Carbon nanotubes is pretty much considered as the world’s next generation magic material. Working wonders in many different aspects of system design and engineering, it is no surprise that its potential applications eventually extended in electronics and computing.
While carbon nanotube transistors have been conceptualized as early as 1998, carbon nanotube-based memory storage is still a relatively new concept. It was only during the early 2000’s, when Boston-based Nantero first unveiled a proof-of-concept idea for Nanotube RAM, or NRAM/Nano-RAM. In fact, at was just as recent as August this year when the company finally partnered with Japanese consumer tech giant Fujitsu, in a commercial venture that will bring NRAM to the tech industry.
But, what is NRAM? Why is it very important as a new memory storage medium? Here are a few pointers that will explain its paramount influence in the future of computers.
Why Current Computer Memory is Inefficient
As we all already know, memory storage is divided into two main categories: volatile and non-volatile. Essentially, volatile memory loses its data when disconnected from a power source, while non-volatile retains data even when it’s not operational. There are several advantages to using both types, which mostly boils down either dynamic access (DRAM) or data integrity and security (HDD).
There have been alternative technologies that allow dual functionality as both volatile and non-volatile memory (design and capability-wise), such as SSDs. But one thing is always constant: there will always be a technical caveat. Just as there are advantages, there are also disadvantages, such as power dependence, operational complexity, data transfer limitations, even implementation costs. Not one memory storage technology today could address each others’ problems without having design issues of its own. So, we usually end up with a weird mix of different computer system configurations, using all of these technologies to optimize what each memory type could offer.
NRAM, and Its Potential
This inevitably leads us to the introduction to Nanotube RAM. NRAM is a type of Resistive RAM (RRAM), which means that it uses varying degrees of electric resistance to generate data values. The binary values are represented by the physical distance of each individual nanocarbon memory cell, higher voltages separates the cells generating a ‘0’, while lower voltages joins the cells together, generating a ‘1’.
Because of the system’s use of carbon nanotubes, NRAM has several distinct advantages over current memory systems, which are namely:
- Ultra Low Power – A functional Nano-RAM chip or card will only consume 1 to 15 μA (microamperes) to write data. To put that into perspective, the already relatively energy efficient NAND flash memory requires at least a few mA (milliamperes) on average to read and write data. Comparing it to volatile DRAM of course, is directly out of the question.
- Improved Read/Write Cycles – Despite its overwhelming advantages, one of the biggest disadvantages of flash memory is its write cycle limitation, which is usually only about 10,000 to 100,000 on most consumer level devices. With NRAM though, even with its prototypical iterations using a 140-nm architecture 4Mbit module, it can already at least crunch more than a billion (1,000,000,000) read/write cycles, without even showing signs of physical deterioration.
- Near Perfect Data Retention – Data preserved in NRAM can last for more than a decade, even when constantly subjected to temperatures as high as a 300°C (572°F). Most memory systems today could not maintain 100% data retention in the same time period even when exposed to standard room temperatures.
The RAM of RAMs
Magnetoresistive RAM (MRAM) and Phase-Change RAM (PRAM) are two similar technologies that currently vie to replace standard memory storage media. PRAM touts its significantly faster write times and more flexible scalability, due to the simpler switching method that only requires temperature changes. MRAM has the more ambitious objective of becoming the “universal RAM”, sharing the same data retention and power draw advantages NRAM offers with its magnetic memory cell architecture.
But despite being decades older than NRAM, both never managed to extend their industrial applications beyond niche roles. Fabrication restrictions aside, both technologies still suffer from inherent weaknesses stemming directly from their basic design configurations. PRAM for instance, will always have issues with its thermal switching process, since cell value integrity may be compromised due to the comparatively gradient nature of temperature itself. MRAM will always be designed with much bigger memory modules, due to the basic complexity of a magnetism-based cell architecture.
As such, the aforementioned positive attributes of NRAM extend not only to currently used technologies within the industry, but also to other breakthrough memory storage concepts as well. The RAM of RAMs so to speak, at least in a directly technical manner.
Memory to Good to Be True?
Ultra low power, heat resistance, non-volatile data retention, and near infinite data cycles. Based on the results previously presented by Nantero, NRAM fills every single category, without leaving room for any inherent operational weakness or design disadvantage. This leaves us with three external categories, implementation, architecture, and cost, and this is perhaps where NRAM finally meets its limitations.
First, carbon nanotube manufacturing is a technical challenge that is yet to be perfected. Specifically speaking, its production technologies cannot compete with silicon just yet. This significantly increases cost, and thus its price point to the market. Second, it does not have the decades long development history other memory technologies had. Its design architecture is yet to be fine tuned properly to be optimized for efficiency. Finally, introducing such new memory medium to various systems and devices will require huge changes in fabrication, manufacturing, and distribution procedures. As such, certain design alterations need to be considered in order to integrate it with existing related technologies.
Despite such challenges however, NRAM is making steady progress at the industry front. Fujitsu plans to introduce a 55-nm architecture design, and use it to manufacture the first commercially available NRAM modules. Nantero has also slightly revealed that development is also in the works to manufacture a DIMM version, a DDR4 chip that might eventually introduce NRAM cards for standard computers.
Most importantly, NRAM’s development schedule outlines its eventual implementation as a three-dimensional multi-layer module, which would exponentially increase its already enhanced performance rates even further.