IT/Datacenter & Super Computing - Page 12

LSI AIS 2012 - LSI beat the rest of the market to release the SAS 6Gb/s controllers that now dominate the datacenter, with over 75% of the channel and 4 out of 5 major OEMs utilizing their current generation products. LSI aims to extend this lead and has shown us some of the initial results for their new 12Gb/s SAS products.

The Invader, referred to as the 9271-8i MegaRAID, sports LSIs latest ROC, the 3108. This chip provides excellent scalability and the potential for speeds that saturate the current PCIe 3.0 bus.

We can also observe the battery backup unit, and also the innovative cache module that provides a fail-safe cache protection for mission critical applications.

Researchers at the University of Southampton have done something that only us mere mortals could only dream of, build a supercomputer from Raspberry Pi's and Lego.

They've called this Iridis-Pi, which is a very small 64-node cluster made from Raspberry Pi's Debian Wheezy distribution, linked through Ethernet. On their lonesome, Raspberry Pi's are not that powerful, but in a cluster with 1TB worth of storage in SD cards, that's another question.

Rackmounting the cluster was done in a very interesting way, where team lead Simon Cox and his son James put the entire array into two towers of Lego. LEGO!!! There are even instructions so you could do this at home if you've got the money for some Raspberry Pi's and some spare Lego around. The entire system cost less than $4,026 or so to make, which is not too bad at all.

IBM has bragging rights at the moment, with the world's fastest server chip clocking in at an incredible 5.5GHz. IBM's new zEnterprise EC12 mainframe cost the company $1 billion in development, and offers 25% more performance courtesy of their hexacore processors.

IBM's zEnterprise EC12 mainframes are available in multiple configurations, with as many as 120 cores available. All models will include transactional execution support, as well as Enhanced-DAT2, allowing 2GB page frames for more efficient utilization of huge quantities of RAM.

Another jewel of the newly-introduced zEnterprise EC12 mainframe is IBM's cryptographic co-processor, Crypto Express4S. It's quite special as its tamper-proof, providing privacy when handling transactions, and other similarly sensitive data. Crypto Express4S also offers multiple security configurations to support the requirements, and needs of bankers and other organizations handling sensitive data. This includes the information on smart passports and ID cards.

Flash Memory Summit 2012 - Micron was demonstrating their newest Micron-designed PCIe SSD controller at the Flash Memory Summit. This controller has allowed Micron to develop the P320h standard PCIe SSD, but also take a new approach to PCIe SSDs.

There are two approaches, the standard SSD that plugs into the PCIe slot, and then a device that looks strikingly similar to a typical 2.5" SSD.

This standard looking SSD is actually connected via a PCIe connection. Even though the connector resembles a standard connection, it is actually an extension of PCIe which allows for superb latency and much higher throughput speed.

Flash Memory Summit 2012 - NVMe is an emerging specification that looks to take over both SCSI and SATA as the go-to specification for SSDs. This new method of accessing SSDs allows for performance above and beyond the current methods of accessing storage. Many of the performance limitations of flash solutions are actually being held back by archaic storage interfaces, and the advent of the NVMe specification looks to remedy that situation.

Of course the first step to achieving this is working silicon, in the form of controllers that are already integrated into usable solutions. IDT has rose to the challenge, beating many others to actually producing the first operating devices capable of supporting this solution that has been in development for years. This demo unit had a rather large fan over the processor to keep it cool during operation.

This board is the higher capacity version that supports more NAND packages onboard a PCIe gen 3.0 compliant interface. The sheer aggregation of flash dies allows for a more linear scaling in performance from the device, leveraging 32 channels via the 89HF32P08AG3 controller.

Flash Memory Summit 2012 - Since we just happened to be in town at the world's only dedicated NAND flash event, we decided a little extracurricular overclocking was in order. Corsair provided us with some of the hottest new technology on the market today, AX1200i, the world's first all digital power supply, eight Neutron GTX SSDs, the world's first Link_A_Media Devices SSDs and two Dominator Platinum 2400C9 kits. Areca tied it all together with a brand new, unreleased ARC-1882ix-12 with 4GB of DDR3 cache, a true PCIe 3.0 RAID controller.

We managed to smash the record with a result of 55,340 PCMarks. PCMark Vantage is the premier storage benchmarking suite used worldwide, so to break the overall record isn't easy.

As per usual when we set out for taking a record there is a large amount of Liquid Nitrogen used, bringing us down to -130C on the processor to allow for much higher CPU clocks. Quite a bit of preparation and test runs went into this adventure, with Mike Guava and Charles Wirth, AKA Fugger, from XtremeSystems taking point.

FMS 2012 - SMART Storage has been making waves lately in the enterprise SSD space with an entire line of drives for every type of application.

The focus at the Flash Memory Summit was on the Optimus SSD, which sports a dual SAS connection. This allows this SSD to pull off 1GB/s in sequential read speed and 500 MB/s in write speed.

A closer look reveals that the SSD has two SAS connectors connecting into the device, which allows the high power of the SSD to be unleashed. This also allows for several types of enterprise features, such as failover and multipath data access.

FMS 2012 - HGST today demonstrated their new 12Gb/s SAS Enterprise SSDs. These are the first 12Gb/s SSDs in existence, and they are in an unbranded case.

We can see the performance improvement between a current tech 6Gb/s SSD, which is pulling down 547 MB/s in sequential read, and the 12Gb/s SSD that is providing 995 MB/s from one device! Nearly 1GB/s of sequential speed from a single SSD is simply unheard of with today's interface.

The 6Gb/s, in the foreground, and the 12Gb/s SSD are both of similar size, so the performance increase does not involve a larger device.

One of the neatest aspects of attending industry shows like the Flash Memory Summit is that you never know just what the guy sitting next to you has in his pocket. While attending the seminar on NVMe (an emerging specification and subject of another article soon), we managed to strike up a conversation with a fellow attendee. After mentioning that we had previously done a write up of the Intel 910, which he had also happened to have read, he produced something that he claims to be much more interesting.

Considering that this device is unreleased and holds 2.2 TB of raw NAND flash, he most certainly did open our eyes. This is the Virident FlashMAX II, which is a PCIe SSD aimed for the datacenter. With read speeds of up to 2.7 GB/s and write speeds up to 1.6GB/s, these cards deliver monstrous performance. 1.5 Million IOPS of random read is simply astounding, especially considering the attention given to solid, sustainable performance over a long period of time. With MLC NAND, and a hefty bit of overprovisioning, Virident is looking to bring sustainable performance to their customers.

Unfortunately their primary competitors can suffer high levels of performance degradation after extended use. The Virident FlashMAX II is designed specifically to avoid these pitfalls, operating at full speed even after years of continuous load.

The flick has been switched for the most powerful GPU supercomputer, Emerald, at the Science and Technology Facilities Council's Rutherford Appleton Laboratory (RAL) in Oxfordshire, U.K., the two systems working together "will give businesses and academics unprecedented access to their super-fast processing capability".

The insane amounts of power will allow researchers to run simulations that range from health care to astrophysics. The supercomputer combo will be used to look at the Tamiflu vaccine's effect on swine flu, Square Kilometre Array project data, climate change modelling and 3G/4G communications modelling. The official launch of the e-Infrastructure South Consortium took place at the same time, coinciding with Emerald's unveiling.

This is a consortium of four U.K. universities, Bristol, Oxford, University College London and Southampton, who will collaborate with RAL and work with the supercomputers. The Engineering and Physical Sciences Research Council (EPSRC) funded the supercomputers with a £3.7 million grant. The EPSRC press release has a tonne of details and specifications for the supercomputers, and also states:

SMART Storage Systems has announced their newest entry into their Enterprise SSD family, the Optimus Ultra+. The 'Ultra' part of the name comes from two central facets of performance, speed and endurance. This SAS 6/Gb/s SSD sports some impressive numbers, with 100,000 random read IOPS and 60,000 random write IOPS. The SSD also supports dual-port SAS, which allows the throughput to climb to an unheard of 1GB/s.

The real story here is the endurance however. The Optimus Ultra+ is rated for 50 Drive Writes Per Day (DWPD) for five years. This means that the capacity of the drive can be written and deleted 50 times every day for five years.

The Optimus family of SSDs all centers around one philosophy; providing SLC-like endurance with MLC pricing. The attraction of MLC over SLC is simple; SLC commands ridiculously high prices while MLC is becoming garden-variety. Even in the consumer market we are now seeing MLC drop below the dollar per GB threshold. This low price level will always be welcome in any market, but creating Enterprise-class MLC is not an easy task.

Just when you thought tape was dead, the National Center for Supercomputing Applications is getting ready to build a new storage infrastructure that will include 380 petabytes (PB) of magnetic tape capacity which will be backed up by 25 petabytes of online disk storage made up from 17,000 SATA drives.

The new storage infrastructure is said to be built to support one of the world's most powerful supercomputers, Blue Waters. Blue Waters was commissioned by the National Science Foundation (NSF), and is expected to have a peak performance of 11.5 petaflops. The NCSA says that they're building the system to:

Microsoft has announced a victory in the MinuteSort test. They claim to have tripled the amount of data sorted by the previous record holder, a Yahoo team. MinuteSort is a test to see how much data can be sorted in just a mere 60 seconds. As more data moves into the cloud, this ability to sort data quickly becomes a bigger and bigger issue.

According to Microsoft's post on TechNet, "In raw numbers, the team's system sorted 1401 gigabytes in just 60 seconds - using 1033 disks across 250 machines." This hardware compared to what Yahoo ran is roughly "one-sixth of the hardware resources" and managed to sort around 3 times as much data. You can see that the Microsoft solution is much more efficient.

Additionally, it's interesting to note that Microsoft Research didn't use Hadoop as one might expect. Instead, the researchers at Microsoft created a new system called "Flat Datacenter Storage." The "flat" portion is the important part of the system. Microsoft explains:

Google and green, it goes hand-in-hand and their next data center will be built with energy savings in mind. Google have previously been good at this with other data centers that are energy-efficient and green. Their latest data center to be built in Taiwan will use thermal energy storage.

Thermal energy storage systems commonly use chilled liquid or ice to act as a thermal battery, enabling a data center operator to run air conditioning at night (when rates are obviously cheaper) and during the day, pump the chilled liquid around the facility for cooling.

Increasing electricity rates in Taiwan will be a big reason for Google to tap the thermal storage solution, where they can skip the peak power rates at night and just use liquid or ice as its also cleaner, and a longer lasting way to store energy rather than using batteries. A Google exec has cited the the increasing electricity rates in Taiwan is a reason for building the new system, and also notes that the new Taiwan-based data center will use 50-percent less energy than typical facilities.

Over the next five years, IBM is set to work with the Netherland's National Institute of Radio Astronomy (ASTRON) where tehy hope to develop a low-powered, exascale supercomputer. Not impressed yet? Hold onto your chair, dear reader. According to IBM, this supercomputer would be millions of times faster than today's high-end desktop PCs, and possibly thousands of times faster than even the most recent super computers.

The exascale computer would be used to analyze data collected by SKA (square-kilometer array), which is a cutting-edge radio telescope set to become the largest and most sensitive of its kind ever built. ASTRON hopes to have the telescope ready by 2024. While it's still a fair way off, the excitement will only build over time.

Now, this is where you don your math hat, and get ready to have your eyes widen a little: to compare to what we know, and use now, exascale refers to a computing device that is just incredibly fast, where the number of floating-point operations per second it can perform isn't measured by gigaflops or even petaflops, but exaflops. Today's highest-end desktop CPUs rank up around 20 gigaflops, not that impressive in terms of scale to this beast.