IBM begins Cisco recon with Blade

IBM quietly has entered into a patent cross licensing agreement with BLADE Network Technologies, Inc. BLADE was formed when Nortel sold parts of their Blade Server Switch Business Unit to Garnett & Helfrich Capital. For the past three years, BLADE has been providing IBM with data center networking solutions for their BladeCenter, iDataPlex, and System Cluster solutions.

Why is this announcement important? First, BLADE’s technology (via Nortel) is decent and innovative. Second, this announcement gives BLADE the ability to create products for the next generation data center to compete with the likes of Cisco and HP. Finally, IBM gets a vehicle (BLADE) to attack Cisco’s flank and create solutions around BLADE’s products.

But wait, doesn’t BLADE produce products for HP? Sure, but HP has all the pieces they need to produce next generation products without third party solutions. How much longer will that relationship last? Wouldn’t HP want ProCurve blades in their switches?

Time, money, M&A, and visionary determination are going to shape the next generation data center. IBM has been doing recon on this space for years and its time for them to act. IT is changing, are you?

Sun fires back at Cisco

Sun released details regarding their new Sun Blade 6000 Network Express Module (NEM). The NEM is a 10GbE switch that requires zero management and provides rapid network virtualization provisioning. The NEM plugs directly into the blade chassis using PCI-Express connectors. This reduces cables, space, cooling, etc. by integrating the switch into the blade chassis itself.

Perhaps the most interesting part of this announcement was Sun’s use of SSD sticks that plug directly onto the motherboard. The SSD serves as a cache drive to speed up performance while drastically reducing energy consumption.

When the rumors of Cisco’s California Server (Now UCS-B) first surfaced, I was convinced that Cisco would create a blade for the Nexus. Instead, my hopes were shattered when Cisco released a separate proprietary box. In the end, Cisco will be attacked on a variety of different fronts; proprietary nature of their solution, storage and server expertise, management capabilities, application knowledge, abysmal software track record, and more.

Sun’s opening salvo on Cisco begins with open vs. proprietary and crescendos with “you don’t know crap about server and storage architecture”. While Sun trails in blade market share, this announcement shows they are still investing in R&D to provide innovative products to the market.

Finally, Sun’s announcement garnered little fanfare. Can you imagine what this would have looked like under IBM’s umbrella? Under the right management team with the right strategy, a Sun acquisition is a gem. Cisco, are you listening?

Is Cisco’s UF and UCS enough?

Cisco likes to refer to themselves as a software company, but this week they showed that perhaps marketing is their real forte. After-all, only Cisco could create such hype around a blade server. While the product announcement was strong on theater, it was weak on substance. Not to worry, they’ll tell us what they really meant at the end of the month.

Fundamentally, Cisco’s proposed architecture and solutions are not new. Cisco themselves have been reselling re-bezeled HP servers within their VOIP solutions. Server vendors such as IBM and HP have offered networking within their respective blade servers. Finally, most networking hardware companies sell appliances that are basically locked down servers running specific applications.

What is new is Cisco’s attempt to control the interaction between networking and virtualization. In a classic Cisco move, they have introduced a new proposed standard to the IEEE called VN-Link. I am hearing echos of EIGRP. Will other vendors support this effort?

Finally, is Cisco overestimating their influence over current datacenter design and thinking? The datacenter is a salad bowl of politics, equipment, and priorities. Sure, Cisco’s core networking equipment is key to many datacenters, but you’ll also find “speciality” gear from F5, Brocade, Woven, Force10, and more. Cisco will need to broaden their influence beyond traditional networking to gain application, storage, and management credibility. Simply launching a new architecture, releasing a blade server, and offering a band-aid approach to management combining Cisco, BMC, and VMware’s offerings is not enough.

IBM may buy Sun: Why?

On September 2008, I asked the question, “Is the Sun rising on Cisco?”  If Cisco truly wanted to redefine the industry, then combining Sun’s portfolio with Cisco’s business and market acumen would have been brilliant.  Ultimately, Cisco chose to go the route of UCS while Sun remains an attractive takeover target.

Today, Reuters is reporting that IBM is in negotiations to purchase Sun for “at least $6.5 Billion.” However, I find it hard to believe that IBM would pay a 100 percent premium for a company that is losing money, talent, and market share. Additionally, the speculation is IBM wants Sun to bolster their server products but the diamond in the rough may lie within Sun’s Virtualization and Software Divisions.

Sun’s has many strengths:

  • Sun has impressive high-end and mid-range servers
  • Sun has wide and deep storage offerings including StorageTek
  • Sun’s OSS Reach: MySQL, xVM, Java, opensolaris, OpenOffice, GlassFish
  • Sun has a solid channel and international strength
  • Sun Solaris is still rock solid
  • Next generation database: Greenplum

What IBM would gain:

  • Sun’s 9.3% server market share
  • Database market share dominance over Oracle (DB2 plus MySQL)
  • Instant credibility within x86 virtualization and cloud computing
  • Increased market share in storage and storage back-up
  • End the Java feud and intellectual property war with Sun
  • Disruption with opensolaris, openoffice, xVM, and MySQL

The Risks:

  • Sun is bleeding money
  • Sun is bleeding talent
  • Sun is stuck in the dot-com bust
  • Sun needs to revitalize their channel (more products service)
  • Sun needs attention and leadership

In the end, I am rooting for Sun.  I wrote that “if there is any executive team on the planet that can revive Sun it is John Chambers and company”, but I am sure that IBM could do the same.  What Sun needs is a machine to sell, market, and develop their vision and strategy across an expanded channel and customer base; IBM fits the bill.

I hope Reuters is correct, and today is a sunny day for the employees at Sun and the industry at large.

Cisco UCS: Let the war begin

Over the past few days, I’ve been reading about the revolutionary nature of Cisco’s datacenter 3.0 strategy. I’ve read everything from “a sharp turn” to “Cisco will shake up the market.” The reality, is Cisco has crossed the demilitarized zone between hardware and server vendors and the war has just begun.

It is clear that BMC is Cisco’s new management vendor of choice displacing HP (Opsware). While BMC has tier I products such as Remedy and BladeLogic, they aren’t particularly strong in network or storage management. Furthermore, it is not clear what are the capabilities of Cisco’s new UCS Manager nor to what degree BMC has integrated their offerings.

Additionally, Intel is walking a fine line by throwing their considerable weight behind Cisco’s UCS-B blade server. After all, who pays the bills at Intel; Cisco or HP, Dell, and IBM. While the Nehalem servers are set to anchor Intel’s push into the datacenter, they’ll need more than Cisco to reap the rewards of such an innovative design.

In the short-term, HP is clearly positioned to be a thorn in Cisco’s side. HP has awesome server products, a full line of storage products, decent networking capabilities, full management functionality, and EDS. In the long-term, IBM’s deep pockets combined with smart M&A activity could derail Cisco’s plans. Juniper may be a tantalizing pick-up, but IBM could yield positive results by purchasing smaller players such as Brocade, Woven Systems, or Arista Networks.

Finally, what will happen to Dell and Sun? Could Dell be shut-out of the next generation datacenter? How does an already weakened Sun survive Cisco’s attack? Will they seek shelter or come out swinging?

Cisco has fired the first shot that started the great datacenter war of 2009. It may take three years or more to declare the winners and losers, but one thing is for sure; IT will never be the same again.

Cisco UCS: A Tech Hangover

Nearly 24 hours after Cisco officially released UCS to the world, I awoke to a “tech hangover.” My excitement of the prospect that “California” would redefine the datacenter for the next generation was shattered with the reality that it was simply a blade server wrapped with management capabilities. Furthering my hangover was the reality that Cisco’s datacenter vision is proprietary, exclusive, and filled with hubris.

To combat this tech hangover, Cisco is offering a marketing campaign, analyst coverage, and cutesy videos/data sheets of this groundbreaking solution. Furthermore, I am sure that Cisco University is busy writing books and papers in support of the eventual launch of certifications (CCIE-UCS).

In the end, Cisco will sell UCS-B systems, but to whom? Will server teams accept Cisco? Or, will UCS-B become the networking server (built for and designed with networking needs)? Is UCS simply a launch platform for Cisco’s software?

Cisco’s real challenge is competing against companies of equal or greater size and market power that represent a portion of Cisco’s overall revenue. Additionally, Cisco is betting the farm on VMware; will this bet pay-off?

As you enter the data center this morning, try to calm the nerves of your IBM BladeCenters, HP BladeSystems, AIX servers, HP-UX servers, and of course, the old man of the group, the obsolete mainframe. Oh, try not to disturb the Junipers, Brocades, or ProCurves because ignorance is bliss.

Cisco UCS: A Let Down; All eyes on Big Blue

Today Cisco launched the third piece to their datacenter 3.0 strategy by unveiling Unified Computing System (UCS). UCS is a essentially a fancy name for a blade server built upon Intel’s new Nedhalem processor platform. Cisco believes by wrapping the datacenter in a web of UCS, Unified Fabric, and good old fashioned networking all managed by the UCS manager will redefine scalability, flexibility, and TCO within a a datacenter and beyond.

Perhaps the innovator’s dilemma has finally caught up with Cisco because I expected more from them then simply launching a blade system with the Cisco badge on the bezel. Why not truly unify switching, routing, server, and storage under a single chassis? Why not unify the backplane for speed and efficiency while offering breath-taking performance and value on a per square foot basis?

Instead it looks as if the old one-stop-shopping theory of purchasing is in full effect. Of course, you’ll be able to manage everything from the UCS Manager. If that was the case, why is BMC in the picture? As Cisco doesn’t have a great track record of software, is this any different?

Cisco is an amazing company with an extremely talented executive management and engineering teams. They are in a truly unique position to change the world as networking, storage, servers, applications, and transport all become one. However, the challenge is how to make this transformation without opening the door for their competitors as they move from the Catalyst to the Nexus. This announcement and product launch is too cautious for the risk taking Chambers we’ve all come to love or hate.

Now, let’s wait and see what IBM’s got up their sleeve to combat his new entrant to the blade market. Do they have another SNA up their sleeve?

Follow

Get every new post delivered to your Inbox.

Join 270 other followers

%d bloggers like this: