Google, ISPs, & Privacy

Two good, independently created, posts on this topic appeared yesterday:

Google and Privacy by Tim O'Reilly, on his O'Reilly Radar blog:

"...In short, it seems to me that Google is being held to a much higher standard than the rest of the world..."

Google and privacy by Matt Cutts (a Google employee), on his Gadgets, Google, and SEO blog.

"...First, I believe Google does more to protect our users’ privacy than any other major search engine. Second, I believe other companies such as ISPs have a superset of the data that Google has, plus they have verified payment/identity, plus they know which IP addresses you are on, even if you switch IP addresses..."

Both Tim and Matt make good points. I do believe that Google respects privacy more than most large companies, and I make moderate (though careful) use of a number of Google services. The ISP privacy risks are significant, whether you care more about commercial or government invasions of privacy. See the Wired blog Threat Level (formerly 27B Stroke 6)'s ISP Privacy Survey. Ironic isn't it, that AT&T (along with Microsoft) is opposing Google's acquisition of DoubleClick on anti-trust grounds?

Google's reach does give them great power and great liability ("with great power comes great responsibility"). I don't believe Google is evil today, and my privacy practices may seem paranoiac to some, but once the trust is breached or some database is hacked, the data is out there forever. I don't think I'm alone in not wanting to have a massive "digital dossier" on me accumulated in the first place without my knowledge, consent, and some control. So despite technology-enabled marketing and post-9/11 security theater, I take issue with Tim's assertion of the inevitability of "moving into a future where what we do, where we go, what we spend, what we pay attention to, will be mined constantly and by everyone."

OK, so what basic things can a savvy consumer do to minimize the accumulation of a digital dossier?

1. Practice diversification, just like with your investments, to avoid having all of your eggs in one basket. Use an email provider separate from your ISP. Don't put all of your online services into one account with a single provider. The telcos and cable companies love the triple-play (voice, data, video), but consider whether one company should know so much about you. The convenience may be tempting, but if you care about privacy, carefully weigh the risks and trade-offs. Change your IP address regularly (you can usually do this by briefly disconnecting from your ISP, then reconnecting).

2. Browsing: Block third-party cookies and delete unwanted cookies regularly. The only cookie I keep across sessions is for my primary bank because without it they require an extra step for login. If you regularly use multiple logins from the same provider, get used to using multiple browsers simultaneously; I often have three or four browsers running at a time. Don't provide your real identity information (full name, address, phone number, birthdate, credit card information, etc.) except to trusted entities, and then only when absolutely necessary and with great care. Never provide your social security number, unless required by law. Adapt readily available lists to block ad and malware servers that you consider offensive. I have mixed feelings about anonymizing proxies (Tor is a different story), as they simply shift the risk from your ISP to another (probably smaller) company.

3. Email: Use a provider besides your ISP. Make sure they offer encrypted POP/IMAP/SMTP/web, especially for the login. Keep multiple accounts for different purposes, preferably through multiple providers. Use variations of your email addresses. Use fully disposable email addresses (e.g., jetable.org, mailinator.com) for one-time or untrusted uses. Never reply to spam, phishing, or other scams. Set up your web and desktop email clients to not display images unless and until you say so. When the image is rendered, it is possible for a server to know that your email address is valid, and when you read the email.

4. Online backups: Store only encrypted files on third-party servers, unless it is public data.

5. Good security is essential to privacy: keep your OS up to date and operate a good NAT firewall with SPI that "stealths" all ports on the WAN side and ignores incoming pings. Be aware of (and block if necessary) any applications on your PC or LAN that needlessly "phone home". Keep critical information on your desktop and off of the web. Use good security practices to keep your local machine(s) safe (passwords, encryption, periodic cache and log cleaning, etc.)

6. Real world: Pay cash. Get an unlisted phone number and a PO Box. I'll leave it as an exercise for the reader to decide whether to get on the Do Not Call lists. If you're really paranoid, get a pay-as-you-go mobile phone; buy the phone and the minutes with cash and change the phone annually. Use Caller ID Blocking (*67). Avoid store loyalty cards and phone surveys. Vote with your dollars by patronizing establishments that have good privacy practices and don't ask for more information than necessary. Opt-out of the data sharing with your commercial accounts (read the fine print when the privacy policies come in the mail, and follow the directions; keep notes of what you've done). Opt out of direct marketing mailings. Some more detail is here (Chris Hoofnagle).

All of these things are easy and don't require great technical skill. In future posts, I'll be providing more detailed tips, and also exploring privacy and security issues with Flash, JavaScript, and Java.

Related Posts:

ISP Account Log File Analysis
ISPs Selling Clickstreams
Basic Browser Privacy
Geolocation
Are You Secure?

Just doing my part

We did not invent the algorithm.
The algorithm consistently finds Jesus.
The algorithm killed Jeeves.
The algorithm is banned in China.
The algorithm is from Jersey.
The algorithm constantly finds Jesus.
This is not the algorithm.
This is close.


[For clarity, see http://blag.xkcd.com/2007/04/19/billboards/]

Yo! Google Mac

I love it when transient stuff shows up in an RSS feed. Looks like Vishal p0wned the blog. Clicking-through the Google Reader link gets a 404, but a Google search yields a partial indexed result (click the image to see a larger view.):

USDA Exposes 63,000 Private Records

The NYT [via BB] reports today that:

The Social Security numbers of tens of thousands of people who received loans or other financial assistance from two Agriculture Department programs were disclosed for years in a publicly available database, raising concerns about identity theft and other privacy violations.

A later USDA press release provides the specifics:

The U.S. Department of Agriculture (USDA) has narrowed to 63,000 the number of people whose private identification information was accessible to the public on a government-wide website and who will therefore be eligible to register for free credit monitoring services.

The 63,000 people were awarded funds through the Farm Service Agency (FSA) or USDA Rural Development (RD). The FSA programs involved are limited to; Conservation Security Program, Emergency Loan for Seed Producers, Emergency Loans, Farm Labor Housing Loans and Grants, Farm Ownership Loans, Special Apple Program, and the Wetlands Reserve Program.

Free credit monitoring. Nice. Thanks.

UPDATE:
Sex Lube Maker's 250K Customer List Slides Onto Net (wired.com)
Attrition.org Data Loss Archive and Database (DLDOS) (attrition.org)

ISP Account Log File Analytics

O'Reilly Radar has a post today about the acquisition of Hitwise by Experian.

Hitwise, the company that performs analysis of log files from 25 million worldwide ISP accounts to provide relative market share graphs for web properties, has been acquired by Experian for $240 million.

For those who don't want to be "pushed, filed, stamped, indexed, briefed, debriefed or numbered", just hope that your ISP is not participating (or, find a good secure proxy, as it sounds like managing your cookies and IP address may not be enough). Of course, it's a little hard to tell what qualifies as an ISP these days (cable? mobile?)

Update: The Hitwise How We Do It page gives an overview of the "network-centric" part of the methodology:

Hitwise has developed proprietary software that Internet Service Providers (ISPs) use to analyze website usage logs created on their network. The anonymous data sent to Hitwise from the ISPs include a range of industry standard metrics relating to the viewing of websites including page requests, visits and average visit length.

Hitwise also combines this rich ISP data with a worldwide opt-in panel to overlay demographic, lifestyle and transactional behavior across the thousands of websites that are reported on every day.

Because of the extensive sample size of network data, Hitwise can also provide detailed insights into the search terms used to find thousands of sites as well as a range of clickstream reports, analyzing the movements of visitors between sites...

...Hitwise only extracts aggregate information from ISP networks and no personal information is seen or captured by Hitwise in accordance with local and international privacy guidelines. Hitwise's methodology is audited by PricewaterhouseCoopers on an annual basis.

Reads a lot like most privacy policies. Maybe it's OK, but there's an awful lot of wiggle room. The clickstream and search term references are the most bothersome (didn't we learn anything from the AOL fiasco?), particularly if tied to some localized demographics.

Related Post:
ISPs Selling Clickstreams

Geek of the Year Awards 2007

Nominations for the Top Geeks of 2007 are now open. In the spirit of the occasion, I'm even wearing my "There are only 10 types of people..." shirt today. Here's my starter list, superheroes all for their contributions to the sport:

Gina Trapani (Lifehacker): For improving our productivity so much, we can now burst out the daily workload in under thirty minutes.

Bruce Schneier (Counterpane): For debunking security theater and showing the world that security and privacy are not mutually exclusive.

Anne Zelenka (Web Worker Daily): For insightful analysis of Office 2.0 trends and tools.

Ryan Singel (Wired): For unrelenting exposure of threats to privacy.

Steven Pinker (Harvard): If the brainy research and accessible writing aren't enough, surely his membership in the Luxuriant Flowing Hair Club for Scientists clinches it.

Daniel Solove (GWU): For proposing much-needed legal reform to balance risks and benefits in the age of digital dossiers.

Now, if only Lifehacker would reduce the frequency of posts, I'd have even more free time.

[comic credit: xkcd]

New High-Tech Book License

This book is licensed for use only in a residential setting, for a one week period. The paper has been designed to disintegrate one week after the hermetically-sealed shrink wrapping is removed. This license gives one end user the right to read this book up to two times, but only under incandescent lighting. To optimize user enjoyment, the publisher has incorporated proprietary color correction technology into the paper, printing, and photos, and each version of the book is optimized for different viewing conditions. Other license versions are available (at additional cost to the user) for fluorescent, halogen, and daylight conditions. There is also a commercial license version available which allows the book to be read away from your primary residence. This book, in all of its versions, editions, forms, and instantiations, remains the property of the publisher.

Copyright ©2007 Media Industry Interests, L.L.C. All Rights Reserved. Media Industry Interests, L.L.C., is a trademark of Media Industry Interests, L.L.C., in the USA and other countries.

[photo credit]

Who's watching the watchers?

Wired has a nice trio of articles out today by Ryan Singel about the various terrorist watch lists out there:

A Watch List Is Born

U.S. Watch Lists Sow Frustration and Fear

How to Get Off a Government Watch List

[via BB]

P.S. For reference, here are Bruce Schneier's past comments on the TIDE, No-Fly (No-Fly update), and OFAC lists, Secure Flight, the OneDOJ database, and the privately-run CLEAR registered traveler program.

[photo credit]

Backing Up Your Data

If you're anything like me, you have an ever-growing appetite for hard disk space, and an ever-growing unease about how to keep all of that hard-acquired data from suddenly vanishing into the ether. The obvious answer is to back it up. But how?

In the old days (after floppies), CDs sufficed. But even with a modest accumulation of digital stuff, 700MB/disc didn't go very far. Then DVD burners became common, but even at 4GB (or 8GB) per disc, it takes a lot of discs, and time, to back up a typical collection of documents, photos, music, and video. On top of that, most burnable CDs and DVDs have a very limited and unpredictable lifespan. High-def DVDs aren't really ready for prime time, tape backup systems are probably most suitable for die-hards, and non-volatile storage devices (like USB flash drives) are still relatively small. So most discussions of hard disk backup focus on having another hard disk, or uploading data to a remote server.

These days, keeping your data safe requires more than a procedure; it requires a strategy. It's important to consider things like what portions of your data are most critical, how frequently various sets of data change, how private some of your data is (encrypt!), how long you need to keep it, and how much time and cost it's worth to keep it safe. Also consider the backup media characteristics: capacity, annualized cost per gigabyte, time and effort, ease of use, robustness, compatibility, and futureproofness. The most solid recommendation is to back up early, often, and to as many places as practical.

OK, you already keep essential data on your flash drive, back up some other key stuff regularly onto CDs or DVDs, and upload certain things to various remote servers. Maybe you even have the Firefox Gspace extension that lets you use your Gmail account as easy file storage. Let's focus on the high-capacity needs (but not forget about the inherent unpredictability of hard disk failures).

First, check out Lifehacker's articles on backups for Mac and for Windows

Don't have a spare drive? Space is cheap these days; the cost/gigabyte sweet spot is with the 300-400GB drives:


[Pricing data is for internal Seagate Barracuda 7200RPM 3.5" PATA hard disk drives from macsales.com as of 12 Apr 2007]

If you don't have a tower with a spare drive slot, add an enclosure, multiple disks, RAID, new controller, etc.; whatever it takes to put together a working subsystem.

Also, to get back to a time-based cost per gigabyte, let's be conservative and assume that you'll always have your data on at least two disks, and that you'll replace drives regularly at, say, halfway through the warranty cycle. So barebones ongoing cost would be ($0.28/GB)*(2 drives)/(2.5 years) = $0.22/GB-year, or about two cents per month for each gigabyte.

Expect to invest some time and effort into setting up hardware and software, backing up, checking backups, and periodically replacing drives. If you want to outsource all of that maintenance, and you're willing to pay for it ($0.20/GB transferred + $0.15/GB-month), you could go with something like the Amazon S3 service. It's a pretty safe bet that Amazon will still be around for the next generation. Windows users can also try Mozy (Mac version coming soon). For online backup, hopefully you have a fat upload pipe.

UPDATE (Apr 23): Mozy has hit the big time.

[photo credit: Wikimedia Commons]

The Great Fluorescent Migration

Today's Wall Street Journal has an editorial [via NetworkWorld] making a case against prohibition of incandescent light bulbs.

But as usual, the real answer isn't at the extremes. Sure, incandescents use a lot more energy, but are we really ready for a complete switch?

I generally dislike incandescents. They're cheap, but they're hot, they don't last long, and they use a lot of power. I switched over to compact fluorescents in as much of my house as I could and saw a dramatic drop in energy consumption. They are a little quirky, like when they don't turn on immediately, take time to get to full brightness, or run dimmer in the cold. But those are minor compared to the energy and environmental savings.

But some problems remain unresolved. The variety of affordable compact fluorescents is very limited. They are readily available in configurations for standard bulb sockets, R30 floodlights, and a couple others. But they are still nonexistent, scarce, or very expensive for other applications (dimmer switches, small bulbs and lower light levels, halogen fittings, etc.)

There are also the environmental considerations. There is an argument to be made that the coal plants that produce the higher electrical levels needed for incandescents release more mercury into environment than a broken compact fluorescent bulb. But this isn't quite apples to apples as the dispersal pattern for broken bulbs is much more localized.

What's really needed is not a ban, but proper incentives. Incentives for innovation and manufacturing to bring down costs and expand offerings of fluorescents (and other lighting technologies like LED). We need incentives for consumers. Maybe even some kind of disincentives for overconsumption of electricity. We need consumer education about mercury handling. And, we need pervasive local recycling of fluorescent bulbs.

UPDATE [13 Apr 2008]: Are Compact Fluorescent Lightbulbs Dangerous?

[photo credit: Wikimedia Commons]

The Post-eMail Era

In the midst of wondering whether email is dead [WWD], we have debates raging today about Gmail vs. the webmail competition [PC Mag, via WWD] and also about desktop email vs. webmail [Wired, via LH].

I've said before I have use for both Gmail and Yahoo! Mail (Plus), but in the end, for precisely the reasons that Lifehacker's Adam Pash points out, "security and control", I won't be leaving my desktop email client behind anytime soon.

Related Posts:
Gmail vs. Yahoo! Mail
Google vs. The desktop