Pages

Tuesday, October 10, 2017

Compuware Delivers Topaz on AWS to Mainstream the Mainframe

By Rich Ptak




Figure 1 – Topaz on AWS      Image Courtesy of Compuware, Inc.
It’s time for Compuware’s quarterly mainframe product announcements. This time Compuware kicks off its 12th quarter (3 years) of new and enhanced product releases by partnering with Amazon. The duo upends mainframe DevOps and mainframe IT by combining efforts to deliver web access to the Topaz DevOps software suite on AWS. See Figure 1.

In an industry first, Compuware provides cloud access to modern mainframe development via Topaz. Developers can enjoy the same user experience on the cloud as if Topaz was locally installed while fully leveraging all the security, performance, flexibility, reliability, scalability and accessibility features of the AWS platform.

Topaz on AWS leverages Amazon AppStream[1] 2.0 technology, a fully managed, secure application streaming service that allows applications to be streamed from  AWS to devices running a web browser. Now, all of Topaz’s rich capabilities and more are accessible anywhere through the most popular web interfaces including IE/Edge, Chrome and Firefox. The power of Topaz is accessible regardless of the device used, be it a Windows, Mac or Chromebook desktop system.

Compuware’s patent-pending technology provides an intuitive, streamlined configuration menu that leverages AWS best practices, and makes it easy for systems administrators to quickly and easily configure their Topaz on AWS infrastructure, customized to their specific needs, in a few simple steps.

Enterprises can scale the number of development environments up or down depending on their needs. Developers have fast access to new features and functionality that Compuware makes available every 90 days without administrators having to distribute, load, recompile, modify and test multiple individual systems or installations. Efforts to modernize mainframe operations and capabilities happen faster with fewer delays and without requiring the involvement of critical IT staff.

Compuware and Amazon have created a highly performant, secure and fluid developer experience. Once developers launch Topaz on AWS, they can access datasets and data files, analyze applications, make code changes and manage other mainframe tasks using the Topaz suite of tools as if the user environment was locally installed.

Some architectural benefits of AWS

An important feature is that this implementation fully leverages all of the unique enterprise product strengths of the AWS cloud architecture. These include:

  • ·         Security – individual secure deployment applied and management on a per account basis – with built-in automated security management services to review policies and monitor compliance with security best practices.
  • ·         Cost optimization – automated optimization assures least-cost-to-user and most cost-effective resource management. Periodic reviews and auto-scaling combine to optimize the operating environment as workload volumes and capacity requirements fluctuate.
  • ·         Reliability – AWS management services work to ensure systems are architected to meet operational thresholds to avoid when possible, and quickly recover from inevitable failures to meet business and customer service demands.
  • ·         Operational excellence – Amazon maintains cloud centers located around the world to assure service response and support.
  • ·         Performance efficiency – optimizes system services for maximum performance using available resources, enabling optimal utilization of IT staff and computing resources through automation, cloud-based services and management.


The pricing is right

All of this comes with no new charges from Compuware. The standard Compuware Topaz licensing charge covers the use of Topaz on-premise, in the cloud or in a mixed environment. If you already have a Topaz license, all you need to do is add an Amazon account with AWS cloud services designed to meet the requirements of your specific enterprise operating environment and workload. As mentioned earlier, Amazon AppStream 2.0 services include an automated function to help users find the right optimized pricing model and configuration for their workload and resource needs. These include highly flexible on-demand pricing, spot-discrete fixed price and reserved instances for dedicated predictable workloads, and combinations. We recommend you review the details with an AWS advisor or check here[2] for more information.

What else is new?

Compuware’s quarterly announcements are never about just one thing. This time is no exception. In addition to the major announcement, Compuware has additional improvements and new product enhancements and capabilities to deliver.  

First up is a collaboration with CloudBees Jenkins Enterprise. Leveraging Compuware ISPW and/or Compuware Topaz for Total Test in conjunction with CloudBees Jenkins Enterprise, large enterprises can streamline DevOps on their mainframes and orchestrate DevOps across all platforms. Compuware will co-host a webcast with CloudBees on October 25, which will identify opportunities and help educate users on the latest in mainframe DevOps processes.


   
Figure 2 Webhook Notifications  (Courtesy Compuware, Inc.)
Next is the addition that extends ISPW so that it can stream information and notifications to web apps through Webhook notifications. Webhooks were designed to allow third-parties, such as developers and apps, to make changes to web APIs using callbacks. See Figure 2. This is how ISPW can communicate with Jenkins and other CI services to trigger actions. In effect, it allows ISPW to integrate with other deployment tools and drive continuous integration processes. Activities can be communicated by DevOps teams as they happen in real time to such tools as Slack and Hipchat.

A bit of risk?

Choosing public cloud service delivery of the Topaz suite may appear to be risky or even premature to some potential users. Considerations that come to mind are those surrounding issues pertaining to security, reliability, privacy, infrastructure control and, unfortunately, more and more government imposed legal and legislative constraints and mandates. Most of these issues have been and will continue to be hashed over and argued about in the press. They remain and should remain issues of, at a minimum, keen awareness. Our conversation with Compuware has convinced us that they are working in lockstep with Amazon to reduce the risk and vulnerabilities as much as possible.

Potential customers should identify potential issues and resolve what needs to be done before making the move to the cloud. Others may find cloud-based but maintained and operated in-house on-premise to be the right solution.

The first step to be taken is to perform due diligence to identify and assess potential risks and vulnerabilities. Then, these can be balanced against the significant potential benefits in the form of client/customer satisfaction, staff satisfaction and cost savings that can result from improved operations, increased efficiencies and simplified infrastructure management. Examine what Compuware and Amazon have done to mitigate the risks. We believe that many will find the decision to move this development activity to the cloud makes sense.  

The Final Word

Compuware continues to deliver solutions aimed at “Mainstreaming the Mainframe.” Their strategy depends upon their ability to identify and overcome structural and operational issues that make mainframe utilization and COBOL code maintenance a complex, slow and intimidating task, especially for those new to the mainframe.

Compuware has delivered significant, game-changing products each quarter for the last 3-years. They have not only improved, simplified and sped up mainframe operations and management, but they have also introduced capabilities that were never thought possible or are radically changing mainframe operations. They appear to us to be on track to continue that success. Congratulations to them. Good luck as they move forward. We recommend examining their latest offering.


Monday, October 9, 2017

Launching a Secure Environment: Applying IBM’s LinuxONE Encryption

By Bill Moran and Rich Ptak

Courtesy of IBM


The other day we attended an excellent presentation by Dr. Rheinhardt Buengden of IBM Germany on applying the encryption in LinuxONE[1]. He provided extensive technical detail on installing and implementing a secure IBM LinuxONE Emperor II system (or one of the other IBM Linux mainframe system). It was a highly informative session.

First, nothing that we learned contradicts our earlier blog[2] on IBM’s announcement. We continue to believe that LinuxONE combined with its associated hardware represents the best commercial alternative for security on the Linux market. But, we did get some greater insight into implementing a high-security system.

 We now have a much better appreciation of the level of effort necessary to achieve a secure operating environment. As one might expect, much of the work revolves around having to choose among the many options in Linux. But, it also requires effort fit the new system into the way business is currently organized and done. To accomplish this requires significant skills in Linux and security methods as well as a detailed knowledge of the company’s current processes.

 We provide some specifics here. There are certain to be others. First, consider the interactions between the security key management and the existing disaster recovery mechanism. Some types of keys are system specific and will not work on another system. Careful planning is necessary to identify and handle inconsistencies and conflicts[3]. The LinuxONE system can automatically recover from an abnormal situation but only if the preparation work has been done.  Similarly, backup and archive policies will need a review for similar inconsistencies. The whole issue of key management will need careful study and decisions made in choosing among the various types of keys that can be implemented. Several types of keys are available; each type has its own different properties, advantages, etc.

There are choices to be made over how to handle the encryption applied to files, file systems and disks. Understanding the relative advantages and choosing the best one requires knowledge of the Linux facilities and their interactions with the security facilities. Failure here could result in an intruder being able to access the most sensitive information in the clear; fatally compromising all system security.

The last topic concerns the Linux kernel. Typically, the Linux kernel included security APIs that invoke certain software functions. LinuxONE hardware will speed up these functions. For this to work, the Linux kernel must be updated with code that supports the LinuxONE hardware. IBM has submitted a fix for inclusion in a future Linux kernel release.

This points to a bigger, more significant problem. LinuxONE relies on some Open source modules such as Open SSL, all such dependencies need to be monitored and updated or modified as necessary if security is to be maintained. We mention this point because the Equifax security breach has been tied to a lack of maintenance to open source module. The lesson is that maintenance for all modules in the system must be carefully monitored and applied. Open source code updates cannot and should not be ignored.

In sum, we think that anyone planning an installation of a LinuxONE system should understand the magnitude of the task they are undertaking and plan accordingly.

For a security project of this scope, seriously consider establishing a security subcommittee of the Board of Directors. This group needs to learn enough to ask the hard questions and supervise security audits of the organization’s activities.

A review of the presentation would benefit any group interested in security. And, be most helpful for groups considering purchase of the new LinuxONE system.  However, nothing will substitute for a knowledgeable and active staff handling the installation and operation of a LinuxONE system. Senior management support is critical. We hope our notes here make that clear.



[1] Here is the URL for the presentation: http://www.vm.ibm.com/education/lvc/LVC0927.mp4
[3] Details on this topic are beyond our current scope. See Dr. Buengden’s discussion on the topic 

Monday, October 2, 2017

IBM LinuxONE Emperor II ™, IBM’s Newest Mainframe Linux solution

By Bill Moran and Rich Ptak

IBM LinuxONE Emperor II

Introduction

On September 12th, IBM announced the IBM LinuxONE Emperor II™, a new, dedicated Linux mainframe with significant upgrades from its z13-based predecessor, IBM LinuxONE Emperor. IBM positions Emperor II as “the world’s premier Linux system for highly secured data serving, engineered for performance and scale.” IBM chose the LinuxONE Emperor II “to anchor IBM’s Blockchain Platform cloud service.” We discuss features and provide some thoughts on evaluating the system for your own environment.


Performance Features

Emperor II is a z14-based Linux-only mainframe system designed as a highly reliable and scalable platform for secure data-driven workloads. Key performance improvements include:
·         A 2-3 x performance boost over the z13-based Emperor.
·         IBM described 2.6 x performance advantage over comparable x86 systems for Java work, a result of IBM moving some CPU intensive Java operations into hardware.
·         Powerful I/O processing capability with up to 640 cores devoted to I/O operations, a benefit for I/O limited applications.
·         Emperor II can operate at near 100% utilization with very low performance degradation. Typical competing systems can achieve 50% or 60% utilization before experiencing significant performance degradation.
IBM’s LinuxONE Emperor II is an impressive, powerful, high performance system. Do keep in mind that all performance numbers are application/environment dependent. Therefore, if performance is critical, do your own testing. Vendor numbers can only provide broad guidelines to potential performance improvement.


Security Features

IBM LinuxONE Emperor I enjoyed significant market acceptance for a variety of workloads. Recognizing the escalating interests in security and high-volume data computing, IBM initiated a large engineering effort to enhance and extend already legendary mainframe system security. The z14-based Emperor II takes security to a completely new level.
IBM states that the system represents the most advanced level of security commercially available today. We believe there exists some justification for the claim. Here’s why.
·         A major block to large-scale encryption has been the extraordinary time and effort needed for encryption/decryption. IBM dramatically[1] decreased both by using an on-chip cryptographic processor (CPACF). This allows users to implement pervasive, end-to-end encryption of all data throughout (and beyond) the system. If a hacker breaks in anywhere in the chain, they only get access to encrypted data, useless without the ability to decrypt. 
·         Hardware protected decryption keys. A hardware-assist feature assures keys are never available in memory in the clear. There is no way for a user, hacker or even an administrator to unlock or make the keys visible and useable.
·         All data can be automatically encrypted and remain so, at-rest, in-motion and during processing – end-to-end – from system to user.
·         Encryption security is implemented with no application changes. Security solutions that require any changes (applications) or actions by developer/user/programmer have been a stumbling block for encryption (and other) security approaches.
·         Finally, IBM has a new architecture called Secure Service Containers. These containers protect the firmware and the boot process as well as, the data and the software from any unauthorized change. A traditional weakness has been the potential for system admins to exploit their elevated system credentials or for those credentials to be exposed to internal or external threats and then used to gain access to locally running application code and data. With Secure Service Containers, the only access is via the web or an API granted to those specifically with access to this environment. This closes a hole long used by hackers gain access to critical and private data.


Other key features

Emperor II delivers enhanced vertical scalability (scale up) possibilities, i.e. it allows a collection of tightly coupled multiprocessors to communicate at very high speed using shared memory. This architecture provides a distinct advantage for applications doing sequential updates to a relational database over scale-out systems, such as most x86 systems.
A typical example would be a banking application handling customer accounts. To maintain a correct account balance, all debits and deposits must be processed sequentially. That is, in the order they were performed, e.g. earliest date and time first. An account can be “locked” to ensure accuracy, having shared memory minimizes the latency and associated delay that results from such lock management.  Attempting this via a scale-out collection of independent systems can result in a very complicated software environment and may also result in performance problems whereas, IBM’s Emperor II would have neither problem.


IBM Strategy

Enterprise concerns about data security have changed, now having dramatically increased in priority. While previously it was on everyone’s checklist, when the final purchase decision was made price and performance dominated. Now, security is a deciding factor, and IBM is positioning the Emperor II to win.

This signals a broader change in IBM’s messaging strategy. No longer is the focus on “speeds and feed” with its reliance on numbers, processing speed, price/performance, TCO, etc. to motivate a change of platform. IBM intends to drive the decision using a business case focused on platform design (architecture) targeting the solution of major business and operational problems, as IBM LinuxONE Emperor II does.

Of course, much depends upon the platforms being compared. In many cases, inherent mainframe security will be decisive. IBM’s Emperor II with LinuxONE security and its vertical scalability far exceeds anything a standard X86 platform[2] has.

While we applaud this change in strategy, it can complicate the selling task. Since IBM’s target is x86 systems, sales reps may find themselves competing with Window systems as opposed to a Linux x86 systems. A security discussion comparing LinuxONE to other systems will require a more knowledgeable sales force. Features and functions such as security, Blockchain technology, etc. will have to be explicitly linked to specific business requirements, problem resolution, etc.

One final word on security. The heavy emphasis on security also represents a risk as bad guys are likely to focus on exploiting weaknesses in applications or lax security procedures as the easiest point of vulnerability. Consumers, businesses and journalists are notoriously quick to indiscriminately point the blame to technology for failure. A successful penetration via, for example, an app accessing an Oracle database when the platform functioned perfectly – can quickly be blamed on the platform and the app overlooked. IBM effectively and economically addresses a real problem area. But, there exists much more to be done by the entire community.


Summary

IBM has done an excellent job in implementing security in this system. Anyone looking to achieve the highest level of security in a Linux environment should carefully examine the Emperor II system.  If they have not done so already, they also need to establish a security department to create and monitor organization-wide security policies.

It can’t be said that any system is truly impenetrable. This is true for reasons relating to the very real threat of internal compromise (e.g. carelessness, poor compliance practices, etc.), technological innovation as well as the subversive efforts of very, very sophisticated and clever people attempting to crack the system. We can say that we think that IBM has done an admirable job in creatively addressing a significant number and breadth of security vulnerabilities and problems. They have made it easier and economically affordable (in cost AND resource utilization) for enterprises of all sizes to use encryption techniques to secure systems and data.

We anticipate IBM’s LinuxONE Emperor II will appeal to high-end enterprises. They are familiar with mainframes and have the staff to manage them. IBM will have to work harder to win over those with less mainframe familiarity and without experienced staff. However, recent surveys indicate that efforts to modernize mainframe management and development tools along with the availability of JAVA, Linux, etc. are attracting new users to mainframes.  

Finally, the security that the system offers will be a powerful incentive for certain customers and the total package of the architecture and its features create a system that can deliver solutions to many customers that they cannot find anywhere else. Congratulations to IBM, we’ll watch and report on how this all develops.





[1] IBM did not provide performance or overhead numbers.
[2] By “standard” we mean that high end Oracle and HPE systems may have a scale up design that eliminates the problem that many x86 systems will encounter.

Monday, September 25, 2017

IBM Research on the road to commercial Quantum Computing

By Rich Ptak




Dario Gil, Vice President AI, IBM Research and Bob Sutor, Vice President AI, Blockchain, and Quantum Solutions, IBM Research recently provided a briefing IBM’s perspective on the state of Quantum Computing. They describe three phases in the evolution of Quantum Computing. They describe IBM efforts and contributions as well as a very recent and significant IBM Research breakthrough on the road to commercializing quantum computing.

The breakthrough is in practical Quantum Computing technology. It marks a significant advance towards commercialization of Quantum Computing. We’ll talk about why in a minute. First a few words about quantum computing. The building blocks of this technology are quantum bits, or qubits, which are the quantum informational equivalent of classical bits, the basis of contemporary computing. Bits have only two states. They are either 0 or 1, i.e. binary – from there all of computing is built.

Individual qubits can exist in much more complex states than simple 0’s and 1’s, storing information in phases and amplitudes. Additionally, the states of multiple qubits can be entangled, meaning that their states are no longer independent of each other. The fact that quantum information can be represented and manipulated in these ways allows us to approach algorithms (instructions that are used to solve problems) fundamentally differently, opening up opportunities for exponentially faster computation. A major challenge to be overcome is how to design algorithms that can make use of these properties to solve problems that are traditionally difficult for conventional machines, like efficiently simulating materials. In this case the molecules at the heart of chemistry and material science.

A cover story article in the September issue of Nature magazine details how IBM researchers demonstrated a highly efficient algorithm that simulates beryllium hydride (BeH2), and then implemented that algorithm on a real quantum computer. This demonstration was the largest molecular simulation on a quantum computer to date. You can link to the article here. Unfortunately, it is behind a paywall, but there are plenty of other highly interesting articles on Quantum technology and other topics available there. IBM’s announcement with a short explanation can be found here. Read the article for more details about the breakthrough.

What matters today to enterprises, business and more

The most enterprise-significant parts of the announcement are in the implication for commercial enterprises. These are exposed in the details of IBM’s vision and focus about the commercialization of Quantum Computing technology. It provides insightful information and structure for making decisions about when to begin investigating Quantum Computing and its potential to affect your enterprise or business.
Image Courtesy of IBM, Inc. 

IBM considers the initial commercialization of Quantum Computing to be within sight. It may be as much as a decade away, but can reasonably be considered to be close enough for some early enterprise movers with interest, resources, and vision to begin exploring the technology and its potential.

Let’s position where Quantum Computing is today. The speakers described three phases of Quantum Computing These are:
·        Phase 1 – development of Quantum Science – interest began in the 1920’s, it wasn’t until the 1970’s that the attention of computer scientists’ attention was captured. This led to a decades-long effort to discover and define the physics of quantum technology and then develop the theories and concepts to build-out the science leading to Quantum Computing technology. Quantum Science underlies the entire field, and will continue as long as there is research to be done to continue to advance the technology.
·         Phase 2 – emergence of Quantum Technology – began May 2016 when IBM provided free access to the first publicly-accessible Quantum Computing prototype, e.g. IBM Q experience, on the IBM Cloud. The opportunity to experiment on a real device led to the creation of new problem-solving tools, algorithms, and even games as real Quantum Computers became accessible to the first wave of users beyond theoretical physicists and computer theoreticians. These new users are practitioners; developers, engineers, thinkers and researchers including scientists, chemists, mathematicians, etc. Their efforts focus on understanding and articulating problems in quantum terms. The phase will end when the now-wider quantum information community discovers the first applications where the use of quantum computing offers an advantage for solving certain classes of problems. This leads to the next phase…
·         Phase 3 – the age of Quantum Advantage – the age of full commercialization of Quantum Computing. It will be marked with the delivery of apps able to fully exploit Quantum Technology to solve commercial problems. Quantum Computing begins to compete, in some areas, with traditional computing methods by offering multiple orders of magnitude increases in processing speeds and computational complexity for certain classes of problems.

Things to keep in mind and conclusions:

Quantum Computing systems that can handle commercial-scale problems don’t exist yet. A considerable amount of research and development work needs to be done before you can begin to contemplate configuring a system of software and infrastructure. But the first serious prototype systems that lay the foundations for the more mature machines of the future do exist. Is it time to begin to develop some understanding of Quantum Computing, how it functions and how it is currently being used?
Quantum Computing will complement, not replace, traditional computing. By its nature, it is best suited to solving certain classes of problems that are traditionally-difficult to solve with conventional machines. These are problems where solving them requires evaluating many alternatives to find the best solution, each of which alternatives may be computationally intensive to evaluate. Today, many problems are addressed (and will remain so) with traditional computing simulation, modeling and statistical analysis, albeit while making simplifying assumptions. For many applications, solutions obtained with traditional computing techniques will be adequate. Also, despite some recent claims, Quantum Computing does not invalidate or decrease the need for recently announced advances in computing security. Such protections will remain critical to secure computing long into the future. 
For other applications, computing alternatives are needed, especially in cases that require simulating quantum behaviors. These include modeling chemical compounds, which requires the ability to predict molecular-level interactions. It is believed that wherever the analysis involves evaluating an incredibly large number of combinations of items, Quantum Computing will have a distinct advantage. Some other examples of nearer-term applications of Quantum Computing include optimization and machine learning.
So, what’s the conclusion? First, as we said, commercialized Quantum Computing is still in the future. It is not ready to address short- or medium-term issues. But, that day is coming. At this stage, most can ignore this technology. But, there also are some that should allocate a portion of their resources (time, budget) to get educated about Quantum Computing. Quantum Computing will realize its biggest advantages when users can define problems in its terms. That requires an understanding of the technology.
Clearly, the level of recommended activity varies with the potential to impact. You need to get a realistic idea of that potential. One approach would be to take advantage of IBM’s offer for free access to its Quantum Computing prototype[1]. Another approach would be to fund a sandbox project, or an off-hours task to learn more about and explore quantum technology.  AND thinking about problems in Quantum terms. IBM is making a considerable amount of resources available to do so, much of which is free, some not.  
In summary, our advice is to concentrate on:
·         Understanding the basics of Quantum Computing approach to determine its potential to impact you and your business. We expect most will find its potential optimization benefits too attractive to resist.
·         Learning about and understanding how Quantum Computing will change how problems are viewed, articulated and programmed for solutions.
·         Considering encouragement of “sandbox” or “off-hours” efforts to learn more about Quantum Computing; formal or informal depending on organizational resources and culture.
·         If the potential impact is significant (and we think it is for many), assign a senior executive the responsibility to keep current on the status of Quantum Computing. 
Finally, there exists no single standard for comparing Quantum Computing status today. The metric of the number of qubits available in an array (that makes up a system) – is insufficient.  For a time, conventional “wisdom” posited it as ‘horse-race’ with more qubits being better.
However, the number of qubits alone don’t work if there isn’t time to execute an algorithm (application) before a qubit array ‘ages’ to a bit and loses the data. A way needs to be found to control/correct such error rates. There are three issues: 1) the life of the qubit array, 2) the time for an algorithm to execute, 3) error correction/avoidance.
Researchers are working on these but no single metric yet exists to measure and relate progress. More about these efforts and other issues appear in IEEE Spectrum and Nature magazine, mentioned earlier. 




Publication Date: September 25, 2017
This document is subject to copyright.  No part of this publication may be reproduced by any method whatsoever without the prior written consent of Ptak Associates LLC. 

To obtain reprint rights contact associates@ptakassociates.com

All trademarks are the property of their respective owners.

While every care has been taken during the preparation of this document to ensure accurate information, the publishers cannot accept responsibility for any errors or omissions.  Hyperlinks included in this paper were available at publication time. 

About Ptak Associates LLC
We cover a breadth of areas to bring you a complete picture of technology trends across the industry. Whether it's Cloud, Mobile, Analytics, Big Data, DevOps, IoT, Cognitive Computing or other emerging trend, we cover these trends with a uniquely deep and broad perspective.

Our clients include industry leaders and dynamic newcomers. We help IT organizations understand and prioritize their needs within the context of present and near-future IT trends, enabling them to apply IT technology to enterprise challenges. We help technology vendors refine strategies, and provide them with both market insight and deliverables that communicate the enterprise values of their services. We support clients with our understanding how their competitors play in their market space, and deliver actionable recommendations.




Friday, August 11, 2017

IBM + Partners breathe new life into Moore’s Law with 5NM chip technology

By Rich Ptak and Bill Moran


When IBM exited the chip foundry business several years ago most industry watchers were sad to see the company go. A key player for decades in semiconductor research, IBM would definitely be missed. We thought that the industry had suffered a real loss. We and others thought statements of IBM’s commitment to making further investments in semiconductor research were to be written off as an essentially meaningless face saving gestures.

As it turns out, we couldn’t have been more wrong. In fact, IBM research continued the work on semiconductor research that it had been doing for nearly 50[1] years.  The IBM-organized consortium of IBM, Global Foundries, and Samsung based at NY State’s SUNY campus in Albany, is delivering a significant breakthrough in semiconductor technology research.  Exiting the chip foundry business was not a sufficient reason for IBM Research to cease its efforts.

Here’s some background on what IBM and its partners have accomplished. Moore’s law[2] says that the number of transistors on a chip will double approximately every two years. The results of that law drove the semiconductor industry for decades.
A meter being roughly a yard.  It might have been more fun if the industry has used the term “nanoyard”. However, as the topic is worldwide technology, the metric system is used, as is the practice in global technical and scientific circles.
Recently, much published commentary (ours and others) discussed how the law was reaching the end of its useful life. A major reason being the physical limits of chip geometry. Incidentally, one of the effects of law is that today’s cellphones (which fit in a pocket) have more processing power than the 1960’s computers used for the Moon visit (which occupied an entire very large room).

To understand what is going on, we need a little computer industry technology background. The industry initially measured processing speeds in seconds. Things moved faster so the term “milliseconds” (one-thousandth of a second) became standard. As the speed-up continued, the “microsecond” (one-millionth of a second) became the standard. One might imagine that things could not get much faster, wrong. Today’s process speeds are measured in “nanoseconds”, i.e. billionths of a second.   

Moving to semiconductors, chip size is measured in terms of the distance between identical features in an array.  The current unit for this distance is “nanometer”, i.e.  a billionth of a meter. The leading edge for productions semiconductor chips today is 10 nanometers.

Moore’s law depends on shrinking the size of the chips while increasing the number of transistors on the chip which increases processing power. Conventional wisdom was that it wouldn’t be possible to push the FinFET technology (which underlies chip manufacturing today) to much smaller chips without losing efficiency. Thus, the comments about the end of Moore’s Law.

However, IBM along with its partners have now developed a process around the Stacked Nanosheet Gate-All-Around Transistor.  This allows eventually building 5 nm chips with improved efficiency, which could not be achieved with FinFETs. The details of the process exceed the scope of this paper. (Those interested can start here[3].) The chart below provides a simplified view of the new IBM process compared with the existing industry standard 10 nm process.


Chart 1 This chart is an adaption of a copyrighted IBM chart.
There are several items of note. IBM's  new chips orient transistors horizontally versus today’s vertical arrangement. This allows transistors to be stacked and have more on each chip. Also, IBM chips use a new way to form the sheet material that is much more efficient in power consumption. It delivers a 75% saving in power compared to existing 10 nm chip architectures.  Finally, the new sheet formation process allows continuous fine-tuning for power and performance of specific circuits during manufacturing. Something not possible with FinFET technology

In summary, it appears to us that the IBM consortium has breathed new life into Moore’s law. With this new architecture, it looks to be applicable for the next decade or so just when many were pronouncing the law dead. However, we expect the interest and investments in such new technologies as quantum, data-centric computing and other approaches to grow.

[1] Publication in 1974 of a paper by Robert Dennard et al on MOSFET scaling rules for improving transistor density. See http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.334.2417
[2] Not actually a law, it was a prediction about how the semiconductor industry would evolve in terms of density and cost per transistor.