• Nginx, web server has seen a huge growth in it's popularity due to its ability to easily handle high-volume traffic. Nginx next release, due in early 2012, will be more pliable for shared hosting environments.
• The OpenStack cloud computing platform has gained support from several technology firms due to its scalability.

• We've discussed various NoSql technologies through out the year. Perhaps 2012 might be the year of Stig a nosql db for social networking sites. The data store's architecture allows for inferential searching, enabling users and applications to look for connections between disparate pieces of information.

• A storage system for the scalable, distributed age GlusterFS is one of the fastest growing storage software systems on the market, as downloads have increased by 300 percent in the last year.

In a recent (youtube) talk, see link below twitter demonstrated the smart use of Sql (MySql) and no-sql ( Hadoop, Cassandra , Vertica) technologies.

→ Read more...

JPMorgan has once again deployed a new FPGA-based supercomputer to help in it's efforts to process and analyse data. The previous supercomputer was used to speed up batch based processes, specifically around the end-of-day risk calculation. They claimed a time reduction from 8 hours to 238 seconds (a reduction of approx 99.2%).

→ Read more...

Details are emerging of China's latest Super computer, called the Sunway BlueLight MPP. It was installed on September at the National Supercomputer Center in Jinan, China. It will use in part 8,700 ShenWei SW1600 chips.

The ShenWei SW1600 is the third generation CPU by Jiāngnán Computing Research Lab. It achieves 140 GFLOPS floating point performance from its 16 cores RISC architecture. The CPU is a national key collaborative laboratory project by Jiāngnán Computing Research Lab and High Performance Services & Storage Technologies.

In addition the Sunway installation uses a sophisticated water-cooling system.

The UK-based mobile chip giant Arm announced a smaller, cheaper, more efficient Cortex-A7. It's efficiency is due in part to the 28nm process that makes it five times smaller than the current offerings like the Cortex-A8.

However that was not the biggest surprise, the new A7 can also be combined with much higher-power cores like the Cortex-A15 on the same chip. This will allow the processing unit to switch between two totally different processing units depending on how much power is needed at the time. Perhaps the least imaginative part is their name of “Big.LITTLE” computing,”

Netcraft server report shows a significant up-tick in the usage of Nginx. This bares out experience where
Nginx is used as a fast reverse proxy and static data server for highly scalable applications

via Netcraft Survey Report

In a spate of liquid nitrogen -200 degree C madness AMD pushed a version of the 990FX and AMD FX processor to a Guiness World Record with a frequency of 8.429GHz. While primarily a bit of dramatic promotion it is still impressive achievement for what is essentially a consumer desktop processor.

Video after the link:

→ Read more...

The default java monitoirng tool JConsole can be quick (cheap) way to get insight into a performance test run or current poduction performance issues. A nice open source addition is the JConsole plugin for Hibernate that will help make greater sense of the Hibernate queries that are being generated.

http://hibernate-jcons.sourceforge.net/

Micron's Hybrid Memory Cube - HMC modules have demonstrated speeds 10 times higher than current DDR3 memory modules where DDR3-1600 is capable of data speeds up to 12.8 gigabyte per second and promises up to 20 faster.

While not ready for consumer release Micron claims to be able to offer 20 times higher bandwidth as DDR3 and do it with only 10% of the energy needed to power modern modules.

I totally missed that pNfs got baked into the new Linux 3 Kernel. PNfs or Parallel Network File System is a huge boost in Network File system access. While the headlines often get written by micro second performance increases in micro processors users (or programs) often wait seconds for data to be loaded of disks.

With pNFS we can split the file server into two types

1. the metadata (controlling) server
2. as many storage servers as you can afford (or as many network ports as you can manage)

In a simple example if we have 10 storage servers for a 10 node cluster, you will see something close to a 10x increase in speed. 100 of each and you’ll see close to 100x increase etc.

And if that hadn't rocked your world, you might be surprised that it is also *backward compatible* with existing NFS data even those stored on a standard PC and transparent to applications running on it.

It was originally hosted on the esa.int site, but seems to have vanished. We've mirrored it, and can download a copy here: javiermartinez.pdf

Original Link : http://www.sciops.esa.int/SD/CSG/EGW08/JavierMartinez.pdf

While dramatically faster for inserts and retrieval NoSql can be very frustrating executing more complex queries, easily performed by their relational cousins. Now the developers of CouchDB and SQLite comes a new query language specification, UnQL (pronounced “Uncle”), which provides an SQL like query language for working with NoSql Databases.

Unql aims to build upon existing SQL like syntax, supplementing it with concepts appropriate for the unstructured, self-describing data formats

The language includes typical SQL actions like SELECT, INSERT, UPDATE and DELETE commands, where they work on collections of unordered sets of documents (and not tables of SQL world). In UnQL, a document is an object that can be described in JSON.

The specification, it is promised, will be placed in the public domain when complete.

http://wwww.unqlspec.org/display/UnQL/Home

Joseph Smarr, a technical lead on Google+ spoke regarding some of the network’s technical details in an AnyAsk interview.

According to Smarr, Google+ uses Java with Guice on the server side and the JavaScript relies heavily on the open source cross-browser Closure framework. Closure is a set of tools for JavaScript including a templating framework (soy), a JavaScript “compiler” which is really a JavaScript optimizer that rewrites JavaScript code to make it faster and more compact and a set of javascript ui elements similar to Extjs. Missing from the Google party appears to be the other google javascript framework GWT, which was used for Wave.

Smarr revealed for speed enhancement the Closure templates are often rendered on the server so that “the page renders before any JavaScript is loaded, then the JavaScript finds the right DOM nodes and hooks up event handlers, etc. to make it responsive.”

Storage and lookup is built on BigTable and Colossus, Google’s file system

Google has said they will reveal more of Google+'s technology in the near future.