| 2013-04 Maths, CPU Debate, Storage, Big Data |
|
| March 2013 | |
|
Compucon CPD Seminars
CPU Update & Technology Debate The Intel Tick-Talk clock is about to talk again any time now. Intel has disclosed some feature and spec info on Haswell which is the 4th generation Multiple Core technology. What else can be new these days apart from having more cores than before? Does AMD sit still? AMD still maintains leadership in graphics from the CPU and will release 8 cores processors for the middle market soon. ARM is twisting a lot of arms these days and is even going into the server space. We will examine a University of Wisconsin debate report on Intel versus ARM. Compucon Storage Direction Compucon has been offering reliable large data storage computer systems for over 10 years. In the meantime, we are preparing to introduce a software based storage systems as an extension of our coverage. It is ZFS and not software or host RAID. ZFS was developed by Sun Microsystems and has been formally incorporated in Solaris 11 for enterprise class of servers. Our team has looked into 3 implementations of ZFS, tested their performance and user friendliness, learned about their eco-systems, and built up our readiness to support soon. Our roadmap is scalable from TB to PB and ZB and includes DAS, NAS and SAN. Edmond will take this opportunity of showing a new and stylist Compucon keyboard. Big Data Computer Hardware
Does New Zealand have a market for big data systems and how big is big? Let us define big here to be a PC system with 10 hard disks or more. Then we know there are many thousands of customers going into tens of thousands over the next ten years. This session attempts to assist our peers to be ready when the opportunities are spotted. We will explain a simple one with 24 disks in a chassis, a design with multiple PC in a single chassis, and finally to blade servers with multiple processing and storage blades in a single chassis. Pedestal takes a max of 10 disks. We rely on racks to take more disks as well as for better economies. We are aware that there is some overlap of big data with high performance computing. |
