With increasing attention, continuing outages and new funding of $15 million, Twitter is providing more insight into its operations. In a recent blog post, the company shared the technical problems it is facing and how it plans to make Twitter more stable.
Twitter is, fundamentally, a messaging system. Twitter was not architected as a messaging system, however. For expediency's sake, Twitter was built with technologies and practices that are more appropriate to a content management system. Over the last year and a half we've tried to make our system behave like a messaging system as much as possible, but that's introduced a great deal of complexity and unpredictability. When we're in crisis mode, adding more instrumentation to help us navigate the web of interdependencies in our current architecture is often our primary recourse. This is, clearly, not optimal.
Our direction going forward is to replace our existing system, component-by-component, with parts that are designed from the ground up to meet the requirements that have emerged as Twitter has grown. First and foremost amongst those requirements is stability. We're planning for a gradual transition; our existing system will be maintained while new parts are built, and old parts swapped out for new as they're completed. The alternative - scrapping everything for "the big rewrite" - is untenable, particularly given our small (but growing!) engineering and operations team.
Providing this kind of disclosure will go a long way toward appeasing the frustrated fans, as long as the dialogue continues...and Twitter performance shows some ongoing improvements. For a translation of Twitter's outage explanations, see this blog by my colleague Charles Cooper.
Dave McClure is becoming the Web version of a streetwise ghetto talker. In his latest post, Dave makes some good points, amid the street talk, in his attempt to define Web 3.0 and identify the winners in the race to colonize the Web.
In his day job, Dave is an investor and adviser to Web start-ups, and the more mild-mannered conference chair for Graphing Social Patterns, and a co-chair of the Web 2.0 Expo, which takes place April 22 through 25 in San Francisco.
I agree with Dave that the entities managing personal profile infrastructure, such as user IDs, social graphs, and online payment systems, will have a major advantage in colonizing and monetizing the Web. In his post, Dave wrote:
Because the Future of the Web belongs to whomever controls Search, Content, & these 3 core infrastructure components:
1. User Logins & Passwords
2. Friend Lists / Address Books
3. Payment Systems
Messaging systems (email, IM, and SMS / mobile phones) are the largest aggregations of user logins. They also have implicit social graph data & targeted friend lists buried in their data stores, but they will take a little mining to get to. The#1 and #2 players in messaging are Yahoo & Microsoft, with Google & AOL duking it out for #3 (note: Gmail is growing a lot faster than AOL). Also, if you consider messages on social networking systems, Facebook & MySpace are also significant players. These two have advantages over the others, since they have already built out Friend Lists, News Feeds, Social Apps, & other viral mechanisms in a way that allows amazingly fast & efficient (if spammy) distribution of content. They will get better at it, but they still need to discover better ways to monetize, and currently Google is doing that best via search, and ecommerce systems like eBay & Amazon are doing it well via traditional online shopping & online wallets (PayPal, Amazon 1-Click). Apple is also doing a pretty good job via iTunes of collecting & storing payment info for millions of users.
Dave asserts that Microsoft (with Yahoo) and Google (with AOL?) are best positioned to capitalize on his version of Web 3.0. Facebook, which has many of the pieces, as well as News Corp. (MySpace.com) and Apple will be factors.
Young Man Intrigued by the Flight of a Non-Euclidean Fly, 1942-47
(Credit: Max Ernst)As I wrote in a post about the proposed Microsoft-Yahoo union, one of the key elements of the deal is the combined reach of users of e-mail, instant messaging and other communications services.
Having contacts and related profile data provides a hub--the personal profile infrastructure--for building out a social network that is sticky, precision-targeted and monetizable at planetary scale. Facebook and MySpace are trying to build it from the ground up, adding richer communications services to their social graphs. Microsoft and Google are focused on engineering the social into their existing large-scale communications services.
They are all driving toward a social and semantically rich Web. It doesn't need a name, but when it arrives it will be the third major phase of the World Wide Web, a web of relationships and meaning, not just pages and data.
(Credit:
Wikipedia)
In a research paper from June 2007, titled "Worldwide Buzz: Planetary-Scale Views on an Instant-Messaging Network (PDF)," Eric Horvitz of Microsoft Research and Jure Leskovec of Carnegie Mellon University analyzed 30 billion conversations among 240 million people using Microsoft Instant Messenger in June 2006. It turned out that the average path length, or degree of separation, among the anonymized users probed was 6.6.
Six degrees of separation posits that a person is a step away from people they know and two steps distant from people known by the people they know--thus the magic number six.
Following is a more in-depth explanation of the phenomenon from an updated version of the research (PDF) posted on arXiv.org:
We present a study of anonymized data capturing a month of high-level communication activities within the whole of the Microsoft Messenger instant-messaging system. We examine characteristics and patterns that emerge from the collective dynamics of large numbers of people, rather than the actions and characteristics of individuals. The dataset contains summary properties of 30 billion conversations among 240 million people. From the data, we construct a communication graph with 180 million nodes and 1.3 billion undirected edges, creating the largest social network constructed and analyzed to date. We report on multiple aspects of the dataset and synthesized graph. We find that the graph is well-connected and robust to node removal. We investigate on a planetary-scale the oft-cited report that people are separated by "six degrees of separation" and find that the average path length among Messenger users is 6.6. We also find that people tend to communicate more with each other when they have similar age, language, and location, and that cross-gender conversations are both more frequent and of longer duration than conversations with the same gender.
Via Roland Piquepaille on ZDNet
See also Nature, "Six Degrees of Messaging"
The social Web is spawning more than millions of widgets, applications, and people connections. It is also has its own themed conference. Graphing Social Patterns got under way today in San Diego with a keynote by Charlene Li of Forrester Research on the future of social networks.
In the future social networks will be like air, Lee said. "No matter what you do, your social networks will be there. The social graph and your identity will be at your fingertips."
She predicted that by 2013 social networks will be open and ubiquitous. Reaching that plateau won't be a technology issue, she noted, but developing a level of trust among users, platforms, and marketers will be the major challenge.
Li broke down social "air" into four components: A user profile; the expressed relationships in a single social graph; the shared activities in a social context; and a business model in which social influence defines marketing value.
Forrester's Charlene Li
(Credit: Dan Farber)Universal identity will be based on your e-mail addresses or mobile phone number, and identity will be federated by a few large identity banks, like Yahoo, Google and Microsoft, she predicted.
One key issue today is that social graphs--the set of relationships in the digital the mimic those in the analog world--are walled off from each other. The walled garden approach is a source of competitive advantage for their owners, but some efforts are underway, such as Google's Open Social and Social Graph APIs and DataPortability.org, to break the lock-in. (I am moderating a panel on the topic later today at the event on privacy and data portability.)
With a more open social graph, new kinds of applications will be enabled. Li gave an example of being able to see what friends have written reviews on Amazon by keying off her e-mail address. Search engines could deliver query results based on what friends find relevant. Users could compare the performance of stock portfolios.
On the business model, Li said that each person will have their own personal CPM (cost-per-thousand impressions), which will be highly influenced by their social network. For example, an individual who is authoritative on a specific topic provides an "endorsed" value, which is enhanced by the quality of their social network, which is relative to the trust conferred by the social network to the individual. We already seen a first take on the personal CPM with Facebook's aborted Beacon service.
Li offer a few recommendations to the Graphing Social Patterns audience of developers, investors and industry watchers:
• Create linkages between services based on individually controlled identity federation.
• Compete on the most compelling social experience, not on lock-in.
• Develop social apps that have meaning, that are more utilitarian. "If you have to explain why Facebook is interesting, it's not going to become more mainstream," Lee said.
• Integrate social graphs into existing activities.
• Design business models that reflect the value created by people's social networks.
I wouldn't argue against social networking becoming deeply embedded at the core of the Web within five years, with more open social graphs meandering through the network looking for connections and dynamically modifying behaviors. Whether the Web gets better at understanding the calculus of trust than in the real world remains to be seen.
- prev
- 1
- next
