Recently several members of the analyst and press community had the opportunity to attend an HP Labs event on Energy and Infrastructure. Prith Banerjee, senior vice president of research at HP and director of HP Labs, laid out the foundation of HP's vision for sustainability. I met with representatives from numerous labs including: color reflective display technology, real-time streaming analytics, information and quantum systems, sustainable IT, exa-scale data center technology and advanced wireless networking. But the big news for data centers was in memristors.
All the meetings offered valuable insight into the future of computing. Stan Williams, HP Fellow and Director of the Information & Quantum Systems Lab, discussed two fundamentally game changing and related technical efforts that his labs are working on:
For the purpose of this blog, I will focus on memristors and its initial application. The original idea of the memristor, which is an abbreviation for memory resistors, was
defined in the 1970's by Leon Chua. HP was the first to build a working memristor in 2008. Today, capacitors, inductors and resistors form the three basic circuit elements with the memristor now being considered the fourth fundamental circuit element. The memristor differs from the other circuit elements in that it can retain memory without power. This leads to three critical questions: Why does this matter? What is the practical application for it? When does it leave the Labs? First, memristors will offer dramatic improvements in storage (e.g. days of video on the form factor the size of a thumb drive). If Memristor production is successful, it could be replace flash memory as we know it. Mr Williams said it would be approximately
3 years before we could expect commercial prototype but at that time he expects memristors to offer storage density of about 20 gigabytes per square centimeter. HP expects memristors to be faster, lower powered and to offer at least twice what comparable Flash technology could offer in 2013. If HP is right, they could change the Flash market overnight and make a mint on licensing fees. Imagine how an increase in storage with a reduction in power consumption could change the capabilities of small factor consumer electronic devices such as cameras, e-readers and smartphones.
But Mr. Williams also offered a longer term, roughly 8 year vision, of how memristors could change the computing landscape by integrating offer both logic and storage. This April, in the science journal
Nature, HP published advances in its research study in nature titled
"'Memristive' switches enable 'stateful' logic operations via material implication"
This means memristors could replace transistors on processing chips. Mr. Williams painted avision of memristors as a bridge that brings storage and computing closer together. Over time, memristors could entirely flatten data center architectures by enabling computation to be performed in chips where data is stored, rather than on a specialized CPU.
This is important because space and power are such huge factors in the cost of data centers. Some sort of memristor revolution could result in data centers that require less square footage and less power even if it is no longer possible to make transitors smaller. In short, it could spark a revolution in the data center as well as in the CE field.
AT&T finished last in the March survey of North American mobile providers. Sprint finished second with 2.4% of calls dropped, and T-Mobile finished third with 2.8% of calls dropped. The trend shows an increasing number of dropped calls among AT&T customers surveyed and a decreasing number of dropped calls for Verizon customers.
