I think some of us are over thinking the Enterprise 2.0 and Enterprise Mashups. The key is to learn from history and apply the knowledge to the present. Jeff Nolan has nicely summarized some of the discussion around Enterprise 2.0. Rod Boothby has raised the question of mashups in the context of Enterprise 2.0 I find Vinnie Mrichandani's post to be the best description of issues we need to consider and resolve.
However, the idea of Enterprise 2.0 seems inside-out to me. People from the enterprise world used to large ERP-style applications and deployments thinking in traditional terms with a willingness to tweak the model to incorporate 2.0 from Web 2.0. This will not suffice. The square (pun intended) peg of Enterprise software will not fit into the round hole (no pun intended) of Web 2.0
A lesson in history
The IT departments and software vendors that had significant investments in client-server tried to simply adapt client-server to the web rather than moving to true N-tier architecture. Slapping a portal on top of client-server software helped SI's and vendors make some money but they were no competition for applications designed from the ground up for the web. A lot of companies are now trying to do the same by trying to 'service-orient' their existing applications. I find the tone of discussion on Enterprise 2.0 suffering from the same problem.
Principles of Enterprise 2.0
The real Enterprise 2.0 applications are the SaaS applications (On-demand) from the likes of Salesforce.com, NetSuite, Siebel On-Demand etc. Some key characteristics of Enterprise 2.0 applications will be SaaS (Software-as-a-Service) delivery model, SOA architecture, simple web services based integration, extensiblity and open-ness.
SaaS & Long-tail: I would argue that all Web 2.0 applications are hosting-capable if not entirely hosted. The benefits of the long tail come into effect only when you have large number of users and but for a few very large companies it is hard to see how this effect will play out if the application is not hosted.
SOA, Web Services & Mashups: The Web 2.0 mashups were enabled by open standards-based APIs with many of them using either web services or other standards such as ATOM. In the enterprise, you need similar capabilities to do end-user integration. And SOA & web services help you expose your internal applications for integration. Rod Boothby poses an interesting question on what will Enterprise 2.0 mashups look like. I strongly believe (that's what blogs are meant for, right?) that Enterprise 2.0 mashups will look just like Web2.0 mashups. Just as your internal web apps for email and procurement now look just like Yahoo! and Amazon, in time your internal Enterprise 2.0 apps will look like the Web2.0 apps. Here is an example- you are browsing a catalog in your procurement application. As you mouse over the price, a bubble pops up (AJAX style) to tell you whether this is within your purchasing authority. This is a mashup of procurement and financials.
Why is this so hard?
If Enterprise 2.0 looks, feels and behaves like Web2.0 then why is it so hard. It is not hard, its hard for the existing enterprise applications that will in five years be legacy just as client-server applications became legacy and mainframe apps before that.
Are we done?
No, we are not done with building out Enterprise 2.0. The CRM space has made a good beginning and there are several others in various stages of development. With a new class of Enterprise 2.0, there will be new problems. Rather than looking for how to morph Enterprise 1.0 (did we ever finish 1.0, btw) into Enterprise 2.0, the key is to look at the Enterprise 2.0 and fill gaps. Some of these gaps are in the areas that have been brought up by Jeff Nolan and others. These include:
- Integration 2.0: How will Enterprise 2.0 applications integrate with each other? Will it be a hosted integration model?
- Identity 2.0: How do you enable single sign-on in Enterprise 2.0? Again rather than thinking about simply connecting your existing LDAP to Enterprise 2.0, look to online identity service providers.
- Customization 2.0: (I am as annoyed with having to say 2.0 after every term that applies to this new class of apps but couldn't resist it.) The question is how do I customize the new applications for specific verticals? Again look for the emergence of experts from verticals and regions. A Chinese company that customizes your vanilla Financials system to deal with national regulations. And then yet another company customizing that offering for clothing manufacturers to account for specialized business processes to deal with QR (Quantitative Restrictions) on clothing imports. And so on and so forth. Essentially, you end up with a network of service providers rather than customization by internal IT.
There are other issues of service-level monitoring and guarantees, analytics, bulk upload, security etc. Rather than trying to solve all these problems a priori before we build Enterprise 2.0, these concerns will get sorted out as we build out the services for the 2.0 world. Those of us who fall into the not good enough trap need to go back and re-read Innovator's Dilemma. The barely good enough for some will, in time, with iterative innovation become good enough for most.
I think this is an interesting discussion and I look forward to more exchanges and posts on this topic.
Update: FASTForward Blog's Joe McKendrick has picked up and extensively commented on this post at Fitting the Enterprise 2.0 Square Peg into the Web 2.0 Round Hole
Update: ZDNet's recent post entitled Facebook for Enterprise = Enterprise appears to agree with my viewpoint.
Tags: SAP Web2.0 SOA Enterprise2.0 wisezen