It’s no secret that I am a neurotic about having numbers in my blogs. I love data and often tell my team that we must have claims and quantitative data points for all of our marketing materials. More than once, I have heard moans and groans from folks when I tell them that we can’t go forward without some stats that support our claims. I also have some OCD tendencies to want to group things into 3s, 5s or 10s. This is also why I love fantasy football. It is fun to sift through data and play with statistics.

My OCD issues aside, numbers are important. Data is important. People trust data. And more than trust, data is meaningful. So when we decided to survey global IT professionals on trends and drivers impacting their data centers, we were hoping to provide meaningful data that would help guide our customers as they address the bandwidth strains virtualization, cloud, big data and convergence are putting on their data centers today, and tomorrow.

It goes without saying that we were pleased with the results of this survey, because not only did it reveal some interesting statistics, it reaffirmed and validated others…such as the fact that the majority of organizations are converging their LAN and SAN fabrics, something we’ve stood behind since the Fibre Channel over Ethernet (FCoE) standard was ratified three years ago. What was most amazing to me was the rate at which the demand for bandwidth is increasing. Not only are almost half of respondents deploying 10Gb Ethernet (10GbE), the majority of the survey respondents foresee needing to jump to 100GbE in just four more years. That’s just unprecedented! These data points perfectly aligned with what we had hoped for; proving that I/O is, in fact, strategic. That is the type of data that excites me. Incredible jumps in data and storage needs and adoption cycles that just keep getting faster (16Gb Fibre Channel [16GFC] is another great example of that – we are seeing adoption of 16GFC growing 25 percent faster than the transition to 8GFC).

Beyond gathering some data that is incredibly useful, this survey was pretty widely received in the IT world (check out eWeek, CRN, InformationWeek, ChannelBiz and Computerweekly, to name a few) fostering discussion, and hopefully, smart, well-educated decisions by enterprise organizations as they deploy new technologies now and in the future. But I must say, my favorite highlight of this survey in the media had to be seeing our infographic (talk about numbers – infographics have become my pet project!) splayed across the jumbotron in Times Square. If that doesn’t say data is important (and strategic!), I don’t know what does.

Posted in Uncategorized | No Comments

Why I/O is Strategic is a new education campaign that Emulex has kicked off (see our press release here) with leading industry analysts, including Enterprise Strategy Group (ESG), IDC, Evaluator Group and 451 Research, to demonstrate the strategic importance of I/O in the data center for four key trends: virtualization, cloud computing, big data and convergence. We wanted to do something different with this series than the typical vendor model of saying…here is a problem; here are a few options to solving the problem and look ours is the best… As marketing people, we obviously get paid to say our stuff is good, but we wanted to take a new approach to this educational effort by creating a virtual panel of experts to bring a wider perspective to IT managers. The new education series, which is based on a survey of 1500 IT managers, a series of briefs by industry analysts, followed up with a series of associated webcasts, helps IT professionals build better strategic I/O plans for these emerging trends in the market. Here are some of the questions we will tackle in the briefs and the webcasts:

• Virtualization
  • How will you support up to 100 virtual machines (VMs) on a single physical server?
  • What are the right virtual Network Interface Card (NIC) and switch plans?
  • How will the purchase of software-defined networking (SDN)/Nicera change the deployment of networks on virtual servers?
  • How do you manage I/O for 1000 virtual desktop infrastructure (VDI) clients?
• Cloud
  • What is the right I/O strategy for deploying your private cloud?
  • What I/O model should your public cloud provider have going forward?
  • How will Network Virtualization using Generic Routing Encapsulation (NVGRE) and Virtual Extensible LAN (VXLAN) change the mobility and scalability of applications?
• Big Data
  • Can I/O make Hadoop run faster?
  • What is the best I/O for all that east-west traffic?
  • How do I use NoSQL in these environments?
• Convergence
  • Can I run HPC workloads on this to skip going to InfiniBand?
  • What is RDMA and why do I care?
  • Fibre Channel (FC) or Fibre Channel over Ethernet (FCoE) what should I use for my next storage network?
  • What are the best practices for moving to 10Gb network convergence?
  • When is 40Gb Ethernet (40GbE) hitting the market?

One of the best parts of our business is that it never stops moving forward and innovation is constant. Sometimes it comes from leaps in technology, leaps in commoditization, leaps in business models, leaps in processes, but whatever the leap is based on, it requires the vendors, analysts and IT professionals to never stop learning.

Our plan is to tackle 3-4 new trends each quarter over the next year to keep the dialog going. I hope you can make the time to read the anthology briefs created by our virtual panel of industry analysts and attend the webcasts with each industry analyst and industry partners. I can’t tell you everything that will happen, but I can tell you that knowing “Why I/O is Strategic?” actually is strategic to you, your business and your career. The new information in this education series will help you be better at your job and smarter than the guy in the next cube and that will make you more strategic to your company.

Posted in Uncategorized | No Comments

Like many technology companies, Emulex has an end-of-the-year shut down. And while this shut down is for most of us, it’s not for all of us. While many of us were focused on the holidays, our sales, support and operations teams have been working on closing the quarter. It seems that every year during the end-of-year tradeshow season (September-November), we announce new technology and say it will ship by the end of the year in calendar Q4. This year was no exception and the new products in question were the new LightPulse® 16Gb Fibre Channel (16GFC) Host Bus Adapters (HBAs).

Leading up to the end of the year, our engineering and ops teams were busy making our commitments come true again. First, I would like to thank them for the extra efforts. Second, I hope you get a chance to check out the new LPe16000 and LPe16002 cards here. These cards are more than just a bump in protocol speeds. They offer a great new tool for optimizing clouds and server virtualization in the data center. The Fibre Channel team is getting ready to launch a new series of Labs blogs on the latest LightPulse HBAs and help you navigate the Romley server transitions, and optimized deployment and management of your FC environments.

It is a new year. We have new products and they are shipping. So, what more do you need? Well, you probably want to know how they work. We have the folks at ESG Labs working on a review and report, but we also want to know what you think. Tell us why you need an upgrade to 16GFC, and we will offer 5 lucky IT folks a pair of 16Gb cards in exchange for your review of the cards. Please note, we need to be able to publish your name, shoot a short video case study and have your company agree to publish the case study. (Sorry, we have attorneys too.) We hope you had a great new year and will help us kick it off with new stories of how 16GFC can help in 2012.

Posted in Uncategorized | Comments Off

In the 10Gb Ethernet (10GbE) market, we will see market expansion at over 50 percent annually during the next four years, according to Crehan Research (see chart below). As we have said many times, the growth of 10GbE will come in many waves. We are about to experience a new phase of growth in the 10GbE market with the addition of the rack server business to our current leadership in the blade server market. According to the last published report by Crehan Research (Q2 2011 Server-class Adapter and LOM market share report), the total 10GbE market was $104 million for calendar year (CY) Q2 in 2011, or about $408 million on an annual basis. This business has been primarily driven by LAN on Motherboards (LOMs) and mezzanine cards on blade servers and Fibre Channel over Ethernet (FCoE) Converged Network Adapters (CNAs) and 10GbE Network Interface Cards.

¹Crehan Research, July 2011 Server-class Adapter & LOM Long-range Forecast Tables

As you can see in the chart above, the green line represents blade server connect rates of 10GbE. This market took off at the Nehalem transition as we discussed two years ago. Given that HP, who controls over half of the blade server market (IDC Worldwide Quarterly Server Tracker 2011 Q2 Report), has chosen to go chip-down with 10GbE LOMs, IBM is offering 10GbE blade configurations for its Virtual Fabric Architecture (VFA) architecture and Cisco UCS blade servers include 10GbE, over 50 percent of blade servers are being shipped with 10GbE capabilities today. By the end of the year, the overall blade market is expected to grow to a 67 percent 10GbE connect rate in CY 2011 and that it will continue to grow, reaching nearly 100%, in the next four years¹.

The next major transition will be Romley-based servers, where Original Equipment Manufacturers (OEMs) are adding modular LOMs and chip-down LOMs that include 10GbE. The orange line is the connect rate for 10GbE on rack servers, which, this year, will reach about 15 percent, according to Crehan Research. The grey line, in the chart above, shows the jump in the 10GbE connect rate at the Romley transition. This is when OEMs will offer almost every server with 10GbE modular or chip-down LOMs. This will drive the overall market for 10GbE from around $400 million today to more than $1 billion by 20141. As we enter this next wave of 10GbE growth, we will see new opportunities in the 10GbE Adapter and LOM market for rack servers in CY 2012 and beyond.

Posted in Uncategorized | No Comments

We recently did a webcast with Dell about the 10GbE iSCSI over Data Center Bridging (DCB)(Solve your data traffic queue problems with Dell EqualLogic & Emulex iSCSI over DCB Solutions) a.k.a Converged Enhanced Ethernet or Enhanced Ethernet or Data Center Ethernet, take your pick. Joining us from Dell was Sarah Cook and Gary Gumanow from Dell and Sean Murphy (yes, he spells ‘Sean’ properly, unlike myself.) from Emulex. We had about 300 folks on this webcast to talk about iSCSI. So what did they hear? Gary took us through the end-to-end solution and the technical details for what is new about 10GbE over DCB and Sean Murphy, our resident VMware product marketing manager, took us through the management and deployment tools for VMware environments.

What is iSCSI over DCB?

DCB or DCE (Cisco’s version of DCB; Data Center Ethernet) or CEE (Brocade’s version of DCB; Converged Enhanced Ethernet) improves the Ethernet fabric irrespective of what protocol (be it iSCSI, NFS, TCP or FCoE) runs over the top of it. It does this by adding lossless characteristics similar to Fibre Channel (FC). This means that iSCSI over DCB can now provide the same enterprise class services as FC at a lower cost and help drive network consolidation via convergence. To do this, DCB added four key technical capabilities to standard Ethernet to bring the best of FC and iSCSI together into one solution.

• IEEE 802.1Qbb, Priority Flow Control, is a PAUSE-based flow control mechanism that extends the legacy Ethernet PAUSE flow control, so it is not full-link based, but each one instead operates specifically on one of eight priority levels within the link. Therefore, a PAUSE on one priority flow does not stop the entire link.
• IEEE 802.1Qaz Enhanced Transmission Selection, defines a technique to allocate bandwidth to entities called “Priority Groups” which are collections of priorities (0-7) within a group. This specification also defines a link-based protocol (DCBx) that permits link local parameter negotiation.
• IEEE 802.1Qau Congestion Notification, provides a message-based, end-to-end congestion notification mechanism that can squelch senders, for example, that overrun the received capabilities of a target. Note that 802.1Qbb is link flow control, whereas 802.1Qau is end-to-end flow control that may travel over several links in a path.
• IETF TRILL, (Transparent Interconnection of Lots of Links), is not a part of the IEEE DCB specifications, but is often discussed in the same context because it provides multi-pathing capability for Ethernet Layer Two fabrics, which is a facility that is provided in FC fabrics.

During the webcast, Gary really takes you into the details of why this is valuable to your enterprise.

Management and Optimization for VMware:

Sean took us through the details of configuring and setting up ISCSI over DCB in VMware environments. Emulex is long-term partner with VMware and have developed industry-leading tools for VMware and vCenter to make deployment fast and simple. OneCommand Manager for VMware for vCenter (OCM for VMware) is a native software plug-in that integrates real-time lifecycle management of Emulex LightPulse® Host Bus Adapters (HBAs) and OneConnect™ Universal Converged Network Adapters (UCNAs) into the VMware vCenter console. This tight integration centralizes and simplifies virtualization management. OCM for VMware builds on Emulex Common Information Model (CIM) providers and established OneCommand Manager features to proactively address key data center issues and improve operational efficiency across VMware hosts and clusters. The core functionality, delivered with OCM for VMware, includes multi-protocol management (FC, FCoE, iSCSI, NIC), online firmware flashing, configuration updates, reporting options and adapter diagnostics, and flexible graphical and command line interfaces.

Sean also touched on our new Universal Multi-Channel (UMC) capability that enables our cards to present as NIC functions (IP, iSCSI, FCoE). They are presented to the operating system, or hypervisor, as a physical port with a separate MAC address and assigned bandwidth. UMC is enabled and managed at boot time.

Most servers are currently deployed with multiple 1GbE physical connections. Typically, these additional ports are used to support virtual servers and high availability, and to provide bandwidth needed for I/O-intensive applications. UMC provides a similar capability for 10GbE networking by using individually configurable partitions of the 10GbE port. With UMC, data centers can save on costs for cabling, adapters, switches and power.

Business Drivers for 10GbE

Since I was the non-techie on this call, I outlined the key market drivers and covered the top ten business drivers for 10GbE iSCSI.

If you have an extra hour, take a listen and find out about the latest in iSCSI.

Posted in Uncategorized | No Comments

Summer is over and it is time to get back to work.  As some of you may have noted, I am a bit behind schedule on my blogs.  I did not intend to “take the summer” off, but it kind of worked out that way this year on the blog front. As we approach Labor Day, the traditional start of fall, football (the American version), back to school…and for the geeks of the world it is “Trade show Season.”  Yes, that time of year when we announce our new products, run around the world giving talks and telling IT managers what new tools we have created for them to use.

This past week in “Fabulous Las Vegas,” as the sign says, we kicked of the trade show season with VMworld.  It was a great week and some very cool things were announced and demoed at this year’s show.   The two biggest things seem to be Solid State Disk (SSD) and overlay networking via Virtual Extensible Local Area Network (VXLAN).   VXLANs will help simplify networks and help integrate data centers, public and private clouds with on-demand overlay networking. The analogy used by VMware is with VXLAN you can now  take VXLANs layer 2 packets and encapsulate them in a layer 3 protocol to span local and remote networks in a single virtual network. You can also do the following:

• Use MAC-in-UDP encapsulation to build the overlay network that can span across L3 networks.
• Use of MAC-in-UDP allows efficient load-sharing with the existing data center networks due to the use of Equal Cost Multi-pathing (ECMP) in the core networks
• Allows overlapping of VMs MAC and IP addresses using VXLAN  Network identifier
• Takes advantage of efficient multicast protocols such as IGMP and PIM for VM’s broadcast and multicast communication needs

More details in this Citrix blog and this VMware blog.

Emulex is working with VMware and a number of industry partners to support overlay networking models.  We see this as a viable option to make virtual networking simple and scalable for the cloud.  We have over 100 events across the globe between now and December.  Trade show season is upon us, and I look forward to seeing you at on the road.

Posted in Uncategorized | No Comments

Each quarter after our earnings announcement, as with most public companies, we attend a number of investor conferences. In doing these events, we update Wall Street on many items, and one of the hottest topics for Emulex is the development of the 10Gb Ethernet (10GbE) market. Last week, we also attended Interop in Las Vegas, where we counted many new 10GbE switching platforms and a number of key technology announcements that help continue the maturation of the 10GbE market. The chart below is one some of you have seen before, but we have modified it, and I want to update you on how we see the 10GbE market maturing.

• Wave 1 – Wave 1 has been dominated by blade deployments. We saw virtual fabrics with the ability to create virtual Network Interface Cards (NICs) drive this market. 10GbE was able to be deployed at lower costs than 1GbE due to the savings related to the blade infrastructure (optics, cables and external switches), and we have seen this market drive Emulex to the number one position in 10GbE ports according to Crehan Research (Q1 2011 Server-class Adapter and LOM market share report). However, server blades represent about 15% 13.7% of the current shipping servers shipped in 2010, growing to 22% by 2015, according to IDC (IDC Worldwide and Regional Server 2011–2015 Forecast, Doc #228060, May 2011).

• Wave 2 – Wave 2, which will come later this year, based on the Romley server transition in the x86 and Unix markets, represents up to 65% of the shipping servers and offers more than 2X the 10GbE opportunity than Wave 1. According to Intel, we should expect to see Romley-based servers announced late in CQ4 of 2011. This will drive the second major wave of 10GbE adoption. In this wave, you will see greater use of modular LAN on Motherboard (LOM) technology, growth of 10GbE storage on NAS and iSCSI over Data Center Bridging (DCB) and the growth of more cost-effective 10GBASE-T switching solutions. Additionally, this quarter, we saw multi-hop FCoE switching solutions announced by Cisco, Brocade and Juniper that each allow any port on their switches to be either Fibre Channel, IP or converged Ethernet (just as we saw in blade switches that drove Wave 1); this universal I/O port technology will be a key driver of Fibre Channel over Ethernet (FCoE) in Wave 3.

• Wave 3 – In Wave 3, we expect to see FCoE lead as the ecosystem has matured, and IT managers have had time to deploy pilots during Wave 2, are ready to transition converged architectures and have managed out older equipment that was still on the amortization schedules and replaced them with multi-hop/multi-technology switching ports. The next major driver of 10GbE in Wave 3 will be the deployment of 10GbE by the Web giants (Google, Yahoo, MSN, Amazon, Netflix, Baidu, SalesForce.com, Facebook) as they require increased bandwidth support for the Web, video and mobile applications. According to our estimates, the servers used by the Web giants in shipping containers and other alternative deployments could account for up to 15% of all servers in the world by 2014.

From the beginning of this market, we have said that the path to 10GbE would happen in multiple waves and take a number of years (three to seven). At the same time we are seeing 16Gb enter the Fibre Channel market, 40 and 100GbE are entering the very high end of the market and into blade mid-plane form factors. The never-ending migration of speed and performance into our infrastructure confines continues unabated. Emulex will continue to lead this market forward in each of these transitions as they take similar paths into the data center and move it into the market at the right time based on meeting business needs, shipping at the right costs and supporting the right applications.

Check out my discussion with Stu Miniman at EMC World earlier this month on 10GbE adoption:

Posted in Uncategorized | Comments Off

As a racing fan, “Drivers, start your engines” is one of my favorite lines; it gets the blood pumping, palms starting to sweat in anticipation, and you are ready to drop the hammer and go. That is kind of how I feel on days when we launch new products and set new strategic directions. Your prep is over, the cars are set and configured for the race and you hope that you have planned for and accounted for every possibility, but you won’t really know until the race is run. This week is a landmark week for Emulex: we are announcing our newest three-year vision and announcing a key new I/O engine for our race with the competition…and yes, we think we are going to win!

The I/O Race Courses

Not every race is run on the same track or has the same type of competition, and for this reason, racing teams often field multiple teams based on the race course they are running. It is the job of the team owners to know the tracks and guide the pit crew to prep the right car to win. This is what our Connect I/O Roadmap does for Emulex. We first presented this Connect I/O Roadmap at our investors’ analyst day in November 2010. We used it to outline our forward-looking technology direction, and it has served as the basis for our updated three-year vision: the Emulex Connect Architecture (ECA).

Let’s look at the type of races we will be running over the next three years:

• Ethernet and Universal LAN on Motherboard (ULOM) – The Ethernet market is in the early stages of transitioning from 1Gb/s Ethernet (1GbE) to 10Gb/s Ethernet (10GbE). As we all know, 10GbE has been around for almost 10 years, but as with all technology transitions, it is the confluence of the right cost, the right applications, the right technology and the right business need that drives these moves in the market. The advent of server virtualization is driving the business needs and I/O demand, 10GbE in server blade chassis gave 10GbE a competitive cost point by reducing the component counts and network convergence-based, fabric-based computing solutions gave us the right technology. This drove the first mass deployment of 10GbE Network Interface Cards (NICs)/Converged Network Adapters (CNAs) and ULOMs. As we move forward with 10GBASE-T, modular LOMs in rack servers and the second-generation (Sandy Bridge) of 10GbE-ready servers that will be shipping later this year, we will see 10GbE grow. Next, we will see the same type of transition happen from 10GbE to 40GbE in future generations of blades and then move to racks again.
• Network Convergence – Network convergence began at 10GbE, and we will see 40GbE and 100GbE run in parallel with the mass adoption of 10GBASE-T solutions launching later this year. Virtual NIC technology has led the first wave of market adoption, NAS and iSCSI 2-4 Socket Rack servers will drive the next wave, followed by volume 1-2 socket servers and then Fibre Channel over Ethernet (FCoE) for storage. These waves will cover three to four years of transition. Emulex is leading this market and will be demonstrating the industry’s first 10G BASE-T Universal Converged Network Adapter (UCNA) and 40GbE UCNA on May 9 at Interop in Las Vegas.
• Fibre Channel – As Mark Twain said, “Rumors of my death are greatly exaggerated.” As a matter of fact, Fibre Channel is alive and well, and we will be doing the industry’s first 16Gb/s Fibre Channel demo at EMC World the same week (May 9) as we set the standard for next-generation Fibre Channel deployments. The transition from 4Gb/s Fibre Channel to 8Gb/s Fibre Channel took three years (to reach 50% of shipments) vs. the two years previous transitions had taken. However, we have to note that this transition was during one of the most challenging IT spending environments in our businesses history. However, given the price bump for optics, the timeline for 16Gb/s switches and 16Gb/s arrays, we expect a similar ramp-up starting in early 2012 after the major server and storage OEMs release 16Gb/s solutions. This places the start of the 32Gb/s Fibre Channel market in early 2015.
• Converged Fabric Adapters – As we have been working with IT managers and OEMs on next-generation solutions, it became clear that 10GbE and Fibre Channel would live in parallel for many years. This led Emulex to create new class of Converged Fabric Adapters (CFAs) that are capable of supporting tradition 8/16Gb/s Fibre Channel and 10/40GbE network convergence on a single platform. The XE201 ASIC that was announced today is the first solution to provide four ports of 8/16Gb/s Fibre Channel and 10/40GbE on one chip. This will enable OEMs to build converged fabric solutions that simplify I/O, drive advanced virtual networking standards (VEB, VEPA, SR-IOV, MR-IOV), lower power needs (fewer server cards, optic and cables and our improve ROI with GreenState™ Technology) and scale I/O performance with RoCE and over 1 million IOPS on a single card. This new I/O category enables Emulex to help IT managers protect current Fibre Channel investments and provides a direct roadmap to network converge at their pace.
  For more details on the ECA and the Emulex I/O Roadmap, check out the ECA Overview White Paper found on the Strategic Direction page.

The Pit Crew – Engineering Optimized for Each Race

The race you run is only as good as the pit crew you have–their ability to adjust during the race to the changes in conditions and the speed of the competition. Emulex’s engineering department is our pit crew, and they never let us down. They have built and race-prepped the industry’s most powerful, cost-effective, fuel/power efficient and flexible I/O engines, and we are dropping those into a multitude of I/O chassis to win each race outlined above.

• Network Convergence I/O Engine – The BladeEngine (BE)-class ASIC family, based on the technology we acquired from ServerEngines, is focused on the LOM, modular LOM, NIC and CNA market races. It has the right balance of Ethernet-centric technologies, costs and fuel/power efficiency to win this race against the bigger players in the Ethernet space. This is why Emulex has over 60 design wins and is leading the market in 10GbE port shipments. We went from nowhere to first place in just three years. Raw speed is not the only requirement in winning I/O races. Emulex must provide enterprise-class reliability and be accomplished across many protocols (IP, iSCSI, RoCE and FCoE), support many Ethernet models (10GbE SFP, 10G BASE-T, 40Gb, 100Gb), maximizing CPU efficiency and virtual machines (VMs) with offload technology, and provide enterprise scalability of I/O with SR-IOV to win race in this space.
• Fibre Channel and Converged Fabric I/O Engines – When we look at Fibre Channel and converged fabric, our Emulex Engines (XE) class of ASICs deliver the raw horsepower to run the biggest races at the fastest speeds with enterprise-class reliability. Storage administrators expect near perfection in the enterprise-class Fibre Channel solutions, and the XE-based solution delivers this reliability. The XE201, announced today, is our ninth-generation of Fibre Channel technology. To this proven 8/16Gb/s Fibre Channel technology we have added full 10/40GbE network convergence on a single chip to create the new CFA category of converged fabric I/O. However, this is more than just a roadmap, it is reality. Emulex already has strong OEM endorsement for the new XE201 ASIC-based technology. We were the first to announce an OEM 16Gb/s Fibre Channel win with IBM PSeries last year. Additionally, we have new wins for 16Gb/s Fibre Channel, 16Gb/s and 10GbE Converged Fabric wins for both host and target applications with tier 1 OEMs. We do not think other competitors are able to match those claims and proof-point in the new market segment.

To learn more about the XE201 and how products based on this best-in-class I/O engine are going to drive next-generation I/O, or to see videos of the first products based on this ASIC, go to the Emulex XE201 I/O Engine landing page.

We Are Entering Turn One…

So we have all revved our engines (at least, Emulex has) and are entering the first turn of the CFA and 16Gb/s race, where the leaders are established and will drive the industry forward. We have announced the ECA and the XE201, we will be demonstrating these new technologies the week of May 9 at EMC World (16Gb/s Fibre Channel) and Interop (10G BASE-T and 40GbE UCNAs), and we will be working with OEMs and ecosystems partners to bring these solutions to market. The FCIA has planned a major 16Gb/s Fibre Channel plugfest for October at UNH, and our engineering pit crew will help us fly through this stop in record time. After this first pit stop, the next phase of the race will begin as OEMs should begin supporting 16Gb/s Fibre Channel and multi-fabric CFAs sometime after this event. We know as this race enters turn two, Emulex will have to keep pouring on the speed to separate ourselves from the competitive pack and continue to pull ahead.

Posted in Uncategorized | No Comments

As I watched The King’s Speech, I was struck by the parallels of how the people of the 1930s-40s had to learn how to use this new “wireless” media and how today, some people are still struggling to understand social media. They are just like Bertie, knowing what to say and they are articulate in their minds, but the new media model is forcing them to change, grow and take risks in ways they never had to before.

It is not earth-shattering or a newsflash to say that Emulex uses social media as both a business tool and a way to build a community of partners and customers. We have been actively working on this for years now. One of the questions we have been asking is, why do some people at Emulex love social media to the point of being addicted, some see it as a means to an end and some see it as a crock–they just don’t get it? Most of those who don’t “get it” think that those of us who love it are off-base in some way. They may be right, but we will leave that for another blog.

Last week, we had one of those dreaded corporate events known as the “all-hands meeting” for the corporate marketing team. We have all been to them, where we shuffle in, make a few jokes about the suits, even though I am one of those suits now, and hope the pain ends soon. The topic was how to review our social media plan for the next fiscal year and what people could and should be doing with social media. Like all good battle plans, it did not survive the first clash of the forces. Why? Social media has not crossed the proverbial chasm yet. During a recent presentation from James Staten of Forrester, he stated that Facebook was up to 500 million users, based on a worldwide population of approximately 6.5 billion, which means they have market penetration of about 7.7%. Needless to say, social media still has a long way to go in terms of growth.

In the room and on the call, we had the social media evangelists, the social media indifferent and the social media detractors. As we started talking, and we never made it past the third slide, it became clear that what we were missing in our social media efforts was four things:

• What Is My Voice – As we learned from The King’s Speech, new media changes the way we communicate and it can be hard to find your own voice. What is right to say, what is safe to say and how do you do it in such a way that the voice is genuine and spontaneous, so it is a conversation and not corporate-driven (or, as some call it, marketing)? Everyone has a voice and they just need to take the time to find it.
• Unlocking the Passion – It may sound cliché, but some people really love their jobs and we have to let them find ways to express those subjects for which they have passion and then get out of their way. As we discussed the way we thought social media should be done, we discovered that we were missing champions for key causes and subjects close to our business (stay tuned for those folks entering the social media fray).
• Rules of Engagement – We are still a public company, and that means we do have to have a few rules to keep us out of trouble, and this was the one thing that made people reticent to engage, because everyone knows someone who got in trouble at work over social media. We need better training and guidance to help people safely find the edges of the box.
• Social Media Information Overload – Some of us who would fall into the evangelist category of social media users, who also do it as a hobby, need to help folks wade through the social media ocean and let them get their toes wet before they dive into the depths.

What was clear today is that many people are still learning what this social media thing is, but they have great voices that need to be unlocked, because they have good things to say.

So what do you think? Are the social media evangelists crazy? How have you found social media to be the most effective tool? Are you a detractor who thinks social media just a fad or a crock? (I realize if you are not into social media, you might never read this, but I felt the need for the question to be complete.)

Posted in Uncategorized | No Comments

Big things are made up of small things, and the biggest small thing in the 10Gb Ethernet (10GbE) world is a cable. There are millions of cables covering thousands of miles of raised floor space in data centers across the globe. Picking the right cables technology is a key driver of switching choices, data center costs and simplifying the deployment of any networking. As we move forward in the 10GbE market, the move to 10G Base-T and copper CAT-6 cabling for Local Area Network (LAN) on Motherboard (LOM) and Network Interface Card (NIC) connectivity will be a major driver of lowering price per port and moving 10GbE into the volume rack server market.

In the chart above, from Crehan Reseach, you can see that the KX connections tied to blade implementations are growing at a steady rate, but 10G Base-T is expected to grow exponentially. This is a result of the growth of 10GbE LOMs, Daughter card LOMs, NICs and Converged Network Adapters (CNAs) that use traditional copper cables. This growth will help drive the cost points for 10GbE and make that the primary networking interface for new rack and blade computers in shipping ports within a couple of years.

This market transition is built on the simple move to lower-cost 10G Base-T that drives the price per port and volume of adoption. At Emulex, we have been telling folks for the past few years that you have to win the 10GbE NIC business to become the leader in the Fibre Channel over Ethernet (FCoE) and iSCSI markets. Like most things in life, it is the little things that matter and a little connector is going to make a big difference in the I/O market for many years to come.

Posted in Uncategorized | No Comments