I am at the Vortex conference today and courtesy of their wireless network will be blogging it.
Please check back for updates as I post them. Please note that in the Q&A sections I am not transcribing directly, I'm paraphrasing and offering my own comments on what I'm hearing. Nothing in this blog post should be taken as a direct quote word-for-word.
Geoffrey Moore Keynote, Orchestrating the Stack:
- the stack consists of business layers (services and applications), compute process layer (app platforms), and compute engines (hardware)
- each layer in the stack has gorillas and chimps. Gorillas are dominant vendors with high switching costs, chimps are significant vendors who are powerful but are not dominant vendors
- of the 13 categories, 7 have gorillas while the remainder do not.
- "openings and instabilities" are abundant, meaning that investment opportunities are prolific
Service oriented architecture, the next big thing
- current stack represents internet-enabled client/server dependent on large monolithic structures
- current leading technologies push the internet itself into the position of being the application bus
- with the internet as the bus, all of the stack elements become modularized as a shared service
- new stack will not happen overnight and it will privilege specific layers
Strategies for exploiting the new enterprise stack
- disrupter: exploit architectural advantage to exploit niches. Target gorilla status in emerging layers. Leverage Innovator's Dilemma
- chimp strategy: neutralize gorillas using niche strategy and dominate specific markets. Exploit opening and instabilities to secure a gorilla position in an emerging layer, creating a defensible position
- gorilla strategy: marginalize chimps and monkeys. Transform layer into platform for services. Marginalize layers not entered.
The agenda for the conference is to talk with 8 of the vendors who claim they are going to be the platform vendor of choice for the new enterprise stack, along with chimps and disrupters, investors, and customers.
other comments of note
- the U.S. is the worldwide leader in technology because we guess first and ask questions later. I agree with this, my own observation from the German side of SAP is there is too much consensus building and too little risk taking.
Moore's take on:
IBM
- gorilla in a number of areas, but ultimately their model thrives on complexity, the more complexity there is the better it is for them.
- want to control many layers in the stack and glue them together with their services business
SAP
- gorilla in business applications, wants to be platform for services
- won't support multiple application server strategies, countering with Netweaver. Vendor specific architectures.
Microsoft
- volume operations model starting from the edge and working inward. Ubiquity and then consolidate complex systems that involve multiple nodes
- keep competitors out by expanding footpring of proprietary technology
Oracle
- looks like part of IBM in that it wants to vertically integrate layers above the database. Compete against IBM using the same weapons.
- Commoditize everything below the database
EMC
- dark horse in the game.
- leverage data center experience to enter systems management software and non-relational content management
- position information rather than applications as focus of systems management
Cisco
- integrate the network to transform it into a platform for web services
- enter systems management beginning with security
- annex enterprise storage
- use visionary consulting to influence the emerging architecture
Sun
- be the SOA disruption accelerator
- evangelize next gen technologies like Java, SunTone and N1 that facilitate SOA
- architecture-focused consulting
- absorb lower cost infrastructures including x86 and Linux
HP
- target systems management layer as the critical platform
- minimize, not maximize, the footprint to leave room for partners
- reinforce the platform through other major initiatives like adaptive computing and outsourcing services
- the anti-IBM strategy, not seeking to control as many layers in the stack as possible
Summary
- enterprise stack is strongly entrenched, if we don't do anything nothing will happen for a decade
- the future state is that SOA is disruptive and characterized by a new set of platforms
- questions that form the theme of Vortex are around what layers are strategic, what platforms make the most sense, and which vendors are most credible.
Questions:
Q: what is the most dangerous assumption to make?
A: IT will end up looking like it always have, different players in the same seats
Q: How did Intuit deliver SOA in Quicken while the enterprise vendors are still struggling with it
A: If God had an installed base he wouldn't have been able to do it in 7 days
Jeff's Note: check out Quickbase for small business
Q: why aren't telcos in the model
A: it's probably a flaw in the model, we will hear from telcos throughout the sessions
Mike Volpi from Cisco is on the stage now, Moore is talking with him
Q: how strong is Cisco's position
A: we feel comfortable where we are, but everything is tenuous in this market. We have done well developing trust with our customers, we are a partner not just a vendor.
Q: what is Cisco's strategy to Huawei?
A: whether it's Huawei or anyone else, we do a better job of creating value on top of hardware. Innovation keeps us ahead of the pack, but it's consulting that enables us to guide customers to achieving more value out of commodity hardware (not clear on if he was referring to Cisco hardware as being commodity... doubt it, but it's happening anyways)
Q: what value are you adding above the router?
A: we started with a network who's sole job was to move data around, it's a relatively dumb mechanism. Where we are today is building a network that enables customers and ISV's to build applications (in essence this is what Moore was referring to earlier when he said that the internet is the new application bus).
Q: are app vendors building in Cisco features into their apps?
A: ISVs are building in telephony and security
Q: communication and collaboration are at least as important as data schemas and transaction mechanisms (a statement rather than a question)
A: in the past these things were different application areas, now they are intertwined
Q: what kinds of things are bleeding edge this year or next?
A: 2 areas that we see are integrated communications (extensions of VoIP)... but he never said the 2nd area.
Q: how would I know when I hit the wall
A: competitive pressures are a sure sign, either through better cost structures or better customer functions. On security it is as simple as something bad happen.
Q: How does enterprise modility fit in?
A: important area. We want to enable full spectrum of functionality to mobile devices, such as voice and video to whatever device.
Q: what happens when you cross networks
A: even with today's technologies dual mode phones are available. WiMax is more universal.
Q: who are other gorillas that Cisco wants to align with?
A: platforms do exist but they are tectonic in that they move in their own direction irrespective of what their partners want. Some of the gorillas are technology providers while others are services. ** he answered this without answering it
Q: is Microsoft a competitor
A: not in routers, but they do want to keep the network as a commodity with little value add. It's like having a car with all kinds of fancy technology but still on crummy roads
The audience is open for questions, but I didn't quite get the first one. Volpi is talking about how the network will do more processing in a grid fashion... my take on this is simply that they want the router to do more processing.
The next question, after a long intro, is how does the world look different when users are still dependent upon vendors like Cisco to build functions into the hardware. Volpi says that competition ensures that customers get what they want... yeah, I've heard that before!
Another question, if Cisco is building 'smart roads', how do the on and offramps look and does Cisco see itself as the provider of the access points or the roads? Volpi went on to talk about how U.S. businesses and consumers are getting a raw deal on broadband compared to other countries. He didn't answer the question.
Another question using the roads and vehicles metaphor, this time about the kind of traffic that is on the network - big rigs versus cars. Volpi answers that prioritization is already well developed in today's routers, there is no reason why routers can't do deeper analysis of the packet payload to do more complex prioritization and security analysis. Now he's talking about XML... Moore asks him directly about what companies Cisco is working with. Volpi answers that they are doing R&D with customers on the XML in routers opportunity.
couple more questions that I'm not fully understanding, plus I have to go to the bathroom so I'm hoping that this wraps up in a minute or two.
My comments:
This was a very good presentation with a follow up discussion with Volpi. Moore laid out a very reasonable analysis of the transition point that the industry is in, at least from a technology standpoint. I think his analysis of the situation with gorillas and strategies for dealing with them needs to be baked into every presentation that a startup gives to investors.
Volpi makes a strong case for Cisco in the emerging stack, however, outside of telephony and possibly using routers in a grid setting, it's not entirely clear how they intend to play with the other vendors. Indeed, Volpi makes a strong point that some platform vendors would just as soon see routers and networks remain rather dumb. All of the talk about competition sounds nice, but we all know that Cisco, like any gorilla, is not going to simply compete with upstarts on the basis of features and functions alone to deliver customer value.
One thing that I am really keen to hear talked about in the discussions today is how the business of software is changing as much as the technology itself.
Next up is a one-on-one with Jonathan Schwartz and a panel titled "Disruption in the Air"
okay, I diligently took notes on the interview and the panel and then lost my fucking wireless connection, which I didn't know had happened until I tried to publish. Everything I wrote from my last post forward was lost. Yeah, hosted systems are the way to go :(
I'll try to reconstitute what I wrote, but the net is that I wasn't impressed with the interview all that much, nor the panel so I may just skip it altogether. It is worth noting that the Cisco people at lunch were impressed with Schwartz so maybe I just missed the importance of what he was talking about.
Geoffrey Moore interviews Shane Robison of HP:
Q: (slide displays the enterprise stack and HP's emphasis) It doesn't appear that HP has a strong position in this slide?
A: HP is targeting not a slice but a spike through the entire stack, they are targeting management
Q: Why is HP focusing on this strand?
A: customers tell them that managing a growing and increasingly complex IT environment is their biggest challenge
Q: how does HP see itself in relation to IBM
A: a real, legitimate alternative to IBM. They are Avis while IBM is Hertz.
Q: What is HP's position on indemnifying Linux customers from the 280 patent claims against the OS?
A: Linux is not going away, we will indemnify our customers directly
Q: how much of HP's enterprise business is services led?
A: "more and more all the time"
Q: Will HP-UX go the way of mainframes?
A: hard to tell, there's a lot of stuff in Unix that still have a high degree of value
Q: what percentage of Openview's revenue line runs on HP gear versus others
A: most of it is on HP gear
So to paraphrase the rest of the interview, it's clear that HP sees itself as #2 next to IBM and #2 to Dell. It seems that they are okay being #2 to IBM but not okay being #2 to Dell. They will continue to compete for market share, they say that Dell is so vocal about it because HP is achieving some success against them.
In the enterprise, which is what I'm interested in, it is evident that they won't be a platform competitor. The focus on being a systems management competitor puts them more against CA and BMC, however the services lead in puts them in IBM's headlights. It is clear that HP does not want to be in Microsoft's headlights, so it was suggested that the partnership between M$oft's dynamic server management tools and OpenView will go deeper.
Next up is an interview with Chuck Phillips
apparently he goes by "Charles" now.
Q: how secure is Oracle in the database market
A: divide the database market into mainframe and non-mainframe markets. They do really well in non-mainframe markets.
Q: what about mysql
A: mysql has utility and oracle has responded with their own 2 processor limited version for $149 per user
Q: what about Microsoft
A: a real competitor, the fact that they have been after it for a decade indicates how tough the problem is.
Q: has Oracle modularized the database for SOA?
A: the database has evolved not because of SOA but because discreet function available as a service is a better way to offer it. Integration will never go away, but a standards-based approach to integration is a good way to go.
Q: what's the game plan for application servers
A: they believe they are number 1 in Unix and gaining in other markets. They also think that customers will use their app server if they want to do custom app work.
Q: what is the relationship between app servers and databases
A: the connection is a lot tighter than the independent vendors will have the market believe. So Oracle believes they will win in app servers by demonstrating that you get more bang for the buck by having both components in the same layer. This is a key position for Oracle in competing against BEA.
Q: the next question is about SAP Netweaver and Oracle
A: Oracle says that SAP is using Oracle as a smart fileserver and not taking maximum advantage of the database.
Q: the question about the PSFT takeover is asked, why?
A: because the multiplier for PSFT customers in Oracle's product line is huge, according to Phillips. The product lines will not be merged. Oracle is aggressively reaching out to customers in an attempt to establish comfort, and customers are now confident that PSFT being owned by a larger company is a good thing.
Q: question about transaction systems and relational databases
A: the important statement was about the vision that they have of the database being the filesystem. The benefit for enterprises is abstraction of where something is stored and how it is accessed. No longer need to worry about where things are stored. Microsoft is moving in the same direction.
audience question about processor-based licensing. Phillips responded with a rather realistic response about the price being what the price is and ultimately everything is negotiable so customers pay what they negotiate. As an aside, the question is really directed at grid and multicore processes (several processors in one package) and how licensing is impacted. It's a good question.
Audience question about Sun's $100 per user per year and the effect on the industry. Phillips said that he doesn't worry too much about Sun, more concerned about IBM and Microsoft.
Another question from the guy that's been asking questions all morning. Long winded narrative that precedes the question, the saving grace being that the guy asks good questions so I guess it's a style over substance issue.
question about mobile computing. Databases on mobile platforms are pretty weak, but when you consider that mobile device computing power has increased dramatically, is Oracle missing a big opportunity? Moore made a comment about two hot trends being embedded and mobile and Oracle is in neither. Phillips doesn't think they are missing, but did use the question to talk about broadcasting of data to mobile devices.
What's new from Oracle, what's the vision. Phillips says there is enough to do so the next big thing is really to simplify all of the stuff that's been happening for 30 years. Complexity is the problem that needs solving, according to Oracle.
The question about acquisitions comes up again, this time about what additional deals Oracle is looking at. Phillips said that they have multiple deals they are working on, and expressed surprise that "so many CEO's don't realize that the world has changed and you just can't respond to acquisition offers by saying you don't want to talk about it at any price". Good answer.
Panel discussion: Analysts - A Front Line Report
John Gantz, IDC
John Hagel, management consultant and author
Dave Passmore, The Burton Group
Interesting approach, the analysts are asked to treat the 8 companies as on Survivor and vote them off the island.
Dave Passmore: all of them are gorillas, they won't go away quietly. They have distribution capabilities, customers, technologies and partners. Stickiness is high in databases and low in commodities, however M$oft is a commodity vendor and certainly not going anywhere (how do you have a commodity and a monopoly at the same time?). Okay, so when pressed, John votes Sun off the island. IBM is well positioned across the board.
John Hagel: there are challenges that all of these companies will have to face, but have not yet faced yet. For all the talk of the new enterprise stack, nothing about the customers has been said and they are the ones who have to make the decisions. CIO's are not the sole decision makers within the executive suite, CFO's are increasingly making enterprise-wide decisions and most vendors don't have great relationships with CFOs (I disagree, most vendors I know go out of their way to build bridges to every C level executive, especially the ERP vendors). Hagel makes a good point about SOA being across multiple enterprises and not solely within the enterprise (somebody still has to pay for it). Hagel votes EMC off the island, puts Microsoft and IBM at the top of the survivor list.
John Gantz: The edge of the network (mobile) is where the interesting drivers for the architecture is coming from. In addition to all the normal criteria, the ability of a company to change shape, charismatic leadership, hidden poison pills, and finally a will to live, keep Sun from getting voted off the island. John questions the ability of Oracle and SAP to move up or down the stack. Agrees that Microsoft and IBM are at the top of the list of survivors.
When asked directly about Cisco, HP, SAP, and Oracle, Passmore gave Cisco a middling rating. Everybody already has a network infrastructure. Cisco has not yet convinced carriers that they are carrier class. Cisco is also vulnerable because they are increasingly viewed as commodity hardware, especially with Huawei coming on strong.
Hagel suggests that a big opportunity exists for connecting enterprises to other enterprises. Passmore says that Cisco won't do that because it would decimate their margins, implying that such a business is in fact a consulting business.
Hagel on SAP and Oracle, says that Oracle will have a challenge in collapsing the stack. Passmore calls them a 'one trick pony' who clings to the notion that everything goes in the database.
HP has the curse of the printer business, according to Hagel. Passmore doesn't know who HP is anymore, calls them a mish mash of a number different businesses trying to sell into the datacenter. Gantz calls them a survivor but a perennial #2.
Hagel puts SAP in the upper tier, although does call them a surprise player. Passmore includes the switching costs factor. Gantz calls them a good company but gives little credit as a "shape shifter". It's interesting to me that SAP just dosn't inspire strong opinions among this panel, don't know if that is a bad thing or a non-issue.
The question on dark horses is asked and Hagel calls Dell and Intuit dark horses, I have to agree (Quickbase rocks). Passmore calls Dell and Huawei dark horses, it's hard to imagine Huawei not eating Cisco's lunch in Asia. Gantz calls Novell a dark horse, also references Tata and eBay, agree.
Comments