Attending VMWare's VMWorld conference this week in San Francisco was an interesting experience - we got to witness a lively ecosystem, a company with a strategy and really passionate - if not even cultishly loyal - user base. But where there is light there is also shadow - and we will describe both in this post.
Monday takeaways
VMWorld kicked off with the usual CEO keynote and Pat Gelsinger did a good job walking through the status and upcoming plans, in one of the most methodological and structured keynotes I have ever seen. Guess the chip engineer background does make a contribution to presentation style here.
Not surprisingly VMware keeps forging ahead with its server virtualization work - making vSphere 5.5 generally available in Q3 2013 - which amongst many other features - gives customers a better handle to deploy Hadoop atop of VMware (courtesy of project Serengeti). And not surprisingly the support for enterprise applications has been increased - SAP and Oracle in the forefront - with a doubling of logical cores to 320, supported memory to 4 TB and a 50% increase in virtual CPUs to 3096. This will allow customer to move even more ambitions payloads to be virtualization.
So far so good - steady state on constant improvement - the main thrust of VMware's R&D though will go into the creation of the Software defined data Center (SDDC) - by itself not too much of a surprise - given the Nicira acquisition and the need for a compelling overall data center strategy. And this of course triggers more complex and extensive management needs - as much more than the conventional server load is being virtualized. The new products are VMware NSX and VMware Virtual SAN, the former is expected to be GA in Q4 2013, the latter go to public beta in Q3 2013.
Noticeable - especially in comparison to last year - was the absence of the so called end user products which had prominent keynote time last year. It looks to me as if VMware may have tried to buy time here - or realized the VMWorld audience is not the decision maker for this product category.
In contrary to end user products - Gelsinger / VMware did talk about hybrid cloud, brought Bill Fathers on stage - and Bill talked about the data center plans - which are now going to be 5 in North America (2 existing and one planned by VMware, and 2 in partnership - surprise, surprise - with Savvis). Moreover VMware offers a connect service with a product called Direct Connect - the enablement on top is for disaster recovery and desktop as a service (DAAS). A good start - but much more is needed.
And last but not least VMware re-iterated it's commitment to Pivotal and Cloud Foundry, both companies will work to deliver a commercially supported Pivotal CF PaaS - that can run both on vSphere and vCloud hybrid service.
Feel free to check out the Storify tweet collection on the keynote here.
Analyst & Press meeting takeaways
This turned out the be one of the better press and analyst meetings out there, with plenty of time for presentation and questions to all the key executives of VMware, kudos to the AR team.
My key takeaways amongst a lot of good learning and information were
- It was good to hear from VMware's strategist Shekar Ayyar that the investment is flowing into the three priorities of SDDC, hybrid cloud and end user computing. Overall VMware sees an addressable market of US$ 50B in 2016.
- The strong dependence of VMware on partners is both an opportunity and a challenge - lower sales costs, but the ecosystem needs to be treated well - and it apparently feels treated well.
- The relationship of VMware with OpenStack seems to get more and better clarified, with Raghu Raghuram clarifying interfacing strategies on a number of layers between the two offerings.
- When I asked Bill Fathers on one of the strongest VMware hybrid cloud sales arguments - that VMware knows the loads in the private cloud better than anyone - and that they should therefore become the adviser on what loads to move to the public cloud when - I got a lot of agreement, but that this - for now - is a professional services opportunity only.
- The future of Socialcast remains in question - but Sanjay Poonen eased concerns here - but VMware still needs to flesh out more. To be fair - it was Poonen's 6th day with VMware and for that he did a formidable job.
- And finally I had the chance to ask Pat Gelsinger where he sees the growth for the likely reduction in compute virtualization revenue (no pushback here!) - and the answer was that in the short term it will be management, NSX and end user computing, then storage and then hybrid cloud.
- Lastly I found all VMware executives remarkably friendly, open and almost eager to answer even critical questions - not all analyst meetings are like that, other vendors take note.
Feel free to check out the Storify tweet collection on the keynote here.
Tuesday Takeaways
The mornings keynote was different than in previous years - when it used to take a product and show me the features approach. So this year VMware tried to pitch new products in the real world scenario of the IT side pitching SDDC to the end user (ironically being played by Carl Eschenbach, VMWare's COO). Personally I liked the idea and execution very much - but was surprised that most of the audience did not - which reminded me that most of the audience lives and breathes very close to the compute virtualization. Not much pitching of a business case needed here - it's known and it's been done.
Not surprisingly the press spin machine slowed down - with only three press releases - one on the Horizon product side that can now deploy desktops to vcloud hybrid service, partners can offer desktop as a service and there is extended smartphone support with new VMware ready smartphones.
Buried in the VMware Cloud Management press release was the little fact that this product was the strongest growing product area of VMware - not surprising - but a good and key datapoint.
And lastly a VMware and Cisco joined customer success press release - well not a surprise here again.
There is money in virtualization
Nothing validates the ecosystem better than a visit of the showfloor. And with VMworld being in Moscone - the comparisons to other shows that take place there - like OpenWorld, Dreamforce, Build etc is easy.
So what I am looking for are usually two things - interest by the attendance and marketing spend by the exhibitors... needless to say that the attendees were all over the place, but even more surprising was the marketing spend, that was way north of what you see even at much larger events like OpenWorld and Dreamforce.
Usually I look at the ridgeline of the booths - and you usually see a Matterhorn like profile - with the organizer's booths being at the center. VMWorld however, had an Ayers Rock like profile - with a lot of multi storey booths reaching out far to the edges of Moscone East and only one row each on either side of the exhibition floor of poor man booths.
Equally interesting - apart from the hypervisor competitors - everybody was there - the hardware vendors, the networking vendors, the data center management vendors, and the software vendors in the ecosystem. Yes, there is a lot of money in virtualization.
Reality Check
As analysts we sometimes need to be careful not to live to much in the bubble that is created by Silicon Valley and by the constant briefings with cutting edge content by vendors. Many fellow analysts were wondering why to even attend VMWorld... as the VMware compute hypervisor revenue stream gets commoditized - and the move to public cloud would end VMware's stronghold on the corporate data center.
But chatting informally with attendees, mostly while waiting in the queues catching a bus, pick a table at the attendee lunches, chat with them at appreciation events, poll the after seeing a demo on the show floor... gives a sense for what is real in the customer base. And to my surprise from the over two dozen companies I spoke with - none had concrete plans to move to public cloud. Didn't even have applications running in the public cloud. The most I got was a beta of Office 365...
And at the risk of running into a self fulfilling prophecy here - of course the VMWorld attendees are there, because they do have a significant compute load in their data center. But I was honestly surprised that there was so little public cloud reality at these randomly polled enterprises.
VMware's opportunities
As previously stated, VMware may have more time for the transformation than previously expected. But the main asset I see for VMware is, that the company has the deepest understanding of how onsite compute resources are used. It really comes back to that knowledge and finding a way to productize this going forward - in connection with the threat the public cloud poses for the company.
More down the road - if VMware's gamble on SDDC starts to pay off - which is a potentially big IF - it has the opportunity to really build differentiating end user computing scenarios - think of e.g. secure social networking, the machine to social relationship, the hybrid cloud load balancing etc.
The other good news is, that VMware has a very loyal user base - I have not seen such a loyal and positive user base at a user conference since a long, long time. Sapphire conferences in the middle 90ies maybe.
VMware's challenges
Based on this, there maybe much more future on running hypervisors for compute in the corporate data center than previously expected. VMware's challenge is the pressure on these revenue streams. The good news is that the executive team is aware of these, acknowledges them and has now a bold strategy in place to move the corporate data center to the software defined data center.
But to achieve this the company needs to master an aggressive product development agenda, mastering not only networking, but also storage virtualization. But ultimately the question will not be if VMware can deliver, but if VMware can deliver enough critical functionality in time to be a relevant player and keep off larger competitors like e.g. Cisco and Juniper for the networking virtualization. The acquired assets of Nicira help a lot here - but there is more work to do.
Finally VMware needs to change the way it goes to market. Great to have a cultishly loyal user base and a very successful channel - but both will not win the battle for the SDDC. The user base is to far removed from the CxO level decision making about SDDC and the channel will have different options to play this game. And VMware's competitors have larger direct sales forces with entrenched relationships to the CIO.
MyPOV
A very interesting VMWorld where VMware laid down an ambitions strategy to get the control of the SDDC. The company will have to change the way it sells to succeed with that strategy, which may change this show going forward - less geeks and more suites. Interesting marketing challenge.
And it looks like VMware days are less counted than I thought, the threat of customer moving load to the public cloud is much less imminent than I suspected. So it will be less the the port to the public cloud, than the replacement of on premaise loads by cloud loads through a vendor switch, like ridding itself of an on premise CRM system and switching to salesforce.com, which poses a challenge to the VMware market.
Finally a very ambitions product development agenda, but you can't fault VMware for that - and it's too early to call the state of progress on that. In the meantime the company deserves credit for going for it. We will have to check in approximately 6 months from now on progress of these plans.
