Is Cisco’s ACI vaporware?

I am wondering if Cisco’s ACI Software Defined Networking (SDN) solution is vaporware. We will start by looking at the traits of vaporware, which include:

  • Product announced as a result of a competitor’s action
  • Loosely or undefined release date
  • Non-Specific set of features
  • Lack of product demos or customer betas

vaporVMware’s purchase of Nicira seems to be the catalyst for Cisco’s stepped up campaign against “software only” network virtualization, a form of SDN. Prior to the VMware’s purchase of Nicira, Cisco had been dismissive of the technology in general. In the past, Cisco has said that SDN was not a feature that customers desired. According to Cisco’s earlier comments, customers did not want network virtualization per say but the ability that it brought. To Cisco, the goal was not abstraction of the control or data plane but the ability to control the network via API’s. By putting this capability in the hardware platform instead of abstracting it via software customers could reap the benefits of SDN such as orchestration without giving up performance.

The announced ACI product has these stated traits. It is but labeled a hardware & software solution opposed to a “software only” solution. By leveraging ASIC’s, Cisco promises to provide the performance comparable to traditional networks while delivering on the operational promises of SDN. Here’s the list of components that make of Cisco ACI architecture according to Cisco’s product page.

Of this list of components, only the Nexus 9000 Series Switches are available for purchase. It raises a red flag when a vendor has non-specific references such as “Software and hardware innovations” and “An open ecosystem of network, storage, management, and orchestration vendors.” To Cisco’s credit they did release an impressive list of hardware and software partners supporting ACI. Their partners and customers rave about how the idea of ACI is fantastic but no testimonials on what they’ve seen in beta or demos. Start to peel the layers back and ask what’s really fully baked, ready for enterprise scale versus what’s still just on the whiteboard.

The Nexus 9000 seems like a fine enough product for physical port connectivity and is a starting point for supporting the ACI approach. However, the APIC is the central part to Cisco’s ACI architecture. Without the controller, the Nexus 9000 is not very compelling if one believes in Cisco’s hardware centric approach to SDN. So, it’s with their APIC that we are waiting to see some hard data on in the form of customer betas and release dates.

Slipping, timelines are another red flag of vaporware. When Cisco announced ACI and the acquisition of their Insieme “Spin-In”, the original release date for APIC was April of 2014 according to most posts. However, I searched Cisco’s site for “APIC release” date and looked at the product page, and the only reference I could find for the release date is the first half of 2014. This gives some flexibility on the release date up to June, a potential two month slip in the initial release date.

So, let’s review. We have a product that seems to be a result of competitive pressure in ACI, we have a critical component that has an undefined release date and, we have no customer or partner testimonies discussing the beta program with early prototypes in hand. I might be missing something, but it is winter in Chicago and my sinuses are starting to clear up looking at this product.

Don’t get me wrong. I think Cisco’s ACI approach is a great idea, but building today’s infrastructure on tomorrow’s products is a pretty tall order. I’d love to hear your thoughts. Is Cisco’s ACI solution just a yet to be released product or vaporware intended to delay customers from jumping to existing SDN products?

Published by Keith Townsend

Now I'm @CTOAdvisor

21 thoughts on “Is Cisco’s ACI vaporware?

  1. Really excellent post Keith,

    The thing that really bugs me about it is Cisco’s approach and how they position and market seems makes me feel like they think the market and customer base are just dumb. Its fine that the release dates are next year, it is an unique approach so they have a lot of ocean to boil. But dont try to tell me yours is the first real solution, dont tell me that this is the only way to do sdn, software only solutions wont and cant scale blah blah blah. What do they think we all live under a rock … like we hadnt already seen the several large/complex NVO implementations that SP’s were presenting back when Cisco was still anti-sdn?

    If Cisco’s solution is great it should win on technical merit but no matter how great their technologists are, the final products all coincidentally happen to align with Cisco’s business interests – like the laws of the universe all just happen to align with funneling money into Cisco’s coffers lol. Dont get me wrong, its not like other businesses are more altruistic, but when you take a step back it is all very rational. Consider the x86 server and the networking markets, or even any B2B solution market of any sort, Market A is highly diversified, very competitive, standardized, multi-sourced etc, and Market B is largely dominated by a single vendor and characterized by the typical patterns we see in monopolistic markets. You dont need to know what type of market this is at all and purely from history and economic science we would expect that the more competitive market would have developed better practices to advance the industry because in the absence of more dominant players more compromise is both required and less threatening to established business models in a beta-max/vhs or blu-ray/hddvd situation where standards are central to enabling the market to move forward and one dominant player cant exert undue influence, compromise happens..

    And that is exactly what we see, in x86 the industry has moved forward and according to even Cisco, left the networking industry in the dust, margins stay in a healthier range for industry consumers, and I believe that those lower margins are also healthier for the vendor – trying to compensate for exorbitant margins is Cisco’s largest business challenge today. That type of excess only causes unhealthiness. There is no such thing as a sustainably differentiated technology, all technologies and technology paradigms fail, the only thing that can be sustainable is a culture of continuous innovation and exorbitant margins are the enemy of continuous innovation.

    Its good that we have differentiation it is critical for the sake of innovation that there are very different offerings. But as Cisco says very clearly with ODL, there is a right place to be common and a right place to differentiate and they understand that with ODL because it aligns with their business strategy in that specific instance. But in their home turf of the switch, their books are still hung on exorbitant margins and that is a huge challenge for them. So in my view, the whole plan is to do something different solely to peddle FUD to try to preserve margins. But I believe customers understand or will come to understand that the methodology used to innovate and advance the x86 and open source communities is far better than the vendor dominated approach that has been used in networking. And then they will have to really live up to their reputation as until now they have been like the kid who has his parent sell all his cookies and wins the big prize at school … their supposed excellence has been fueled by the monopolistic, unhealthy market they were bleeding dry. Lets see how they can keep up now with the rest of the players that have already been slugging it out in commodities 🙂

    1. Art, great thoughts. I knew you’d find this post.  I wrote the whole “VMware One Hit Wonder” post. However, Cisco isn’t too different in the sense that they are a really big organization that is driven by the need to maintain their revenue and size. The industry is in the middle of a period of disruption. Cisco is more than capable of producing a product that will meet the technical needs of their customers. The influence of the business drivers may be too much of an inhibitor. This is a question that I’ve wanted to tackle in general.

    2. Hi Art, few thoughts here. not really a response probably…
      I’m ex-Cisco that said I agree with you that ACI is an attempt to secure the dominant position of Cisco. What would you expect them to do? Go with full SDN SW based solution and fight the VMware and everyone else on their turf? From business point it doesn’t make any sense (at least to me)
      If you look at at technical aspects of the solution – 2 schools of thought here. Both have merits, both have disadvantages. Depends on use case, scale and customer’s idea on how to do networking. I’ll take the middle road and say there will be customers for both NSX and ACI. And stuff like OpenStack. First let’s wait till both solutions are fully available.

      Do I believe that time of vertical vendor solutions is going away? More yes than no. Stuff like OpenStack and OPC project is very interesting and has great potential. There is a customer base that is looking to get out vendor shackles. Some did already (Googe, FB)
      I think we are witnessing transition similar to mainframe-to-x86 in networking. Early stages of it. No idea how long will it take. It will largely depend on new payers coming in and ripping the old guys with better/cheaper solutions.
      I am personally looking forward for Cumulus Network and the whole white box networking trend.

      But majority of customers will be conservative if not hostile to that transition (just based on my experience in the field) and that will actually plays to Cisco’s advantage here. There’s far too much hype around SDN/OpenFlow. It got so ridiculous that just having it as part of solution automatically gives some weight to it. I’d rather not be in a position of choosing a network solution in this market 🙂 So would you pick a dark horse or go with a proven leader like Cisco in such climate?

      1. Great points Kanat,

        I agree, I continue to be a little surprised by just how much and how rapid industry changes are happening. Its reminiscent of green energy where it has been difficult to invest in anything becasue say you do a solar implementation, by the time its implemented they already have better and cheaper solar panels and that is just within one tech lol.

        With networking first I was looking at the network market, then with the converged infrastructure trend we have to consider how all of these segments come together and nobody could really know the players across all these domains as well as a specialist within a given domain would and that fuels unpredictability, but it gets much worse as consumer tech is intersecting with corporate tech, public clouds are maturing with enterprise features at the same time as enterprise vendors are finally figuring out how to deliver private cloud, and then an even bigger wildcard is just how amazingly fast the developing world is coming online, clearly we are just seeing the beginning of how that will impact the future tech landscape. then things like the NSA spying come to light and its becoming just how much that and similar could damage our economy as globalization continues to speed forward at a breathtaking pace.

        Now I am not on the list here to try to blindly promote Dell, but I do choose to work for Dell for a reason. We are far from perfect, but compared to what else is out there I think we have done about as good of a job of future proofing under the current challenging dynamics as one could do. Force10 started with a strategy essentially to adapt public cloud type of deployment and provide the software that could enable enterprise users to leverage cloud-class equipment and operational models. And I think that point specifically is of crucial importance. When it comes to white box hardware, I think that market will grow very well for web 2.0, technology startups and service providers whose business models support greater investment in strategic and educated purchasing … this class of customer has always gotten a fair deal on equipment and have purchasing practices that can combat exorbitant margins. Where I am more concerned is about the average enterprise customer for whom IT is not a core competency of the business model … that segment of the market rightfully pays some premium as their business model leaves lots of opportunity for suppliers to provide value add. And it is here I think customers can be much more easily mislead as it is this segment that has been paying exorbitant premiums for commoditized technology for years and have had a real hard time understanding what the problem was when large network operators and software giants formed the ONS.

        I think this segment is the target for both UCS and ACI, or at least the most important middle of the road type customer that can tip the scales during standardization efforts. At the end of the day I think the more that enterprise tech is aligned with cloud tech, the more that enterprises will stand to benefit from the purchasing power of larger and more sophisticated network operators and that would help to keep the enterprise market more competitive and healthy. The more that enterprise tech is different from SP tech, the more opportunity their is for obfuscation and misleading by vendors. UCS and ACI are clear examples of this IMHO and really give Cisco the license to do unique and creative pricing strategies which will only grow if these solutions grow in power … its setting the table for the foibles of the past 20 years all over again.

        But in any case back to Dell … if you bought what we were leading with in 2010 …. S4810 and Z9000 with Open Automation, you would today have something more similar to what Cisco is now positioning than if you bought what Cisco was peddling in 2010 (or even 6 months ago). We know nearly all of our deployments are heterogeneous, so we build for that. All the gear is optimized for leaf-spine or single tier topology, all of it running merchant silicon with single system on chip architectures providing all line-rate, non-blocking, low-latency performance. One important point here is that this physical architecture is the optimal recommendations for whichever upper layer control protocols you are using. So you put in our preferred architecture that is about as future proof as you can get … then look to the l2/l3. Here we provide Open Automation which unlike similar competetive paradigms has the benefit of Active Fabric Manager which I think is about the best way of delivering the benefits of SDN to the average enterprise – at least in the physical switching infrastructure. Our traditional control plane software is solid, unix based, modular, and proven. Open Automation is strong, we have a longer and stronger track record with OpenStack than any of our key competitors and today can deliver full transparent automation for our networking solutions for OpenStack via Crowbar and/or Active Fabric Manager and in March we will be talking about yet another very cool openstack integration. For NVO we have a very strong relationship with both Microsoft and VMware … our decision to not make our own vswitch to this point has been very intentional here as we seek to align with each vendors preferred methodology and our overall architecture is flexible enough to allow that within a single common infrastructure solution. So with Dell you dont have to choose between your virtualization vendors preferred path and your infrastructure vendor … clearly every vendor still needs to mature more in this area but we are already well underway in delivering this more open posture and our track record is a great example here with solutions like Dell Virtual Server Networking which for some time now has been able to provide transparent network automation services for Citrix and VMware environments and we dont even require enterprise plus licensing … standard is fine. You look at Cisco and they dont even have any coordination between a N1k and a physical nexus switches unless you are running a full UCS solution they already decided a long time ago to not allow their switches to leverage the open VMware API that our VSN and other vendors similar solutions are built on, and I do see how that could forward Cisco’s agenda but I dont see how it is good for industry consumers. And with HP, they are also clearly aware that having a more differentiated solution could be more lucrative and I think they are intentionally being overly aggressive with OpenFlow which could force more custom extensions than would be ideal and I dont see the commensurate benefit for premature adoption. My opinion on HP’s use of openflow goes back to my earliest interviews with Martin Casado to the discussions I have with him today … we both shared the opinion that, assuming it is a natural evolution of virtualization to leverage NVO’s, how can you plan for the appropriate enterprise DC networking solution today if the vast majority would eventually fundamentally alter their network with NVOs… e.g, what would the ideal SDN model be in an environment that was 100% virtualized running all NVO’s … clearly planning for this type of environment would have fundamentally different requirements than for today’s fabrics where NVO hasnt really been deployed yet in the vast majority of enterprises. And we also both shared the opinion that NVO’s would be more mature and would likely be deployed before OpenFlow or SDN in physical switches would be ready. (I am not trying to speak for Martin here but rather reflecting on our interviews that have been posted on NWW). And then when I compound this with what Dell research group has been doing … some of the most sophisticated SDN experiments bar none … and being able to do this under the “extensibility without extensions” paradigm of the Split Data Plane approach, it seems to me vendors could be doing a whole lot more innovation in a much more open way if they wanted to.

        Sorry for being long winded … in summary I think the Dell solution puts the right physical hardware and architecture forward regardless of the control logic, and then at the control we have really robust support for traditional, programmatic, NVO, OpenFlow etc, we are a primary source behind CloudNFV and I think that while not perfect we do have the most open strategy of any major infrastructure vendor.

    3. Art ..
      [Disclosure: I work again for Cisco but below are my personal comments].
      It is continuously amazing to me that
      a) making money is seen as a bad thing for a publicly traded company
      b) aligning product roadmaps with business strategy is labeled a coincident
      c) technologists have the misperception that technical merits alone will win
      Sounds old but what matters is focus on what customers want and will buy – technology and product will follow from there if a company can actually execute.

      Thomas (@thomas0002)

      “If Cisco’s solution is great it should win on technical merit but no matter how great their technologists are, the final products all coincidentally happen to align with Cisco’s business interests – like the laws of the universe all just happen to align with funneling money into Cisco’s coffers lol.”

      1. Hey Thomas glad to continue the conversation here … i do have a couple disagreements on these points:

        “a) making money is seen as a bad thing for a publicly traded company”
        IMHO today wall street is completely obsessed with short term over long-term value. When you consider what is best for the entire populace of investors, or just the nation or world it seems very clear that if one were to pursue short-term success at the expense of long-term success that would be a very bad thing for the vast majority of the populous. So I am not saying it is a bad thing to make money, I am suggesting that all of us who have been to biz school these days understand how important differentiation is to making money. And we all understand how industrial & community commons (standards, & norms) can help guide innovation in a more consumer-friendly manner and while this can lessen short-term gains for a few select players, overall it is far better for the ecosystem. But if every company and every CXO and investor were measured on long-term investments I believe to my core this alone could substantially aid in addressing the biggest problems affecting our economy and society. Life isnt about making select few rich imho, it is about doing the best thing that creates the best impact for the most people … as much as possible I think ODL could be a great example of this and applaud Cisco’s increased investment into openstack but overall the strategy looks to me like it is meant to meet the most sophisticated network operators with an attractive solution while continuing to drive unnecessary differentiation in the enterprise market for the purposes of maintaining margin, and it seems to me this is a strategy that would benefit Cisco while doing the opposite for customers and the overall market whether that is realized by the consumer or not. Not all of us spend our life’s analyzing the network market and not just Cisco but all marketers have advanced knowledge on how to manipulate consumer bases and where that type of manipulation is most successful is statistically correlated with less diverse markets. I do hope Cisco has continued success in a more diverse and healthy market and with that I hope that continuing market pressure keeps pushing Cisco in the right direction. There is a saying ‘lest not the fires of war rage so intensely the spoils are ruined for the victor’ … and I think this characterizes today’s wall street to the tee including one of it’s crowning jewels … Cisco, the king at a time where wall street is absolutely crazy and out of touch with the populace.

        “b) aligning product roadmaps with business strategy is labeled a coincident”
        I am not intending to imply this so allow me to clarify. I believe consumers approaching a purchasing decision should do so with clear, objective perspective and there are far too many who fall in line and blindly believe the hype here. Sure it is understandable that Cisco has aggressive marketing and I cant fault them for that but monopolistic and unhealthy markets get that way with full participation from consumers. Management science only recently came to understand disruptions and concentric waves in the economy so there simply havent been business strategies to this point that have been able to accommodate for the actual realities of how innovation happens until very recently … this is one of the biggest challenges for all incumbents and established businesses, when you understand how reality actually works, that we do fall into predictable patterns (I highly recommend this talk: )
        So today we understand that we as a society dont really understand long-term sustainability well, we now know clearly that all paradigms fail, disruption has to happen from time to time and in the future we will anticipate this better but in the past very few understood this well. A good way to understand this is consider Henry Ford’s famous quote “If I had asked people what they wanted, they would have said faster horses.”. When you understand this and human sociology the clear outcome is that markets in and of themselves dont bust paradigms. And even capitalism by itself doesnt either which is why we need more innovation in governance models. When one paradigm begins people cling to it and does amazing things for them and the fitting response for a long time is to continue to build on this paradigm with incremental innovations and by that time the entire market and all the power in the market belongs to the old paradigm that eventually really needs to be busted. We as a society now need to come together and form a governance model that smooths this process as while disruption itself is fundamentally important we as a society need to do more to minimize the negative human toll that can happen with disruption while maximizing the positive benefits. Long story short, some of us can see clearly how our technology and business strategies align, but many do not and I think it should be more apparent to consumers exactly how those align and regardless of fault, I do not think that is happening very well today.

        “c) technologists have the misperception that technical merits alone will win”
        I agree with you on this point entirely, I think all of us, the entire market, all suppliers and consumers would all benefit if consumers were more educated, objective and open minded about purchasing decisions MOST IMPORTANTLY how to align technology strategy to business model as there are so many misalignment in IT shops today. The flip side of this is that greed – an unhealthy level of greed – its very common among most sales forces and sales leadership and these archetypes all have wet dreams about fat dumb and loyal (FDL) customers, many businesses make it an objective to make FDL customers, but the presence of this type of customer is actually like a cancer to the supplier that over time create the innovators dilemma. Centering business around at least a 10-year plan, I think would be among the most effective means of addressing the problem.

        But something big needs to change for this to really happen, we all know that if next quarters numbers are anything less than exceeding optimistic analyst expectation CXO heads start to roll, and this factor is probably the most responsible for damaging the global economy so severely at a time where we should all be enjoying the fruits of technological and human renaissance. I think Cisco has made a lot of positive changes, but not nearly enough … I hope market pressures will continue to push them in a positive direction though and with time we all compete in a much more diverse and healthy market.

  2. [Disclosure – I work for Cisco but have not direct involvement in ACI, consider this as my personal two cents]

    So, your timeline is a bit off–the market became aware of Insieme in March of 2012–the Nicira acquisition was announced in July 2012. If you take the reasonable assumption the Insieme had been in the works for some time prior to the leak in March, then you cannot really claim that it was in reaction to Nicira being picked up.

    As far as your other criteria:
    RELEASE DATES: a window is provided, so its not undefined. It may not be as defined as you liked, but I don’t think a 2Qtr window is unreasonable. We are not exactly a start-up, if our CEO goes onstage to announce a product, as a publicly traded company its a safe bet there is something behind it. If you look at the execs behind Insieme, you will have to admit they have a track record of getting things done. Finally, I like how you took a rumored ship date, looked at the availability window announced, then construed that as a “slip”.

    NON-SPECIFIC SET OF FEATURES: Here is a web page with (at last count) seven white papers on the ACI policy model, data model and operations: Granted I am biased but that seems reasonably detailed to me. 🙂

    DEMOS: They are in the works–folks will be pleasantly be surprised.

    Omar Sultan (@omarsultan)

    PS Hey Art, I believe the UCS is #3 in Blade Server revenue behind HP and IBM, so I think we can hang in the commodity space OK. 🙂

    1. Hey Omar,

      Hope your having a great holiday!

      I work for Dell BTW, sorry I didnt mention on the last comment.

      You guys are definitely doing well and know how to execute. But lets consider an earlier example as hindsight is more clear. Cisco’s path into VoIP was extremely accelerated because of the financial aspect. Anyone involved in selling VoIP in those early years knows exactly what happened … Avaya and Nortel would come in selling VoIP which would almost always end up being sold with Cisco switches. So Nortel/Avaya would present a bill of materials where only the VoIP equipment would be considered, Cisco would be able to present a solution with usually with equal parts VoIP and Switching, they had almost no margin or technology pressure on the switches so they discounted the hell out of the VoIP gear and used the margin in the switches to consistently undercut Avaya and Nortel while retaining strong blended margins. I am not saying Cisco may not have won anyway, who knows, but the financial angle was the lynchpin. Very much like when you were back in grade school and there is some competition to sell cookies or some junk and we all know, some of the kids are going to have parents that sell a ton of it, the rest of us were just screwed, never a fair competition. Cisco will know real competition when coffers arent still being filled up with easy, unsustainable money and despite recent competitive pressures Cisco is still having a pretty easy run selling at huge margin levels particularly in campus. I am not saying it is a bad thing to be selling 6100 series line cards designed over a decade ago … hey if they get the job done who cares, but when we see what margins these old amortized boxes are pulling, there is no doubt that it is and really must be an unsustainable source of funds if we hope to have remotely healthy markets.

      And now look at UCS … samey samey. Not only does UCS (chassis) necessitate an all Cisco access fabric, but really, we all know UCS is for Cisco loyalists. How many UCS are running on anything but a 100% all Cisco end-to-end network? So once again we have the same dynamic … most DC opportunities involve both Network and server, Cisco gets to use the excessive margins in their switching infrastructure to fund the UCS implementation and undercut competitors while still retaining overall very attractive profit margins. True Dell or HP could add storage in the deal but that isnt the same as our storage products all exist in healthy, competitive markets without exorbitant margins so other brands cant play this game, it only works like this when you have some unhealthy source of excessive cash tilting the scales.

      At the surface there is nothing wrong with this tactic – the problem in this case was the extremely unhealthy and unsustainable levels of margin that existed in the Cisco switches … and if markets are working well, then no market should be characterized by lopsided market share and exorbitant margins. You know this Omar, we all do, not that this cant happen for a period but should be unsustainable over time if capitalism is working the way most of us including Cisco customers & employees think it should be when we take a step back or consider any market other than network infrastructure. You take your average Cisco customer and ask them about economics using some industry outside of tech and most of them would say lopsided market share and associated effects are bad, but they have the blinders on when it comes to Cisco. And that is a cognitive dissonance that I think will cause a lot of problems for Cisco as the blinders come off and people become more open minded.

      At the end of the day one thing we can say clearly because Cisco themselves has repeatedly acknowledged this … the networking market is all screwed up and way behind where it should be and behind other domains of technology in terms of innovation. So how is it then that Cisco gets off acting like they are the only ones in the world that can fix things practically in the same breath where they were acknowledging just how bad things had become under their reign. But hey lets give Cisco the benefit of the doubt and look at their plan – and regardless of what you say about the tech, overall the Cisco solution mandates a level of vertical integration we haven’t seen since the mainframe.

      In my view what Cisco is trying to do here is – they feel they need to lead the market so they are using an approach that is differentiated solely for the sake of being differentiated … the sole purpose is to try to keep the perception that Cisco is leading the market when in fact it has not. By the 2nd ONS half of the attending SP’s were already putting pure software-based SDN service offerings into production and by the 3rd ONS they were reporting on their efficacy. And yes things need to evolve but today when we consider how the market will advance it is like the Cisco way vs the rest of the market. So how do consumers choose, do we go with the company who has acknowledged the market they led became very unhealthy with all the hallmarks of the mainframe as they continue to push increasing vertical integration? I think that is what Cisco wants, but SDN was never supposed to just be about the technology, it was supposed to be about the market conditions that lead to the industry becoming so unhealthy and I think Cisco is betting that most people only think of SDN in terms of the technology itself, but in my view the #1 most significant post about SDN is James Hamilton’s “Networking: The Last Bastion of Mainframe Computing”

      This post describes better than any other work what is wrong with the industry and in my view, Cisco hasn’t come close to addressing the problems with the industry that Mr. Hamilton so eloquently nailed. My view … as a longtime networking guy I participated in the unhealthy market for far too long, I am looking to the history of x86 to OCP to the best of open source to try to learn better ways to advance the industry around industrial and commercial commons… and one thing I believe to my core is that competitive, diverse markets are healthier for everyone and I hope ODL will prove to be a good example of that balance … I just wish we saw more similar behavior across the board from Cisco but looks to me that much of what Cisco is doing is quite the opposite.

      1. Hey Art…

        Hope you are getting some downtime over the holidays too. Better rest up, 2014 is shaping up to be an interesting year. 🙂

        As far as UCS goes, I think you are underselling what Cisco has done there, unless you believe the networking guys/gals are now making the server buying decisions, “Cisco loyalists” has nothing to do with it. Different buyer and usually different budget. Much like with storage networking, we started with 0% market share and moved into a market with strong incumbents and very loyal buyers. Cisco delivered a better mousetrap and customers recognized that as demonstrated by revenue share numbers. As Thomas pointed out elsewhere in this thread, the better mousetrap entails more than just cool tech, its about addressing the complete problem–as they say people don’t buy shovels just to buy shovels, what they really want is a hole. You may view this as unfair competition, but I view it as really understanding what a customer wants. At the end of day, I am not sure how many customers care about the deal structure and where we make money–in my experience, they look at bottom line, expected TCO and see if the investment made is justified by the value they expect to receive in return. If the only thing someone touts is that “we have lower margins than Cisco” I am not sure how much progress they will every really make.

        As to the state of networking–no doubt, it is ripe for change–things like virtualization, cloud and open source are driving a sea change, which I think is a great thing for everyone involved–customers and vendors. Perhaps one thing we disagree on is the rate of that change. My experience is that customers like to do the disrupting, they don’t like to be disrupted. You and I live on bleeding edge of this stuff and that can sometimes lead to a loss of perspective. The fat part of the adoption curve for SDN is still ahead of us. SPs are certainly aggressive with SDN and NFV, but they can afford to be. The cost model for SP/CP infrastructure is completely different than enterprises, public sector, etc. They can make the investment and take a collection of raw technologies and do interesting and valuable things with them. Most enterprises I have talked to are looking to solve problems not start science experiments–they want to improve user experience not sit there programming flows. The gap between those two is a huge opportunity for whomever is savvy enough to exploit it.

        To that point, our persecutive has always been about the outcomes, not the tech. That outcome is about infrastructure programmability by whatever means and whatever level: OpenStack, OpenDaylight, OF, BGP-LS/PCEP, I2RS, onePK-pick the right tool for the job at hand. One of the reasons you see ACI launch with such a broad ecosystem is because its concerns itself with what we think customers really care about–helping their apps run better. For other customers with a different set of concerns, OpenStack or OpenDaylight might be a better starting point.

        I would love to say our movement on all this stuff is going to be smooth and linear, but, we are a big company and different parts are going to move at different rates. However, if anyone reading is counting on us to dig our heels in or bury our collective heads in the sand, they are going to be disappointed.

        At the end of the day, I whole-heartedly agree with you that “competitive, diverse markets are healthier for everyone.” With that, I wish you a Happy New Year.



    2. Fair Points Omar but still a few things I see a little differently, but I respect your opinion.
      I think UCS market success is first and foremost is a testament to Cisco’s ability to execute in marketing and sales, my opinion. UCS did put forward some new management capabilities, but I think from the time UCS was launched – by that time public clouds were already delivering agile infrastructure without the whole proprietary end-to-end thing. I really think that is unnecessary and there are ways to innovate in a more consumer and ecosystem friendly manner. Now no doubt Cisco has put forth a ton of money and leadership into standardization efforts and they deserve credit for that. But what I am suggesting is that the governance models for innovation and standardization need to mature, and that the computing market has been leading with a more mature governance model for innovation and advancement of standards and ecosystems. Regardless of who is leading any initiative, I think the important thing is that we as an industry use the most effective known approaches for advancing the industry.

      I dont see any problem with Cisco pursuing a vertically integrated approach, in a diverse market I love to see all different types of approaches and that is a healthy thing. The problem I have thouch is essentially the same problem I had with Apple – let me explain. Back in 2004 I was really excited about the progress of open-source business models such as Sourcefire and Canonical. The key reason why Clayton Christensen’s work took off the way it did is because there has been such a trend for businesses to grow & diversify – with the wall street craze businesses have had to expand into adjacent markets and pursue growth like never before. But the one problem was that Christensen i essentially proved there was no reliable or consistent way to ever take on any incumbent, which means these businesses facing pressure to expand into adjacent markets had a huge uphill battle. Prior to that, all management strategists centered around the work of Michael Porter which led to the belief that a business that simply out-executed another business would win. Christensen and contemporaries such as Wiggins & Ruefli (see Schumpeter’s Ghost: essentially proved that superior execution did not result in challengers beating incumbents. The incumbent advantage won in all types of circumstances and the incumbent advantage was strong enough to where the disruptor would lose to an incumbent should the incumbent eventually get on track with the new tech, stealing the spoils from the disruptor.

      So by the early 2000’s awareness of the ‘innovator’s dilemma’ was spreading like wildfire, incumbents had a critical need to expand into adjacent markets but knew the odds woudl be stacked against them. Rising awareness of technology disruption combined with the incumbents advantage offered assurance that even a bumbling, slow incumbent could stay in power over an innovative new market entrant so long as they at some point eventually would respond to the disruption – leaving little options for an expanding or diversifying corporation. Enter Sourcefire and Canonical … while the companies hadnt experienced the huge wall street gains that proprietary models had, it looked to offer a viable model an enterprise could leverage to successfully break into a new or adjacent market. This thinking is the exact place that most who likes the idea of free market capitalism actually want to be. This means that in order for Company X to enter into a new market where company Y is currently dominant, company X would have to not only innovate to cause technology churn, but they would also need to drive interchangeability of components, standards, industrial commons and ecosystem support.

      And then the iPhone happened – and the eyes of every greedy wall street exec lit up like a child on christmas morn. Once Apple started taking off, you just couldnt talk to any executives anymore at most companies about anything open (Circa 2005). I predicted (and hoped) apple would fail – I have immense respect for them but in my opinion they were trying to advance a model that would have been less healthy for the industry … they didnt want their technology to ever commoditize. They want to keep the gold plating on it forever. Consider an example with automobiles … maybe Mercedes Benz introduces a new technology and offer it to their customers a few years ahead of the market, but eventually the tech trickles down to lower end brands and lower end models. They know if they introduce a new tech even when based purely on their own R&D, with a bit of time it will move towards to low end. But for Apple, they wanted to own everything. They didnt want anyone to have any way to access comperable technology unless it came through them, and they used all types of hype and BS to gain a temporary lead but eventually it played itself out and today we see the myopia of that path – no company has a stranglehold on innovation. Christensen never said that no company had ever beaten an incumbent, there are plenty who have, Christensen’s work simply showed that it was an anomoly when they did and no leading management theory showed consistent results.

      Today Christensen talks a lot about what he calls ‘the corporate laws of physics’ … and imho some of those laws are simply that market diversity is needed to ensure a healthy market. The Apple way to break into a new market wasnt simply innovation as management science had already shown that all to often the spoils of innovation went to incumbents and not the entrant. Apple’s model was innovation + proprietary … sure it used a lot of standard components but the solution was just as vertically integrated as if the product were entirely in-sourced. That appears to me to be Cisco’s model. And there is nothing wrong with that model but in my view another ‘corporate law of physics’ is that in a healthy market highly vertically-integrated business models belong in the niche segment.

      Today we are learning a new way – a more sustainable way … it is based on awareness of the corporate laws of physics, that when you want to break into a new market that has a strong incumbent, the way to do it is not innovation alone, not innovation + proprietary … the entire aim is off. As Dr. Ruefli always said, there is no such thing as sustainable, technology-based differentiation. The only thing is a culture of continuous innovation. In a truly sustainable strategy you cannot plan to have some type of proprietary cheat or lock-in to act as your hedge.

      And yes, a company can try to drive this culture independently, but the challenge with vertical integration is that no company own innovation, no person owns innovation … as the open source has shown us … microsoft, sun etc … no company could create systems that could handle the internet, the open source took off becasue it took the community coming together around commons to make the internet possible. The community working together as a whole towards a common objective should always win out in the end, that is my hope.

      Often we lose perspective that business and society lives in symbiosis, society is the host and business is the parasite. The only reason why any business ever came into existense or could stay in existence is an inherent offer of value to the community, but with the level of complexity that exists and the pace of change … Christensen proves that people generally dont gravitate towards newer or better technologies unless other factors are also met. Anomolies happen, unhealthy markets spout up, but in the end for all of us I hope that the market pushes toward a more healthy place.

      1. Art:

        Its good we see things differently–would be boring otherwise. 🙂 Kidding aside, I think the diversity of opinion is good for the market–as you note, no one company, organization or person had the monopoly on good ideas. The thing to looks is if the market is structured to be open and welcoming to diverse viewpoints or if it is exclusionary.

        I think we have to be a bit careful–having a project simply be Open does not imbue it with special powers of goodness. I think the value of open projects is that they can create a venue that supports collaboration, synthesis of new ideas and rapid innovation–or it can lead to projects that grind to a halt in a quagmire of ego, politics and the IT version of Game of Thrones.

        I think the disconnect happens when we start confusing technologies with solutions (back to the shovel/hole analogy)–which is a caution to be heeded by open and closed source acolytes like. I think if you asked the typical smartphone buyer why then went with iOS or Android, you would probably get reasons like price, carrier incentives, apps availability, UI, service–very few are going to care about open vs closed source.

        Maybe one nit to pick is that the Internet took off because of open standards, not necessarily open source software. IP and Ethernet created a stable substrate with low technical and financial barriers to entry for folks to go play and to do interesting, differentiated things. I think its an significant point to make because its important to recognize that embracing open is not automatically a race to the bottom.



  3. Hi Keith.
    I’m ex-Cisco, so therefore I’m not here to preach the Party Line or be a champion of Cisco.

    On your points
    Product announced as a result of a competitor’s action – heard some rumours about Mario cooking up something in beg of 2012, so assertion that ACI is a direct response to Nicira acquisition is kinda thin. I think it was more like response against pure SDN companies in general.
    Loosely or undefined release date – That’s just Cisco and industry in general. Does not automatically mean that ACI will suck.
    Non-Specific set of features & Lack of product demos or customer betas – it’s hot out of the oven and as far I understand ACI will fully function only on 2nd gen HW with ACI ASIC. Which is announced to be released in 2014. I’d expect all the demos & detailed paperwork with that launch.

    It’s true that Cisco has margins/empire to defend and with that comes an understanding of why ACI looks the way it looks. It’s an honest attempt to offer the best solution for customers within the HW based business model that is the core of Cisco’s success. With SDN elements in it.
    To me it looks like an attempt to adapt to changes in the game. It’s not as radical as Nicira, Big Switch or say Cumulus. But can we realistically expect that kind of move from a company like Cisco? The rest is just marketing chest pumping, but you can’t fault Cisco there – everybody does that…
    I have mixed feelings towards ACI. I love the concept of service profiles/UCS manager for entire network where you define elements and their relation. It’s neat. No more CLI to do day-to-day stuff. There will be a customer for that.
    On the other hand I’m concerned that this will be another case when there is a gap between pretty picture on the slide and the real product/feature capability of the solution. The good news is that it’s a brand new play and there’s very little legacy/old stuff it needs to be compatible with (like Borderless Network).
    Then there’s the question of – what’s the positioning of ACI?
    in DC play you got Nexus 2K,5K,7K,9K for HW.
    for SW there’s APIC, ODL, One PK…
    This is starting to look like HP portfolio. So some clear communication both internally and externally needs to happen.

    To summarize – it’s too early to label ACI as vapor. same goes for VMware NSX. We’ll just have to wait and see. 2014 will be the year we got our answer.


  4. I think this is a typical Cisco product plan. Product is announced, it eventually is released, and it’s a definitive 1.0 product. I’m not worried about the release date so much as the version that becomes useable. Look in the security space: FWSM, CSM for CiscoWorks, ASDM, etc. All were horrible products on release. FWSM eventually got shitcanned for ASA blades, CSM is finally starting to be usable but still has a lot of warts (especially compared to competitor products like JunOS Space), and ASDM is now usable.

    That doesn’t mean ACI will be horrible or even just unusable out of the gate, but it does make me wary. Will 1.0 work? Or will I have to wait for 2.0, 3.0, or something further down the road? Will any of that delay require new hardware as well as software? And what’s the timeframe for improvements? Cisco tends not to follow a rapid, iterative build and deployment process (and historically that has been the correct decision, since they’re targeting core enterprises like Tier 1 and Tier 2 network backbones where “fail fast, fail often” is a good way to lose customers), but they’re going to need to if they want to match up with NSX, OpenStack, CloudStack, public clouds, etc.

    There’s one big problem with ACI, though, that will exist even if ACI is the bee’s knees – their competition is software-based. I can play with NSX, Openstack, AWS, etc., on my laptop for just a few dollars. How is someone who doesn’t have an existing Nexus/UCS setup supposed to learn about ACI and see if it fits their situation?

      1. That wasn’t meant as a slight, I was trying to say the release doesn’t look like vapor ware, it looks like a typical, real release. After that, yeah, I’m using past behavior to indicate future trends, but it’s the best I can do…

        As for the emulator, I will bookmark it. It’s the first step in building the user base. I don’t think that eliminates the training issues entirely (can’t do 1000 node scalability test for $5 like on AWS) but it’s a good start.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: