RETHINK big news roundup: new memory technology, silicon photonics and software trends
In this edition of our news roundup, we bring you news of innovative memory technologies from Intel, Micron and HP, an update on silicon photonics - and its implications for the field - and a look at Big Data software trends.
HP’s first computer using memristor technology
On October 15 2015, HP joined forces with SanDisk to bring HP’s latest memory technology, the memristor, to market. In July, Intel and Micron also brought a new memory technology which they called 3D XPoint. Both technologies seem to be very similar and also offer similar advantages over previous memory technologies as shown in this Ars Technica article.
HP is competing against Intel to bring their newest memory technology to the server market. The company is incorporating the memristor technology in their newest computer prototype “The Machine” which is intended to be unveiled this year. This computer has several advantages over its predecessors: apart from the memristor, it offers specialized cores and photonic connections. For more information, visit the IEEE website.
3D XPoint Steps Into the Light
In July of last year, Intel and Micron joined forces and released a new kind of memory offering higher performance and lower latency. The first 3D NAND (negative-AND) and 3D XPoint samples will be released and tested soon, meaning that mass production of devices using these technologies will likely take place before long.
Read the full article on the EETimes website.
Samsung begins mass producing world’s fastest DRAM
Recently, Samsung has started producing a new breed of incredibly fast DRAM (Dynamic Random Access Memory) based on the HBM (High Bandwidth Memory) interface. This technology will enable higher performance and energy efficiency, which makes it especially well suited for the HPC (High Performance Computing) and the GPU (Graphics Processing Unit) markets. This will undoubtedly be of special interest in the Big Data world because of the recent hardware shift towards GPU-based machine learning processing, as explained in our previous news roundup.
For more information, visit the Samsung website.
Light at the End of the Silicon Photonics Tunnel
Silicon photonics appears as a very promising area that could enhance several fields; in particular, integrated circuits and networks. Companies such as IBM, HP and Intel are already experimenting with optical components, but further research needs to be done before this technology reaches the mainstream.
Read the full article on The Next Platform website.
Light Chips Could Mean More Energy-Efficient Data Centers
For several years, researchers have strived to create a microprocessor working with light instead of electrical components. Now, the first prototype of a hybrid microprocessor has been developed at Massachusetts Institute of Technology. According to researchers, the data transfer rate inside the microprocessor is from 10 to 50 times faster than the common electronic microprocessor. While optical components have been successfully applied to develop faster networks, they have never been successfully used in microprocessors because of how hard it is to integrate these sorts of components in classic semiconductor-based circuitry.
Read the full article on the Technology Review website.
Spark is overtaking MapReduce. Are you ready?
In a survey of 3,100 information technology professionals carried out in December last year, it was noted that the 22% were using Spark. Apache Spark seems to be overtaking Apache MapReduce since it is both much faster and very flexible. Even IBM called Spark “potentially the most significant open source project of the next decade”, meaning that we can expect MapReduce to be replaced rather sooner than later.
Read the full article on the Tech Target website.
New Hadoop Survey Identifies Big Data Trends to Watch in 2016
A new survey conducted by Syncsort identifies three big trends for 2016. As suggested by the previous news item, Apache Spark will become the next framework to deal with future Big Data challenges; the platform of choice is Hadoop and companies are moving from their current platforms to Hadoop; and Hadoop is becoming increasingly important for advanced use cases.