ReadWrite Real-Time Web Summit and a forthcoming research report on the topic, we've now had extended conversations with nearly 50 companies in this space. The breadth of offerings, technologies and strategies is amazing.The phrase "real-time web" may make you think about Twitter, Facebook, or perhaps real-time stock market trading, but there are actually hundreds of companies all around the world working on building and leveraging different types of real-time delivery of data online. In preparation for this week's
We offer below one way to think about this broad market. We hope it's useful and interesting.
After talking to all these companies, one of the ways we decided that diversity could be explained is with a chart illustrating the continuum of real-time companies and use cases ranging from those that facilitate human-to-human communication to those that facilitate machine-to-machine communication. In between these end points there are services that mediate communication between humans through machine analysis or social objects and other services that do machine-to-human communication. What do you think, is this a fair way to describe the real-time web market?
On the far left we've got services that facilitate human-to-human communication, like Twitter itself, Instant Messaging and a new service called Olark (our review). Olark integrates with your existing instant messenger client to facilitate real-time chat with visitors to a particular page of your website. It's document-centric (another potential axis of analysis) but the value proposition is in direct communication between people.
On the near left we've got services that derive their value through human-to-machine/object-to-human communication. These forms of communication are particularly mediated by technology (though obviously in absolute terms even the human to human examples here are). For example: Aardvark is a service that cross references what you say you know about, what others say you know about, your network of contact networks, everyone's present availability and past performance in answering questions from Aardvark users. The result is a "real-time web of people" accessible through a very intelligent bot that delivers any question you have to the most qualified person socially near you who is available at that moment.
On the near right you'll find machine-to-human systems. These are like real-time robots who do work for you and then notify you immediately when something you're interested in has changed. Notify.me is a white-label real-time alert service (among other things) where the value comes from the machine communicating things to you.
On the far right are machine-to-machine systems. The two examples we have here are entirely theoretical so far. Both come from PubSubhubbub co-creator Brad Fitzpatrick of Google. Fitzpatrick says he imagines a future where PubSubhubbub or some other effective real-time communications protocol is used by services that want to subscribe to stay synced with our social media profiles made available through the WebFinger protocol. He also envisions a future where the entire web may be Hubbub-enabled, allowing Google to simply subscribe for updates whenever they're available from webpages, instead of going out and pro-actively spidering the entire web to index it.
These kinds of machine-to-machine communication can have all kinds of benefits. Fitzpatrick told us, for example, that the early implementation of Pubsubhubbub for shared items in Google Reader helped aggregation services get updates far more efficiently than the old "is there anything new yet?" method of periodic polling. He says, "When we enabled Hubbub [real-time] subscription to Blogger and Google Reader shared items, FriendFeed's traffic to us dropped 85% and latency changed from minutes to seconds."
My theory is that the technologies on the left have traditionally not been a big focus of development but are in this current era being heavily built out. On the right side, however, the financial world and others have traditionally put a lot of resources into machine-to-machine communication - but in this real-time web era, this kind of communication is just beginning to be developed in a lightweight capacity.
The above is one way to understand the breadth of the real-time web; but there are many others. What do you think of this model? Where would you place your favorite real-time services on this continuum? What take-aways do you get from a visualization like this?
We hope you'll join us on Thursday to talk about questions like these. Either in person in Mountain View, California (register now before prices jump on the day of the event!) or right here on ReadWriteWeb, where we'll be live-streaming selected sessions compliments of Justin.tv.
Finally, check out our fabulous event sponsors below. Where would you put them on this continuum? There's all kinds of different ways to discuss and think about the real-time web, and the more clearly we can think about it the better we'll be able to take action and build the web of the future with this important new part of it well-utilized.
We'll see you on Thursday!