Web 2.0 and the Enterprise discussion

As announced a couple of weeks ago, I had the pleasure to moderate a panel organized by the TiE Software SIG, at the invitation of Prashant Shah, Nimish Mehta and Rajeev Motwani (thanks again guys). Because the topic was “Web 2.0 and the Enterprise”, I had chosen to invite three accomplished bloggers Jeff Nolan, Charlene Li and Ross Mayfield. Because of their experience of having been on “both sides”, I thought that they could bring a uniquer perspective to the issue. I took a few notes during the event which I am going to share here, and will point to other posts related to the session.

About 150 people joined us for the session, which made the TiE Conference Center fill pretty full. After a yummy Indian dinner (TiE serves the best food out of all event groups I attend), I introduced the topic by explaining that our goal was to introduce a few basics about Web 2.0 and then spend time chatting about why and how enterprise architectures would be impacted - with a goal to bring the discussion way beyond “We use blogs and wilkis” - not that there is anything wrong with those. I also asked the audience to join the conversation as soon as they had something to contribute - and boy did they do it. This was actually as close to an “un-panel” as I have seen a discussion get, and I really enjoyed it - even if it seemed a bit out of control sometimes. I will actually blog about this first “un-panel” experience separately.

Even though I did not want to spend too much time on it, I offered three characteristics of Web 2.0 that were relevant to the session:

• Rich User Interface and User Experience, where simplicity and appealing design are definitely prominent
• Architecture of participation, where users get involve in producing, commenting, rating, reviewing,...
• Vertical applications and mash-ups, where functionality is segmented into specific applications that serve a limited purpose, is made available to other applications through web services, and can be remixed into mashups

I also mentioned a few key enabling technology such as LAMP, RSS and Ajax.

Jeff mentioned the power of scriptable clients that are synchronized with back-ends developed in real-time, and the blurring of consumer and enterprise stacks into a common infrastructure - with two sets of naming conventions for the moment. Charlene mentioned the bottom up adoption of Web 2.0 applications behind the back of IT, which is fearing a loss of control - similar to the one they showed when instant messaging and browsers were introduced ten years ago. She insisted however that by giving up on control and letting people collaborate with limited/no constraint, collaboration/efficiency/problem solving happened. Ross challenged my “alphabet soup” and insisted on the fact that Web 2.0 was about social computing, establishing new social values and interactions and creating an empowerment of the group. He added that one of the main drivers of adopting social computing is to get information which is stuck in individual email databases, and people’s head, into widely accessible archives.

We then went into a mode of interacting with the audience *a lot*, letting most often attendees drive the topic to be discussed. This made the whole thing extremelly lively and fun, but a tad disorganized. A number of interesting points were discussed, that I am going to try and synthesize/comment on:

• There is a clear tension between legacy IT which is based on a command/control structure, and a need to minimize/manage risk, and social computing which is rooted in open communication, trust, and participation. There were a lot of discussions around the risk surrounding hosted applications (even though the success of Salesforce.com proves that large companies actually adopt it even for highly sensitive data) and the needs for proper security infrastructure. Besides, or on top of, the necessary layers of defense to keep the bad guys out and let the good guys in, a key piece of the puzzle is a portable identity and reputation system that can be used internally to establish the authority of a contributor - and there seems to be an agreement around the fact that the reputation of a person will actually be an aggregate of several different reputation implementations/incarnations (e.g. Technorati inbound links, linkedin profile, hits on google, eBay rating, etc.).
• Because of the wide accessibility of Web 2.0 applications, they are deployed by small groups of users in the enterprise - around specific projects. IT will at some point hear about these open source applications that have been deployed somewhere internally, or these hosted services that are being used “behind” their back, and will either try and shut them down or deploy them more agressively. An example mentioned by Ross was DKW (Dresdner Kleiner Wasserstein), that deployed a bunch of SocialText appliances to run wikis inside the firewall as a way to get people, and information, off of email.
• SOAs and Web Services have been around for a long time in the enterprise, with limited deployments and success to date - which led people to question the potential of Web 2.0 services based on similar technology. The panel’s take on this was that standards like RSS were reasonably simple to implement and that a lot of information unlocking and collaboration could be achieved as a “low hanging fruit”. There are also a number of micro-formats and structured blogging initiatives under development, which will hopefully work for both the enterprise and the consumer market.
• An interesting question came almost at the beginning of the event (from someone from Oracle): Will there be a new class of Web 2.0 enterprise software players that will eventually take over the current class of players ? My answer was that established players were not sleeping away as new companies and business models were developed, and therefore the most likely scenario was one of M&A campaigns (like Oracle's rumored roll-up strategy of open source startups) seeing these newcomers acquired as they reach scale.
• Ross shared his strategy for building an enterprise software startup: start “on the side” an open source project of value in a business context, attract a community of users and eventually developers that will embrace and contribute to the product, offer support and services around it, and if relevant develop commercial extensions, an appliance offering, distribution and hosting partnerships, etc. on top of it. SugarCRM was mentioned a couple of times during this discussion.
• There was a very pertinent question about what business models Web 2.0 was enabling, as compared to Web 1.0 that introduced Amazon, eBay, Google, etc. Our answer was once again centered around social computing, collaboration and the opportunity to “have a voice” in a way that was not as accessible. But that answer did not seem to strike a chord amongst participants who are not into blogging.
• There was one question about non-collaboration infrastructure, i.e more standard transactional applications and services. Unfortunately we did not dig into as much as I would have wanted to, but yes there are also changes happening in that area as well. I mentioned the fact that Tibco, the largest independent EAI player, has jumped on the bandwagon and has released an Ajax toolkit which I am told is pretty decent. It is a natural evolution for these guys to add a Web 2.0 layer to their stack. There are then companies building similar layers “from scratch” - not having the issue of dealing with legacy code and design.

There was still a lot to be discussed when we brought the session to an end because we had ran out of time. I would have wanted to talk more about transactional application, enterprise mashups, the evolution of BPM, etc. But it sounds like “Enterprise Collaboration 2.0” was big enough a topic to focus on. I have also suggested to Rajeev to organize a new session around the same topic in 12 months time, in order to figure out where the industry has moved.

If you are blogging the event, or posting photos/videos, please use the tag web20enterprise.

More:

• Jeff Nolan blogged his thoughts about the session. Very interesting comments.
• Zoli Erdos did a fantastic job liveblogging the event, which was more than challenging.

Tags: web20enterprise

Comments

[Kintan Brahmbhatt]

Thanks Jeff.
I like your point about the tension between the command/control structure of legacy IT and the open/participation-oriented nature of Web 2.0 applications. But looking at it from the other side, what are the pros of an "enterprise" setting that would make it a better case for an application to be deployed as an enterprise application, and not as a consumer application ?

For example: Active Directory and the group policies within the enterprise can lead to a lot of potential bvenefits. What's your opinion?

Kintan

[dahowlett]

As I suspected Jeff - but I expect to have a couple of cases on wich I can go public in the next week. How 'good' they'll be remains to be seen. Keep fingers crossed.

[Tony Karrer, Ph.D.]

Great stuff! We are looking at the use of Web 2.0 in the context of eLearning. What I'm finding (and everyone must be) is that there is considerable complexity in trying to decide on appropriate components and interfaces.

One of the problems is that the terms and design patterns don't seem to be established, e.g., add-in, component, badge, widget, etc. And how the different components get integrated and then how you interface with them also seems to be poorly defined.

Or am I just missing some good definitions of these?

Tony

[James]

How come folks who talk about the enterprise at conferences think it is OK to have analysts stand in as proxies when their view of the enterprise may not be entirely accurate.

Wouldn't it be better if you got a real employee from a Fortune 100 enterprise to participate on future panels...