-term vision for the electronics industry. They identify this HIR effort as the key technology going forward that is needed to support the other technologies. They envision a process where standard and separately manufactured chip components can be integrated to produce the chips needed to serve the various fields of technology.

The IEEE has created 19 separate technical working groups looking at specific topics related to HIR. This list shows both the depth and breadth of the IEEE effort. Working groups in 2019 include:

Difficult Challenges

  • Single chip and multichip packaging (including substrates)
  • Integrated photonics (including plamonics)
  • Integrated power devices
  • MEMS (miniaturization)
  • RF and analog mixed signals

Cross Cutting Topics

  • Emerging research materials
  • Emerging research devices
  • Interconnect
  • Test
  • Supply chain

Integrated Processes

  • SiP
  • 3D + 2.5D
  • WLP (wafer level packaging)

Packaging for Specialized Applications

  • Mobile
  • IoT and wearable
  • Medical and health
  • Automotive
  • High performance computing
  • Aerospace and defense

Just a few years ago many of the specific technologies were not part of the HIR process. The pace of technological breakthroughs is so intense today that the whole process of introducing new chip technology could easily diverge. The IEEE believes that taking a holistic approach to the future of computing will eventually help all fields as the best industry practices and designs are applied to all new chips.

The effort behind the HIR process is substantial since various large corporations and research universities provide the talent needed to dig deeply into each area of research. I find it comforting that the IEEE is working behind the scenes to make sure that the chips needed to support new technologies can be manufactured efficiently and affordably. Without this effort the cost of electronics for broadband networks sand other technologies might skyrocket over time.


A Better WiFi?

Regardless of the kind of ISP service you buy, almost every home network today uses WiFi for the last leg of our broadband network. Many of the broadband complaints ISPs hear about are actually problems with WiFi and not with the underlying broadband network serving the home.

Luckily the engineers that support the WiFi standards don’t sit still and are always working to improve the performance of WiFi. The latest effort was kicked off a few weeks ago when the 802.11 Extremely High Throughput Study Group of the IEEE initiated an effort to look for ways to improve peak throughput for WiFi networks.

This group will be investigating two issues. First, they want to find ways to increase peak throughput on WiFi for big data applications like video streaming, augmented reality and virtual reality. The current WiFi standard doesn’t allow for a prioritization of service and the device in your home with the lowest bandwidth requirement can claim the same priority for grabbing the WiFi signal as the most data-intensive application. This is key feature baked into the WiFi standard that was intended to allow the WiFi network to communicate simultaneously with multiple users and devices.

The Study Group will also be looking latency. We are now seeing applications in the home like immersive gaming that require extremely low latency, which is difficult to achieve on a WiFi network. Immersive gaming requires fast turnaround of packets to and from the gamer. The sharing nature of WiFi means that a WiFi network will interrupt a stream to a gamer when it sees demand from another device. Such interruptions are quick, but multiple short interruptions means a big data stream stops and starts and packets get lost and have to be resent. Changing this will be a big challenge because the pauses taken to accommodate multiple applications is they key characteristic of the sharing nature of WiFi.

This Study Group effort is a perfect example of how standards change over time. They are trying to accommodate new requirements into an existing technology. We’ve never had applications in the home environment that require the combination of dedicated bandwidth and extremely low latency. In a business environment any application of this nature would typically be hard-wired into a network and not use WiFi. However, businesses now also want mobile performance for applications like augmented reality that must be supported wirelessly.

The Study Group is taking the first step, which is to define the problem to be solved. That means looking in detail at how WiFi networks operates when asked to handle big data applications in a busy environment. This deep look will let the engineers more specifically define the exact way that WiFi interferes with ideal network performance. If they have one, the Study Group might suggest specific solutions to fix the identified problems, but it’s possible they won’t have one.

The end result of the work from the Study group is a detailed description of the problem. In this case they will identify the specific aspects of the current WiFi specifications that are interfering with the desired performance. The Group will also specifically define the hoped-for results that would come with a change in the WiFi standard. This kind of document gives the whole industry a roadmap and set of specific goals to tackle, and interested labs at universities and manufacturers around the world will tackle the problem defined by the Study Group.

Most people in the industry probably view standards as a finished product, as a specific immutable description of how a technology works. However, almost the exact opposite is true and standards are instead a list of performance goals. As engineers and scientists find ways to satisfy the goals those goals the standards are amended to include the new solutions. This is done publicly so that all of the devices using the protocol are compatible.

I just had this same discussion a few days ago concerning the 5G standards. At this early stage of 5G development what’s been agreed upon is the overall goals for the new wireless protocol. As various breakthroughs are achieved to meet those goals the standards will be updated and amended. The first set of goals for 5G are a high-level wish list of hoped-for performance. Over the next decade the 5G standard will be modified numerous times as technical solutions are found to help to achieve those performance goals. It’s possible that some of the goals will never be met while others will be surpassed, but any given time the 5G ‘standard’ will be a huge set of documents that define the current agreed-upon ways that must be followed by anybody making 5G gear.

This Work Group has their work cut out for them, because the issues that are interfering with large dedicated data connections or that are introducing latency into WiFi are core components of the original WiFi specification. When the choice was made to allow WiFi to share bandwidth among all users it made it difficult, and maybe impossible to somehow treat some packets better then the rest. I’m glad to know that there are engineers who are always working ahead of the market looking to solve such problems.