Fulcrum Microsystems Blog

By Gary Lee

In my last post, I introduced the new FlexPipe™ packet-processing pipeline, a high performance packet engine for Fulcrum’s Alta High-Speed Ethernet switch architecture that can provide sustained performance of a billion packets per second while maintaining less than 300ns latency.  In this post, I want to cover the microcode support built into FlexPipe that delivers a new level of flexibility for supporting current and future standards and protocols without sacrificing performance.

Data center standards are constantly evolving, leaving chip makers to either wait to hard-code the protocols in the device, or to use lower performance programmable silicon and count on a software-based enhancement after the chip has been designed into a system.  Neither is ideal. But the support for microcode built into FlexPipe gives future Alta chips the performance of a hard-coded solution with the flexibility of a software solution.

Typically, when one hears the term microcode and programmable, it is assumed that the architecture is a run-to-completion or non-deterministic model.  FlexPipe operates in a deterministic manner, meaning that with any possible microcode implementation, the engine will maintain more than one billion packets per second of throughput performance while maintaining less than 300ns of latency in cut-through mode.

With the use of microcode, we can provide customers with a rich set of flexible features that can be adapted to changing market needs. Our customers can get to market early, before industry groups finalize standards, allowing future-proof system designs. It also allows us to anticipate standards that have not even been thought of yet.   Some of the known standards where the programmability comes into play include RBridge/Trill, Shortest Path Bridging, VEB, VEPA, VEPA+ multi-channel, VN-Tag, Fibre Channel, LISP or other proprietary switching headers for vertical markets.  We also plan to support OpenFlow, and will be able to keep up with this implementation as it evolves.

FlexPipe is the key differentiator for future Alta chips, giving them the performance and flexibility they need to deliver the top-of-rack switch performance necessary for high-port count virtualized networks.

Comments welcome: feedback@fulcrummicro.com

By Gary Lee

A billion packets per second.  That’s the performance of FlexPipe™, the new packet-processing engine that’s part of the Alta high-speed Ethernet switching architecture announced by Fulcrum CTO Uri Cummings at Hot Interconnects last week.

What network switch needs a billion packets per second performance?  Let’s stop first to recognize just what we’re talking about because a billion of anything is hard to visualize.

In months it would equal 82 million years (back when dinosaurs ruled the earth).  In people it would equal 1/8 of the world’s population (or the population of China in 1981).  In apps it would equal the number of downloads from the iTunes app store in the first nine months of operation.  In dollars, it’s the taxes paid by you and 103,424 of your friends.

A billion packets per second is 3X the performance of Fulcrum’s current generation of FocalPoint switch and it means that Alta-based switches will have the performance needed for applications such as virtualized data center networks.  Virtualization is increasing the density of server farms and allowing data center operators to efficiently deploy cloud services. Virtualization technology is rapidly changing, and is spreading from virtualized servers to virtualized storage, and its being pushed down into hardware by technologies like Intel’s SR-IOV.  The market is attracting acquisition dollars, and also venture capital.  There are a lot of big bets on the future of virtualization.

Those bets can pay off if the network performance is there.  FlexPipe is a pipeline that’s wide enough to support full rate, non-blocking performance for a large number of virtualized flows. And it can deliver these performance levels with all pipeline features enabled. FlexPipe brings with it FocalPoint’s 300ns of latency, which has been on the market for five years and is still five times lower than other data center Ethernet switches.

There’s much more to the FlexPipe, including the ‘Flex’ part, which refers to new programmability features that will be discussed in my next blog.

Comments Welcome: feedback@fulcrummicro.com

By Gary Lee

At Hot Interconnects today, our CTO Uri Cummings outlined the next-generation Alta 40Gb switch chip architecture. Today’s architecture-level discussion gives us the opportunity to focus on key design issues that really make a difference in the performance of networks, but often get lost in the focus on speeds and feeds that occurs when a switch chip is announced.  Fully provisioned throughput, low latency, efficient memory architecture and programmability matter as much (or more) than total port bandwidth.

This is especially true now that there is a sea change in the Ethernet market and the need for high-performance, low latency switching in interconnect applications is expanding as more data centers embrace Ethernet as a compute fabric, and as the economics of Ethernet becomes more essential to building out cloud computing services.

So, what is the Alta architecture?  Alta is a 40Gbps switch with integrated 10GB PHYs that keeps all of the benefits of the FocalPoint switch family, including 200 ns cut-through latency, which is still the lowest latency in an HSE switch to date.  The architecture has full support of data center bridging features, advanced header processing functionality and the parallel-multicast capability that makes it popular in video and Wall Street applications.

Alta marks the debut of a significant performance innovation called FlexPipe™. FlexPipe is the industry’s fastest packet processing engine with performance greater than a billion packets per second even with all pipeline features enabled. But FlexPipe is also programmable, providing the ability to reconfigure various locations in the frame-processing pipeline using microcode.  This enables flexible support for evolving data center standards such as TRILL, FCoE, MPLS and others.  Another feature of FlexPipe is deterministic performance.  FlexPipe is such an innovation that we will break down its features in two upcoming blog posts.

With Alta, FlexPipe is combined with the output queued, shared memory architecture we’ve used in FocalPoint for several years now.  This shared memory design is based on our proprietary RapidArray packet memory.  RapidArray is a key reason for the latency and throughput performance of FocalPoint.  More on this memory design is in another blog post located here.

Alta is designed to support a data center evolution that is moving from 10Gb to 40Gb, moving to support thousands of virtual servers and merging separate data and storage networks into one unified fabric.  Look here for more on Alta later this week and next week, and for the first Alta-based chip announcements in Nov.

Comments Welcome: feedback@fulcrummicro.com

By Fulcrum Microsystems

The Hot Interconnects show is coming up next week, and Fulcrum co-founder and CTO Uri Cummings will use the opportunity to debut the next-generation FocalPoint chip architecture.  More to come about this in this blog, but if you are going to Hot Interconnects, plan to go to Uri’s presentation (Aug. 18 @ 3:15 pm).  Here’s the abstract from the show website:

Introducing the Next Gen FocalPoint Architecture

In 2006, Fulcrum Microsystems launched FocalPoint and set a new standard for Ethernet by delivering the first high-scale, high-density, low-latency 10G switch architecture. FocalPoint ushered in the era of data center Ethernet and opened up the possibility of serving HPC, clustered computing and server interconnect with the same fabric. Since then, the market has taken off and the original architecture has evolved to offer higher scalability, routing, and new data center specific features such as those defined in the IEEE 802.1 DCB Task Group. Ethernet has become firmly established as the leading data-center interconnect technology. But the data center is the place to push the envelope for performance, and Ethernet must evolve both in its throughput and also in its scalability. Uri’s presentation will cover the next-generation FocalPoint architecture, called Alta, which delivers industry’s highest density, line-rate high-speed Ethernet (HSE) switch that comes with a full-speed deterministic programmable pipeline. Come see the future of next-generation data center Ethernet.

By: Gary Lee

In a notable data center fabric partnership, Voltaire, Ltd. and BLADE Network Technologies (BNT) announced they will team up, combining Voltaire’s chassis-based core switch with BNT’s top of rack switches in a new bundled data center fat tree networking solution.

The combined offering can deliver more than 3,400 10 Gigabit Ethernet ports with non-blocking, lossless switch capacity of 11.52 Terabits per second.  This development is notable for a couple reasons:

First of all, while it has ventured into Ethernet before, Voltaire is better known as an InfiniBand vendor.  The fact that the company is publicly announcing working with an Ethernet top of rack switch vendor like BNT is a clear indication that the data center is shifting toward all Ethernet fabric topologies.

Secondly, the choice to partner specifically with BNT demonstrates a clear commitment to flattening out the switching architecture in the data center.  Thin is in, and the days of bloated monolithic routers are numbered.  From the press release:

“Unified fabrics and flat network topologies designed for end-to-end performance hold great promise for enterprises seeking to achieve top efficiency for their critical business applications and server virtualization environments,” said Lucinda Borovick, research vice president, Datacenter Networks, IDC. “Datacenter customers benefit with the ability to choose the most effective platform for the environment when providers such as BLADE Network Technologies and Voltaire jointly validate best practice designs.”

Both of these developments point to the trend of Ethernet taking over as the fabric to unify the data center.  Additionally, this type of switching architecture accentuates the need for an ultra low latency switch with the ability to scale for advanced data center applications.  It requires FocalPoint.

By: Gary Lee

As announced today on the Fulcrum website, I will be speaking at this week’s Linley Tech Spring Seminar.  From our press release:

Much as has been written about the advantages of output queued shared memory switch architectures, but few switch devices have been able to take advantage of this due to performance and design constraints.  My presentation will outline the unique set of technologies that has enabled FocalPoint switches to get around these constraints and make a reality out of what had been little more than an abstract academic concept before.

During my presentation, I will be discussing the FocalPoint output queued architecture compared to the combined input/output queued (CIOQ) architectures employed that you see in other switches.  Because of its use of the OQ architecture, FocalPoint stands above the competition in a number of applications and for a number of reasons:

• Low cut-through latency
• Low multicast jitter
• Eliminates queue under-run for TDM traffic
• Allows creation of virtual switch planes for traffic isolation of
  • Control plane traffic
  • Storage traffic such as FCoE
  • TDM traffic
  • Voice, video and data in telecom systems

Translation: higher performance for lower cost.  If you will be at the event, come by session seven (10G Ethernet in the Data Center) and ask a question or two.  For those of you who can’t make it, I’ll put together a more in-depth blog of the conference when I get back.

By: Gary Lee

Recently, we had an article published in CommsDesign, in which I discussed the issues surrounding constructing switching fabric architectures for cloud-ready data centers.

The article contrasts data center fabrics with traditional telecom-style fabrics.  Although telecom-style fabrics can scale to high bandwidth, the complexity and component count necessary to accomplish this leads to high cost—too high for the data center.  Traditional 10GbE fabrics have not been designed to scale in the data center, and like telecom-style fabrics, have high latency that impacts data center performance.

However, with devices like FocalPoint on the market, advanced 10GbE switches are available to network architects.  These feature-rich switches offer the low latency and efficient scalability required for large cloud data centers while meeting the emerging IEEE standards for data center bridging.  They’re not nice to have—they’re necessary.

Read the full article…

Comments welcome: feedback@fulcrummicro.com

By Gary Lee

Network World recently ran a commentary by a network manager at a $3 billion high-tech company talking about choosing the Cisco Nexus switch in 2008 to build a Ethernet-based fabric for his greenfield data center.

“At the time, our options were rather limited. The Nexus 5000 was the only top-of-rack 10GbE switch around.”

But now, he goes on, there are many credible choices.  He points to data from a Clear Choice test done by the magazine comparing four different top of rack datacenter switches and concludes:

“Flash forward to today and my rule proves again; competition is a great thing. There are now several vendors with top-of-rack 10GbE switches, many of which went through a very nice Clear Choice Test on NWW.com. The overall picture: the Nexus 5000 is playing catch-up at this point.”

What follows after his short article is about a dozen or so comments from readers either supporting Cisco or venting about the high cost and complexity of Cisco networks.

One note, though, stated that Cisco is in a class of its own because it is the only switch to support Data Center Bridging (DCB).

This is going to change.  Our FocalPoint switch supports DCB and has been tested for compatibility in an Ethernet Alliance-sponsored test conducted at the Univ. of New Hampshire Interoperability Lab.  In fact, FocalPoint demonstrated DCB interoperability with the Nexus switch (as well as with switches from Dell, Finisar, Intel, NetApp and QLogic).

DCB is incredibly important for the future of integrated data center fabrics, which is why it’s important for the industry to understand the extent of its availability and the progress on interoperability.  More details about the FocalPoint DCB feature set are here: http://tinyurl.com/yaz6q3t.

Competition is a good thing, especially when it comes to a technology as important as DCB.

Comments are welcome: feedback@fulcrummicro.com

Happy New Year from all your friends here at Fulcrum Microsystems!

As 2009 draws to a close, we have taken a minute to reflect on our accomplishments from this calendar year and look forward to the next.  Perhaps most indicative of our success has been The Los Angeles Business Journal story naming the company as the area’s fastest growing private company.  Despite the adverse economic environment, Fulcrum continued to stay true to its mission of innovation and maintained its success.

With a number of design and customer wins set to come to fruition in 2010, along with new product announcements, we look forward to the year ahead.  The sales team has been out on the front lines driving the discussions of solutions that will enable networks to bring the next generation of real-world applications both in and out of the datacenter.

Keep up to date on all things Fulcrum right here at our blog.  Thanks and, again, Happy New Year to all!

I just got back from the SC09 show in Portland, where I saw a number of interesting HPC and data center interconnect-related emerging trends. Here are a few snapshots:

• Fulcrum, Arista Networks and Ixia announced the first ever performance test of a high density (48-port) 10GbE switch, which showed that it had the low latency, throughput and data center bridging features necessary to scale data centers for cloud applications. The FocalPoint silicon inside the Arista 7148SX performed extremely well, demonstrating completely lossless, full line rate throughput. As data centers continue to expand to support cloud-based applications, port densities of 10GbE switches will continue to increase, but must do this while maintaining the very low latency and high throughput proven by this test. Full test results are available on the Arista site. We’ll also have another post with much more detail on the ramifications of this testing.

• One anticipated highlight of all SC shows is the announcement of the Top 500 supercomputer list. This year’s announcement proved dramatic with a new Cray XT5 computer taking the top spot. On the interconnect front, the InfiniBand Trade Association tried to create its own drama by pointing to a rise in the use of the InfiniBand (IB) interconnect. The association said that IB is now the interconnect technology for 23% of the Top 500 and is gaining ground on the leader, 1G Ethernet (GE). Many supercomputer designers used IB because of its latency advantage over the open interconnect standard GE, but IB is almost proprietary, as only two silicon vendors are left (and even they are looking hard at adopting 10GbE). Now, with Fulcrum’s low latency 10GbE switches with built-in data center bridging, supercomputer designers can take advantage of the economies of scale Ethernet brings. With current generation 10GbE switches competitive in latency and scale, it’s only a matter of host adapters catching up in these areas before 10GbE will push IB to the fringes of this list in the coming years.

• Cloud computing is driving many data center and HPC managers to consider whether to rent or buy HPC capabilities to service smaller workgroups or individual high-performance users. This paradigm was on display at SC09, where Cray made a big splash about its CX1 deskside “baby” supercomputer, which comes on wheels so it can be moved from office to office. The company said Dell would be reselling the computer and also that HPCWire gave it a 2009 editor’s choice award. But HPCwire also called HPC cloud technologies a technology to watch, and another industry commentator said that HPC and cloud computing are “twins separated at birth.” Gains in networking technology will continue to make cloud-based HPC a performance competitor with deskside computing and make the rent vs. buy decision even more difficult.

• Another interesting story that points to the broader industry trend of demand for low-latency switching is the emergence of blades featuring graphics processing units (GPUs). For applications like data mining, which require multiple traffic streams, the low latency of FocalPoint silicon can help isolate GPU traffic from other traffic, while still enabling data center architects to benefit from the power and cost savings offered by using GPUs.

And finally, congratulations go out to Fulcrum partner Arista Networks, which was named “Best Interconnect Product/Technology” in HPCwire’s annual Reader’s Choice Awards! Read the full article that includes all award winners in all categories on HPCwire.

Comments welcome: feedback@fulcrummicro.com.