On CouchSurfing, members upload profile information, pictures of themselves, and add friends, much like on other social networks. When you want to stay in a foreign country, you examine people from that country who have a couch ready for you.
So far, other than being single-purpose, CouchSurfing has normal social networking traits. Where it is special is with members rating each other after every stay, much like you'd rate sellers on eBay:
Do they have many friends? Are they active in their local CouchSurfing community? Do they look creepy? If they have no pictures of themselves, why not?!
In my opinion the most critical part of CouchSurfing security is the above safety feature. It comes in two forms, references and vouching. References are given by fellow members much as we described. Vouching is an extra measure added purely for reasons of security:
What do reviews by people who stayed with them before say about them? How do they review people they stayed with? How do people they stayed with review them?
Further, CouchSurfing provides with a verification system (optional). Members donate a bit of money which supports the site and the community. At the same time this provides CouchSurfing with your real name and address. They then mail you a (real) postcard with a verification code for you to enter back at the site, completing the process.
"When someone is vouched for, it signifies an elevated level of trust in the community. The only way to become vouched for is to be extremely trusted by someone who has been vouched for by three other members. You must know each other in the real world. If you are a trusted friend, the vouched member can vouch for you. Once you have been vouched for three times, you can in turn vouch for members you sincerely believe to be trustworthy."
This creates a lively, worldwide community where friendly tourists can feel safe, and travel the world among friends, while saving a buck. And, perhaps -- wink, wink -- make a friend.
In our industry, building trust has the added complication of sharing sensitive information, as well as dealing with paranoid people. Having run several trust-based operational and research groups during the years, I can attest that we mostly limit ourselves to members who already trust each other.
This creates a circle very difficult to break. Limiting the membership considerably, we create a closed club of sorts based on profession, industry, and usefulness to the group, with a difficulty for new people to enter and be accepted. How do we expand trust and accommodate new members when each addition detracts from established trust?
Online, how do you establish someone's identity even if you did trust them? You look for consistency and identifiable patterns, such as names (or nicknames), email addresses, and PGP keys. These could easily be manufactured, which brings us back to taking it on faith and historic experience.
"On the Internet men are men, women are men, and children are FBI agents."
Another form of online trust is the blacklist, or DNSBL, which contains IP addresses seen involved in malicious activities and can be filtered by users who want to avoid attacks or spam. Naturally, listing strategies differ between blacklists, and some do a better job than others. Mistakes happen, both by listing those who did nothing wrong (false positive), or not listing someone who did do harm (false negative).
In email, we have seen attempts, such as SPF to DomainsKeys, where people can agree on a certain distribution system or database where they can be listed and verified. Similar systems that failed include Website certification business (selling digital certificates) and so far, DNSSEC.
Trust is a key challenge for anyone who works online. Being a security professional who likes to connect people, I have the distinct pleasure and continued pain of trying to make large groups of people with competing interests (and a reason to be paranoid), cooperate. It is always a long process.
It is interesting that CouchSurfing, a social site from long before the Web 2.0 buzzword era, with friendly people who simply want to travel on the cheap side, does a better job of it than we do.
But if you think CouchSurfing has it easy -- not needing to share sensitive information -- think again. Members may not be sharing data on organized crime, but a 19-year-old girl who travels to another country has other types of worries, which is why CouchSurfing and the model it made work are so impressive. And it's cool.
The main difference between CouchSurfing and trust established in our industry, is that CouchSurfing's system is open and scalable. What can we learn from them?
Follow Gadi Evron on Twitter: http://twitter.com/gadievron
Gadi Evron is an independent security strategist based in Israel. Special to Dark Reading.