Data center efficiency is a hot topic as enterprises add more servers and make them into virtual servers to squeeze as much processing power as possible from their infrastructure. Network scalability is a critical success factor in this evolution, and switch chips from Intel® are playing an expanding role in making possible large-scale data center networking.

 

Highly scalable networks are critical to both the cost effectiveness and the performance of server installations. The networking industry is converging on the “fat-tree” architecture as the way to build data center networks that provide scalability and fine-grain fault protection. Fat-tree networks are flat and provide any-to-any connectivity. Intel’s low latency Ethernet switch chips are built to meet the needs of large-scale data center networking.

 

For example, our first-generation, 24-port, Intel® FM2000 series 10 Gbps Ethernet switches pioneered the notion of building low-latency, non-blocking fat tree networks using Ethernet. These layer-2 switches feature latency of 200 nS and can be used in two-tier networks that can scale to 288 10 GbE ports. In the second-generation, layer-3 Intel® FM4000 series switch ICs, we dramatically improved the efficiency of distributing flows to minimize hot spots, while also implementing in-band management. While scalability was maintained at up to 288-port fat trees, all of these innovations dramatically improved the performance of the data center fabric and set the stage for the third-generation Intel® FM6000 series.

 

With the Intel FM6000 series switches, latency is the same, port count (or the radix of a scalable fabric) has tripled to 72 ports and we’ve made further improvements for performance and scalability of the fabric, delivering up to 2500 non-blocking 10 Gbps ports in the same two-tier fat tree structure. Additional Intel FM6000 features include programmable hash algorithms with multiple active profiles, enhanced in-band management, 10 Gbps serial links with embedded PHYs, robust congestion management, and programmable tunneling capabilities. Others in the industry may claim that the only way to achieve scale is with proprietary fabrics; we have proven that it’s possible with plain old Ethernet.

 

Scalability is a critical issue and Intel is delivering it in a way that will enable Ethernet to power the next great generation of high scale data centers.

Arista Networks & Intel Present:

Hadoop for the Common Man:

A practical working guide to building a cost-effective 10GBASE-T foundation

Thursday, March 29, 2012
10:00a.m.-11:00a.m. PST (GMT -8)

Duration: 1 Hour

Register Now

Business analytics represent a significant IT challenge. Large volumes of high quality data across disparate systems, with computational based search, analysis and data summarizations are necessary for providing actionable data to business groups. Over the last several years many companies have begun deploying Hadoop clustering technologies in response to these challenges.

While these Hadoop clusters offer faster and more reliable data results, several of the high performance networking and compute technologies, often cited when deploying these clusters, have been cost prohibitive and operationally disruptive for the broader data center consumer.

This webinar will discuss key networking technology shifts that substantially lower the barriers to entry for deploying Hadoop clusters. Specific topics include advances in 10GBASE-T server connectivity, scalable switching, and I/O optimization within the newest generations of Intel® Xeon® processors. Additionally, this webinar will discuss how to get started with a real world, performance proven, cluster example. Intel and Arista Networks will lead these discussions and will end with a web based question and answer session.

Attendees of this Webinar will also learn about:

  • The need for deep switch packet buffers when moving data files
  • I/O optimization requirements for MapTask and ReduceTask operations
  • Best practices and distance limitations specific to 10GBASE-T cabling
  • Topology designs for east/west Hadoop traffic patterns
Featured SpeakerFeatured Speaker
mark.jpgMatt.jpg
Mark Berly
Sr. Systems Engineer
Arista Networks
Mathew Eszenyi
Product Marketing Engineer
Intel

RegisterButton.jpg

     Time, tide, and embedded waits for no one. It’s been a while since we updated our release package for Windows Embedded operating systems, and I’m happy to announce that it is live. Our partners at the EDC were kind enough to host the file again, so start your downloading!  Okay, hold onto that thought for a second and let me explain a couple of things first.


     1)      Windows Embedded Compact 7 support.  We have native NDIS6-based drivers for WEC7. The drivers for WEC7 do not match the usual driver naming convention.  We have placed several technologies into one binary, the e1i.  See the file “DriverSelectionGuide.txt” for details about which silicon is supported by the e1i. If your favorite isn’t listed, continue to use the CE6 drivers. For Windows Embedded Standard 7, WES7, please build the image using the provided builder, THEN install the normal Win7 drivers from our normal release media.

     2)     Embedded Teaming support. We have XP Embedded teaming support. Just use these INFs with the normal release media ANS binaries. If you run into issues, reproduce with normal XP first since XPe can be built without a lot of packages that can cause false failures. Again build the image first, then install ANS.

     3)     Legacy 10/100 products get a swan song. While they are still on this release media, they may not be for long. Officially we have ended their life, and that means they can end up off the media at any time. I’m working on getting the source for a few 10/100 drivers published, but that is a ways out yet. If you want to be a beta tester for it, send me a message. This is only for 10/100--don’t expect or ask for others yet, please. Gotta shake out the system and an EOL product line is great for that. Thanks in advance for your patience and super kudos for many happy 10/100 years.

     4)     When is the next Embedded release? Good question. If you use Embedded Windows with Intel® Ethernet, call your disti, Intel sales rep or local Intel field person and demand they keep the support coming. Without your inputs, it might be a while, and it was already too long this time.


     Okay that is enough for now.  Here is the link again. Thanks for your patience, and leave a note if you use the package and want to see it regularly updated.


Thanks for using Intel Ethernet.

dougb

Even more ECC updates

Posted by dougb Mar 16, 2012

ECC has many real "definitions" - error correcting circuits, error correcting code, or error correction code - but they all do the same thing.  It helps keep data intact within the chip memory.   ECC uses a special algorithm to encode information in a block of bits that contains sufficient detail to permit the recovery of a single bit error in the safeguarded data.  This protocol will not only detect single bit errors, but will transparently correct them on the fly.   Double errors will be flagged as an error and the device will try to get software’s attention about it.  Related to ECC is parity.  Parity will keep track of the quantity of bits in total and track them as either even or odd.  Should this parity change while it is in the chip memory, it will be flagged as an error.  Since you can’t tell which bit went rogue, this is a poor man's protection. Also, if more than 1 bit changes, parity check can miss it.

(Warning HTML table!)

 

Product

Packet Buffer

(In band Traffic)

Managability

(out of band traffic)

Datasheet Info

X540

ECC

ECC

7.14.1

82599

ECC

ECC

7.14.1

82598

ECC

ECC

Look for DHER/PBUR

I350

ECC

ECC

7.6

82580

ECC

ECC

7.23

82575 and 82576

ECC

ECC

7.6

82571

ECC

Parity

13.7

82573 /  82574 / 82583

None

None

n/a

82546

None

None

n/a


Both ECC and parity have a basic limitation in that if the error is large enough, it will look okay.  ECC is far more resistant to this.  We try to make sure bad things don't happen to your data, but it still might happen.  And while it will try to tell you when it does go bad, sometimes it still won't notice.  That's why our lawyers care about articles like this.  Multiple bit errors are very rare and probably will cause other problems to the machine.  Data integrity isn't made with a single point safety net.  If you want to guarantee your data, use a multiple layered approach since it’s unlikely that all of them will fail.

Today we announce the arrival of our newest product:  the Intel® Ethernet Controller X540

 

What is it?

The Intel® Ethernet Controller X540 is the next generation of 10 Gigabit Ethernet (10GbE) controllers from Intel. Using a 40 nm manufacturing process, the X540 is the first 40 nm, dual-port integrated Media Access Controller/Physical Layer (MAC/PHY) single chip designed for reduced power and package size that is designed for use as a LAN on Motherboard (LOM) network controller. It also powers the Intel® Ethernet Converged Network Adapter X540.

 

Key Features

  • The world’s first fully integrated single chip 10GBASE-T Ethernet Controller specifically optimized to bring 10 GbE networking to server boards as a LOM.
  • Designed for LOM in mainstream rack and tower servers. – Support up to 100m over widely implemented Cat 6A cables
  • Backwards compatible with existing 1 GbE infrastructure, providing a seamless (or easy) upgrade path to 10GbE
  • Provides industry leading features for I/O Virtualization and Storage over Ethernet, including iSCSI and FCoE.
  • Low power <12.5 W
  • Small package 25 x 25 mm,
  • Designed using the latest 40 nm PHY technology with industry-leading integrated Electrical Mechanical Interference/Remedial Frequency Interference (EMI/RFI) filters.

 

Intel® Ethernet Unified Networking Principles

Intel has delivered high quality Ethernet products for over 30 years, and our Unified Networking solutions are built on the original principles that made us successful in Ethernet:

  • Open architecture integrates networking with the server, enabling Information Technology (IT) managers to reduce complexity and overhead while enabling a flexible and scalable data center network.
  • Intelligent offloads lower cost and power while delivering the application performance that customers expect.
  • Proven Ethernet unified networking is built on trusted Intel Ethernet technology, enabling customers to deploy Fiber Channel over Ethernet (FCoE) or Internet Small Computer System Interface (iSCSI) while maintaining the quality of their traditional Ethernet networks.

 

Intel’s unified networking solutions are enabled through a combination of standard Intel® Ethernet products along with trusted network protocols integrated in the operating systems. Thus, unified networking is available on every server either through LOM implementation or via an add-in Converged Network Adapter or Network Interface Card (NIC).

Filter Blog

By author: By date:
By tag: