Amazon/ Cloud Customers: “Trust, but verify”

While Amazon continues to recover from their Cloud outage, it seems that some in the industry are throwing some FUD (fear, uncertainty, doubt) their way.  After all, Amazon has been on an amazing run within the Cloud business as they continue to build-out new datacenter regions while continuing to reduce prices to their customers.

Based on the many articles written on Amazon I have the following observations.

  • Confusion rains supreme within the world of Cloud computing
  • It’s time to stop pontificating and start solving problems
  • Cloud computing isn’t simple
  • Putting Cloud in front of every product’s name isn’t helpful
  • Availability zones are misunderstood
  • Cloud does not detract from personal responsibility
  • We need more information / facts
  • Those who think this outage proves Cloud computing isn’t ready for prime time are missing the boat

While Amazon has the ultimate responsibility for this outage, why didn’t their customers have a contingency play for this scenario?  Did everyone simply think Amazon could never go down?  Haven’t we learnt anything from past outages of Google, Microsoft, and others?

Earlier this year I wrote a blog post entitled “When the Cloud Goes Down” detailing an experience I had with a provider.  The black box mentality of the Cloud needs to be replaced with an openness and transparency that does not exist today.  A dashboard showing status and health of the Cloud is simply not enough.  We need the ability to monitor and manage our slice of the Cloud independently of the Cloud provider.

Ultimately, we may have seen the perfect argument for the Hybrid cloud; defined as the ability to provide some resources on a private cloud while accessing additional resources on a public cloud.  In a Hybrid cloud model, customers would have the ability to swing services from the public cloud to their private cloud or to other public cloud providers to avoid outages.  Where is the Amazon VM Export capability?

In the end, I’ll borrow a famous phrase from President Ronald Reagan, “Trust, but verify.”  Your business may depend on it!

Advertisements

When The Cloud Goes Down

Has everyone seen the ‘Slip Slidin’ Away’ Bridgestone Blizzak television commercial? It brilliantly depicts what it’s like to lose control of an automobile on a snowy or icy road.  Anyone who lives in a winter wonderland or visits one can easily relate to that terrifying feeing of having no control and helplessness of something you depend on.

Imagine having a similar feeling of helplessness when your Cloud provider has a major outage that affects your entire business.   Perhaps it’s your email system, Intranet, documents, CRM site, or another key application.  Worse, what if it’s your IAAS provider that houses development or customer facing systems?

Naturally, you quickly turn to your Cloud provider for help.  Suddenly, your love of instant messaging, web forms, status pages, and social media turns into shear panic as you realize you’ll never get to speak to a human being.  You are greeted with a message saying, “We’re sorry.  Our services are temporarily down.  Our technicians and engineers are hard at work to resolve the problem. Please check back or follow us on the myriad of social networking outlets we support.  We’ll be updating this site every 15 minutes.  We appreciate your business and your patience.”

Don’t worry; the next 15 minutes will pass faster than a quantum torpedo detonating against a Borg cube, as you’ll be spending it fielding calls from people of all levels in the organization.  After hitting the refresh button on the browser you’re greeted with another message saying, “We are in the process of restoring our services.  The approximate time to completion is 5 to 6 hours.  We apologize for the outage but remember we haven’t had one in 2 years.  We appreciate your business and your patience.”

Far fetched?   I thought so too until I had the pleasure of experiencing it first hand.  In the aftermath, no guaranteed SLAs or credits could make up for the headache I had.  As someone who ponders, evangelizes, analyzes, and designs the next generation datacenter (the Cloud), this was a first hand lesson in the importance of continuing to radically re-think how we design, manage, monitor, predict, and recover the Cloud.  In other words, it’s time to stop putting lipstick on the technology and ideas of yesterday and make room for something different and innovative.

Finally, I’ve never really liked the term Cloud.  It implies simplicity or ease of use that may be prevalent on the front-end (users) but masks the reality of the complexity on the back-end (administrators).  The reality is nothing is 100 percent and “even the best laid plans go awry.”  The key is to understand that while technology is awesome, it pales into comparison to the power of being human.

IBM Throws a Pebble at Cisco UCS; Buys Blade Networks

In the wake of increasing competition from the likes of Cisco, HP, vEMC (VMware plus EMC), IBM responds by purchasing Blade Networks.  For those who have never heard of Blade Networks, Blade was mercifully spun out of Nortel Networks and has hundreds of customers and several hardware OEM deals.  Coincidentally, I think not, Blade has been a long time partner of both IBM and IBM/Netezza.

After years of transforming themselves into a software/services company, IBM is being forced back into the networking business.   While some have postulated that “IBM has turned their back on Juniper Networks”, the reality is Juniper’s baggage may be too big for even IBM to swallow.  Additionally, IBM’s purchase of Blade Networks is a pebble across the bow of Cisco and will do little to anger one of their most strategic partners.

Blade gives IBM a converged networking fabric company while eliminating their competitors from Blade’s technology; namely HP, NEC, and SGI.  Additionally, Blade provides IBM a way to ‘dip their toe in the water’ to see if the market, customers, and partners approve of this new direction.  If IBM is truly looking to challenge UCS or Matrix, then they need additional pieces to this puzzle.

What IBM needs is a new platform ala Cisco UCS that eliminates the baggage of the original blade systems; optimized for density and space.  They must examine how to better integrate their storage platforms with their blades using FCoE and perhaps should look towards a true Multi-Hop FCoE solution.  They must revolutionize virtualization and I/O as perhaps no one else on this planet has the experience, patents, and real world deployments as IBM.  Finally, IBM has the opportunity to rethink management by acquiring, integrating, and refining their current solutions.

If IBM needs a little inspiration, then they can look to their long time bitter enemy Oracle.  While virtualization, fabrics, networking, server chassis, and storage is interesting, applications are still king.  Oracle’s vision is clear; you can run our apps on any server or virtualization platform you want, but it just runs better on Oracle.

The last time I checked, IBM is still Big Blue and they have an arsenal of technology at their disposal.  The question is  ‘if’ and ‘when’ IBM will wake from their slumber and lead the industry once again.  Aside from a blockbuster merger between IBM and Cisco, … hey, one can dream… your move Dell.

VMware Two Years Later: A Microsoftian Transformation

It’s hard to believe, but two years ago I wrote a blog post railing against VMware’s assertion that the operating system was dead.  This was about the same time that the great VMware exodus occurred and VMware replaced Diane Greene with Paul Maritz an ex-Microsoft executive.

What do you do when you are backed into a corner?  You release an upgrade that breaks your install base.  You fire your CEO and begin to lose key personnel.   Hire a Microsoft Executive to become your new CEO because he’ll be out of a job soon.  Pick a fight with the largest software company on earth.  Awake the sleeping giants in IBM, Sun, HP, and more as they want their slice of the pie.  Finally, you bring your most loyal customers to VMworld and proclaim that you are still the king.

Looking back, I failed to understand the transformation that VMware was about to begin.  As a technologist, it’s hard to accept that great technology does not equate to great profits.  VMware’s leadership understood this and set down a path to become the new Microsoft.  I write this with the highest respect and admiration as few software companies in the world have achieved the growth rates, sustainability, profits, and broad market presence as Microsoft.

Today, VMware is a marketing driven technology company.  Look no further than VMworld 2010 as an example of this transformation.  While there were labs and hard-core technology sessions, the message was always on-point and precise, Virtual Roads. Actual Clouds.  While cultivating a legion of developers and partners, VMware repeated that “if you want it cheaper, faster, better” then you’d buy it from them.

While I do not always agree with VMware’s vision of the future, it would be foolish to dismiss their plans and underestimate their marketing efforts.  VMware’s plans fall heavily on “The New Infrastructure”; it looks and feels a lot like the old Microsoft as “The New Infrastructure” is made up of VMware’s lucrative products and powerful partners like EMC, Cisco, Google, and Salesforce.

Give credit where credit is due, Maritz and VMware have done a masterful job in transforming the company while painting an obtainable vision of the future.   In order to make the transformation complete, VMware must cultivate one final group, the application vendors. For once the “old guard” embraces the “new cloud” paradigm, the world will forever change and the provider of this technology is going to be worth billions and billions of dollars.

Cisco’s New Data Center Products Trump HP’s New Avaya Partnership

HP continues to plug-the-holes against Cisco by signing a three-year agreement with Avaya.  The agreement calls for HP sales and services teams to be trained and certified in Avaya call center and unified communications.  HP sales teams will have the ability to resell Avaya and offer outsourced management.

What’s interesting about this announcement is that HP now has agreements with multiple competitors for the next generation data center including Brocade and Avaya.  If you add the potential overlap between Avaya and Alcatel-Lucent mixed with a bit of Microsoft then you have the danger of some explosive conflict.  Of course, HP is no stranger to handling this type of conflict.  However, would a future Avaya acquisition make better strategic sense for HP?

Meanwhile, as HP strengthens their partnerships, Cisco strengthens their next generation data center arsenal.  Once again, Cisco has trumped their competitors by introducing FabricPath, a superset of the emerging IETF standard called Transparent Interconnection of Lots of Links (TRILL).  Remember, Brocade is committed to TRILL within their recently announced Brocade Virtual Cluster Switching (VCS).

FabricPath is an upgrade to NX-OS that combines the best of layer-3 routing and layer-2 switching allowing for scaleable data centers with predictable network performance.  Take the following example from Cisco that was featured in an article by Kevin Komiega in InfoStor:

“With spanning-tree you have multiple links which are blocked and a high level of oversubscription. With FabricPath you can build a scalable, flat, non-blocking network with two layers and no oversubscription with a 16X improvement in bandwidth performance,” says Nikhil Kelshikar, product marketing manager for Cisco Nexus 7000 Solutions.

Additionally, FabricPath and Cisco’s new F-Series modules for the Nexus 7000 allow Cisco to combine six Nexus 7000 switches into a single product.  Cisco is offering a pre-packaged solution called FabricPath Switching System (FSS) that can be grouped in clusters of eight to allow for 160Tbps of raw switching capacity.  Wow, did I just write that?  Not to mention the fact that Cisco just took the air out of Juniper’s Stratus Unified Fabric.

If that wasn’t enough, Cisco announced the availability of a software release of WAAS that can be run as on on-demand service on the Cisco ISR.   Next, Cisco rolled-out new cloud deployment professional services and Cisco introduced a new Catalyst 4948-E Switch with increased capacity, performance, and automation.

Finally, Cisco is introducing Cisco Intelligent Automation Solutions for IT Services.  Building on Cisco’s acquisition of Tidal Software, Inc., they are releasing new versions of the Tidal Enterprise Scheduler and Tidal Enterprise Orchestrator.  As any reader of this blog knows, I am very interested in the autonomic aspects of the next generation data center and I hope to obtain additional information about this solution.

In this latest round of the battle for the next generation data center, Cisco’s products trump HP’s partnerships.

Quick Alert: HP and Brocade; “…Be Nice Until It’s Time to Not Be Nice…”

The HP Technology Forum 2010 is in full swing as HP and Brocade continue to “be nice” and reaffirm their decade old partnership.  According to Brocade’s press release, “…Brocade and HP enjoy the largest SAN customer base with over 3.5 million HP B-series ports installed.”  Note: HP OEM’s Brocade’s technology under the HP StorageWorks brand.

What did they announce? A new 64-port 8Gbps Fibre Channel module for HP’s StorageWorks SAN switches, a new host bus adapter card for HP’s BladeSystem C-class servers, and a the availability of HP StorageWorks P2000 G3 Smart Array that creates a deployment ready bundle of storage arrays, SAN switches, and HBAs.

Brocade has a two-pronged strategy: Arm the likes of HP and IBM with OEM’d products and provide a competitive vision of the next generation data center within Brocade One.  This “have your cake and eat it too” strategy is needed to combat Brocade’s largest rivals while shielding them from Cisco.  However, whatever revenue benefits this brings Brocade in the short-term may be overshadowed by an inevitable showdown between Brocade Foundry and HP ProCurve in the long-term.  After all, if you are Brocade you don’t spend $3 billion on Foundry Networks to play third or fourth fiddle and if you are HP you don’t spend $2.7 billion on 3Com to “margin share” with an OEM.

In the end, I’m reminded of a scene in Road House where Dalton says; “I want you to be nice until it’s time to not be nice.”  And how will you know when not to be nice?  When Mark Hurd says so.

Enterasys Joins the Datacenter Fight; Bringing a Knife to a Gun Fight

Enterasys joins the growing list of network equipment vendors unveiling their vaunted datacenter strategy with the hopes of derailing their larger competitor’s plans.  Enterasys’ hope hinges on a “vendor agnostic” datacenter strategy and is based on their S-series switches with their policy-based operations model.  By supporting multiple virtualization and storage vendors, Enterasys is banking the farm on best-of-breed architectures and emerging IEEE standards.

While Enterasys’ strategy looks good on a press release, it may have been their only option.  Faced with the reality that Dell has decided to team with Juniper Networks, IBM is remaining “neutral” (for now), HP does not need another switch, Brocade purchased Foundry Networks, and Cisco is Cisco, there is little room for Enterasys to partner with a major U.S. manufacturer.  However, would an International partnership make sense?  Aren’t they stronger outside of North America?

Additionally, does Enterasys really believe that Cisco, HP, and Brocade will not support 3rd party vendor solutions?  The last time I looked, Cisco’s MDS supports EMC, HDS, IBM, HP, NetApp, and more.  Cisco’s UCS-B supports multiple hypervisors including offerings from both VMware and Microsoft. Finally, do you remember Cisco’s Open Partner Ecosystem that “helps stimulate technology innovation, augment service delivery, and accelerate market adoption of Unified Computing”?  The same could be said, in varying degrees, about HP and Brocade.  Therefore, what is new or different about Enterasys’ solution?

Perhaps, Enterasys’ will attempt to differentiate their virtualization support via their superior policy-based security and templates.  However, isn’t that simply a different way to do virtual machine profile mapping ala Cisco and Brocade?  Don’t get me wrong; Enterasys has some very interesting and innovative technology but their lack of market power and limited product portfolio puts them at risk of getting lost in the shuffle.

In the end, Enterasys faces an uphill battle against their larger and higher profile competitors.  The war for the next generation datacenter is in full swing and Cisco and HP are proving plenty of “shock and awe” while Juniper, Brocade, Dell, and IBM have their guns locked and loaded.

In the future, Enterasys’ relationship with Siemens Enterprise Communications may yield additional firepower.  However, right now Enterasys is bringing a knife to a gunfight.

%d bloggers like this: