Fulcrum Microsystems Blog

By: Gary Lee

Last week, I traveled to San Jose to speak at the Ethernet Technology Summit, where FCoE was one of the focus areas at this year’s show.  I was part of an interesting panel discussion on FCoE, with panel members from key stake holding companies like Cisco, HP, Brocade, QLogic, Emulex, NetApp and Amphenol.  Several themes emerged from this discussion.

It was clear that the panel felt that, with its current extensive installed base, Fibre Channel will have a long half-life, but that FCoE will take a large market share in greenfield installations.

Another point was that data center bridging (DCB) standards are becoming stable with several early implementations on the market today. The server side of the network is the current FCoE focus with converged network adapters (CNAs) available from several vendors.  NetApp pointed out that it has an FCoE storage target available today, and it expects several other vendors to follow suit shortly. This would eliminate the need for any FC switches in the data center fabric.

My paper was entitled “Fabric Requirements for FCoE in the Data Center.”  It described how FocalPoint can provide all the key DCB features today along with FCoE frame processing (pdf download), allowing our customers to develop FCoE-enabled solutions. I pointed out that there are only a few trusted FC system vendors in the industry today, which is not good for the market, but this can change with FCoE. Throughout the conference, I sensed some resistance to embrace FCoE, but I posed the question: “If FCoE becomes mature in five years, and you need to build a new cost-effective data center, would you have any FC switches in your network?”

Comments welcome: feedback@fulcrummicro.com

By Gary Lee

I was in San Jose last week to speak at the Ethernet Technology Summit. My topic was FCoE, but another focus area of the show was the status of 40/100 GbE.  In one of the keynote addresses, John D’Ambrosia of Force10 Networks discussed “Current State and Future Plans for 40/100 GbE.” As one of the drivers of the IEEE standards effort, he described how 40/100 GbE is stable today and the standard should be finalized by this summer. Standards are one thing, but what about real products?

We see 10 GbE as the bandwidth leader today and we feel it will be 1-2 years before 40 GbE starts to take off in the market. 100 GbE will have further hurdles to overcome. The industry is showing opposition to using 10 fibers in each direction to achieve this performance level. A much better configuration is 4 x 25G, but it will be a few years before 25G serdes reach the correct price point. Because of this, it will probably be 2-5 years before we see cost effective 100 GbE links in the market.

In a recent Network World article, end customers like Facebook are complaining that 100 GbE is not enough and they need 1,000 GbE right now.  This will obviously take a while, and I don’t expect to see cost effective products at this bandwidth until around 2020.  In the mean time, people will have to be satisfied with high bandwidth LAG groups, or new technologies like TRILL, which take advantage of normally unused spanning tree data center network bandwidth. With our FocalPoint Ethernet switch products, we will stay on the forefront of these technologies, implementing them as they become cost-effective industry standards.

Comments welcome: feedback@fulcrummicro.com

By Gary Lee

I’m at the Ethernet Technology Summit today to present on the fabric needs of FCOE data center switches (more on that in another post), and in looking over the keynotes for the show, I was struck by the number of discussions centered on video traffic.

It is clear that we are entering a world in which video distribution is becoming a key network traffic component. A 2009 report issued by Cisco stated that all forms of video are expected to exceed 90% of all consumer Internet traffic by 2013. To support this trend, though, switching infrastructure needs to efficiently support multicast traffic.

We’ve been anticipating this demand and considering what features need to be built into the switch to support this traffic.  The FocalPoint family of 10GbE switch silicon has been architected with video distribution and multicast traffic in mind.  FocalPoint’s output queued shared memory architecture queues multicast traffic only once, reducing internal switch congestion and memory requirements.

In addition, video traffic can be assigned a unique traffic class along with a dedicated memory partition in the switch, allowing it to be flow controlled separately from other data traffic using IEEE Priority Flow Control frames. This minimizes both video latency and traffic loss. At the switch egress, video traffic can be giving minimum bandwidth guarantees, which translates to bounded latency jitter.

Latency jitter is a key parameter in video distribution networks as it impacts the size of memory buffers required at the location where video is being reproduced, which in turn drives the cost of video distribution systems or set-top boxes. In addition to bounded latency jitter, FocalPoint can also operate in cut-through mode with absolute latencies around 200nS and P-P latency jitter of less than 50nS. All of this leads to cost savings for both the service provider and consumer.

Comments Welcome: feedback@fulcrummicro.com

Recently, I spoke at the Linley Group’s Carrier Ethernet Design Seminar, discussing the migration to packet-based fabrics in the context of supporting legacy time domain multiplexed (TDM) traffic in mixed service system architectures.

In all, the show provided an interesting look into the migration of carrier networks and the role Ethernet fabrics will play in this evolution. The presentation was timely because acquisitions have shaken up the market for cell-based backplane fabrics, leaving companies in the ecosystem wondering what the long-term health of cell-based fabrics might be.

During my presentation, I pointed out that silicon solutions for TDM are too expensive at 40G and above.  Additionally, with Ethernet at 40G and moving toward 100G, a number of vendors are adopting it because of its cost advantage.  I also laid out a three-phase evolution that we’re seeing in the market; from completely TDM based networks, to hybrid TDM/packet networks and finally to completely packet-based networks.

Phase one is the expensive, legacy architecture with TDM-only fabric necessitating the SARing of all packet traffic.  Phase two adds a second packet fabric for native IP traffic, but also adds system cost.  In phase three, we move to an Ethernet backplane that features advanced congestion management providing bounded latency for efficient TDM transport.

This low cost Carrier Ethernet-based solution is only possible with the proper features built into the Ethernet fabric.  By providing TDM traffic with separate memory partitions and watermarks, it can be effectively isolated from packet traffic. In addition, advanced egress schedulers using minimum bandwidth guarantees can provide bounded latency, which is required for the proper reassembly of TDM flows.

Packet-based traffic will inevitably dominate the carrier networks.  My presentation outlined one way in which cost effective Ethernet backplane fabrics can evolve to meet the needs of these networks while also providing transport for legacy TDM traffic, but we would be interested to hear your thoughts on how this evolution will take place.

Comments welcome: feedback@fulcrummicro.com

You might have seen the news this week that Fortinet® has chosen FocalPoint silicon for its recently announced FortiSwitch-1000 chassis-based switching platform.  Fortinet is a leading network security provider and worldwide leader of unified threat management (UTM) solutions.  As Fortinet Senior Director of Data Center Strategy Marshall Bartoszek observed:

“One of the keys to this [new switch family] is the FocalPoint architecture, which provides high performance at the core of our architecture and powers the vScale congestion-avoidance features that give our switches a distinct, low latency competitive advantage.”

The FortiSwitch-1000 is another example of a FocalPoint-enabled fat tree network architecture providing non-blocking Ethernet fabrics for even the largest data centers.  Data Center Bridging features built into FocalPoint provide the FortiSwitch family a silicon solution that enables the high-performance network fabrics required by data centers for latency sensitive tasks and high-speed interconnect applications, including server virtualization, data center consolidation and cloud-based computing.

Read the full release here…

Comments welcome: feedback@fulcrummicro.com

Congratulations to Arista Networks and Blade Network Technologies, whose switches performed best under a recent Network World Clear Choice data center top-of-rack Ethernet switch test!  The magazine summarized the value of TOR switches like this:

“As data center managers consolidate and virtualize their servers, the next order of business becomes moving all that traffic. Enter top-of-rack data center switches that offer speed, scalability, redundancy, virtualization support and other features not available in garden-variety Ethernet switches.”

This intensive evaluation that included six switches that feature at least 24 10-Gigabit ports, resulted in Arista’s DCS-7124S and Blade’s G8124 switches being deemed the top choices for the data center. The testing was summarized as follows:

“With the best combination of features and performance – especially in the areas of latency and jitter – both switches earn Clear Choice awards”

The detailed test results can be found here.  The general consensus was that the Arista and BNT switches offered the best overall performance, with list prices that are about one-third that of the competing solutions.  This price advantage can be attributed to the switches’ use of merchant silicon like FocalPoint, which has the functionality and low latency required in the data center.

Comments welcome: feedback@fulcrummicro.com

As we look forward to the coming year, the market for 10GE switches is expected to heat up—driven primarily by increased competition in the data center.  This article in Network World notes an increase in large vendors like Avaya, IBM and HP that are challenging the current Cisco-centric paradigm, and they’re doing so by developing cost-effective data center bridging (DCB) switch fabric solutions.

To compete with Cisco on cost, these vendors will not be able to justify the large internal chip development model made famous by Cisco, and the result will be a heavier reliance on merchant silicon to keep costs down.  However, for these fabric-based architectures to have the necessary performance, it is vitally important that the switches have extremely low-latency and high bandwidth along with advanced DCB features.

FocalPoint provides the combination of features necessary for large DCB fabrics designed for scaled out data centers that require cost-effective Ethernet switching solutions.  With Fulcrum at the center of the switching architecture, vendors now have the merchant silicon solution they need to break through Cisco’s stranglehold on the data center market and legitimately shift the course of the data center evolution.

Comments 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!

The IEEE is defining an important new specification for the cloud data center that defines how to connect multiple virtual machines to a data center bridging (DCB) fabric through a single high bandwidth port. An earlier proposal by the PCI-SIG for single root complex IO virtualization (SR-IOV) has not gained much traction in the industry. Although Cisco has released the Nexus 1000 as a virtualized bridge within the blade server rack, it is in effect a proprietary solution that forces the use of Cisco equipment throughout the data center if unified management software and network policies are desired. This can be an expensive and limiting solution.

A new proposal called Virtual Ethernet Port Aggregator (VEPA) is being developed by the IEEE, as described in this article from Network World. This standard moves the virtualization function from the blade servers into the DCB fabric. Many companies are attracted to this proposal because the management and network policies can be easily applied to both the physical and virtual fabrics using a single software stack running on an Ethernet bridge. With the Cisco proposal, part of the fabric is inside the blade server system, which of course, Cisco wants to supply as well. It looks like Cisco has now acquiesced, and will join the VEPA standards effort.

FocalPoint switches have several features to support the VEPA standard. A key feature is the use of Global Resource Tags (GloRTs), which can be used to identify the locations of thousands of virtual machines connected to the FocalPoint fabric. Translation between VEPA tags and GloRTs at the ingress and egress of the fabric using a Q-in-Q type approach allows easy address translation across blade server domains while maintaining support for all other DCB traffic such as FCoE. This highly scalable solution can be achieved while maintaining the low cut-through latency pioneered by Fulcrum.

For more details on DCB and CEE support in FocalPoint, download this PDF whitepaper.

Comments are welcome: feedback@fulcrummicro.com

At the recent SC09 show, the Ethernet Alliance brought together equipment from 14 vendors to demonstrate Data Center Bridging (DCB) using a number of different storage protocols that leverage cost-efficient, ubiquitous Ethernet as a converged fabric.

The interconnect was a major element of this unified fabric showcase, as it incorporated different technologies such as 10GBase-T and SFP+ 10GbE Direct Attach Cables into the demonstration. It also demonstrated how network convergence takes advantage of high speed Ethernet to deliver client messaging, storage and server application communications over a single network.

Fulcrum’s Monaco reference design, which features the FocalPoint FM 4224 24-port 10GbE switch chip, was the 10GbE switch used for all of the tests. The switch was one of the few available that supports the data center bridging (DCB) features required for converged fabrics including support for the Fibre Channel over Ethernet (FCoE) standard.

For more information on the demo, visit the Ethernet Alliance Website and download the SC09 Whitepaper.

Comments welcome: feedback@fulcrummicro.com