The Outbound Index is about providing a service useful to businesses in
eliminating the problems caused by spam.
It is not a search for a
theoretical utopian solution to email spam - that is the purvey of
fine discussion groups such as ASRG, IETF standards groups and others -
who are doing a much needed service which will benefit the global
community in the future.
I come from the "practical applications of technology to a business problem" perspective. My primary goals are:
- Make sure all revenue generating emails arrive to the business and get seen
- Employees waste very little time each day on emails such as
spam, viruses, scams - energy or actions which may harm or at least do
nothing positive for the business
The Outbound Index was conceived and designed to function on a global
scale. It has been in use for several years in the service of business
customers of a small hosting company (a few thousand users.) Beyond
this fairly local use, developer/users from Malta, France and Romania
have successfully queried the Outbound Index millions of times.
We have long term experience using the Outbound Index with real
business customers - as the only decision factor in tagging
"suspicioius" mail. Customers tell us they "love it" - with the user
interface of MailCruncher.com. So from our viewpoint: Does it solve the business problem? YES.
Can the Outbound Index, and should it, be expanded to help more ISPs,
more businesses and more end user recipients? That is the topic this
article explores.
( Screenshot showing results of queries to the Outbound Index )
As you consider whether or not the Outbound Index is valid or has
value, I'm imagining the thought process or objective reasoning,
falling mainly into three categories:
1. Central authorities are a bad thing - they represent a point of
failure, a place where ethical corruption potential means trust cannot
be placed in them.
I would agree in the sense that one should stay very well informed
about who owns and who directs any central authority one trusts. I hope
you know who is behind the root DNS servers you depend on, and every
DNS server you depend on, because they control very close to
everything. You should also check up on who controls the internet
backbone data flows as well as who can turn your electricity on and off.
Look over those behind the Outbound Index carefully. There are major
technology corporations who we would not accept as partners or
investors. We see that this role requires a precise vision of service
to recipients, not to service to senders or highest bidders.
Senders could self-declare what servers they authorize, and some do by using SPF et al.
Business recipients still need additional verification beyond that, for two
reasons. One, businesses will lose mail they want - if they reject mail
without an authorized server match. (See article explaining Non-Malicious Forging.) Two, spammers are now the most likely to
authorize their own servers - if you accept all mail from
self-authorized servers, you'll accept a lot of spam and be right back
where you started with lost employee productivity.
2. Scaleability. I've heard it said that it is impossible to list all email servers or to catalog all the relevant data.
Our practical experience has been that cataloging the data necessary to
make the system work "pretty good" (as in Pretty Good Privacy or other
modest, realistic and pretty damn good products and services) is very
possible. We've done it and are doing it, and going further or all the
way wouldn't be that big of a deal. Even at the present stage - it
accomplishes the goals relevant and useful to businesses.
3. Anything you can do, I can do better - myself, alone, on my own email server, for free - without a central authority.
CACHING
First consider caching. If the Outbound Index has handled a million
unique IP+domain requests today and your own server has handled 1,000 -
would the Outbound Index cache be more likely to have the next answer
you need in its cache, or would your own server be more likely to have
the answer you need cached?
The Outbound Index was conceived from its inception as a
community-beneficial cache. We are only dealing with data related to
email servers and domains used in email - a fraction of the larger
Internet world of IPs and domains. One of our top technical goals is to have a super fast
cached answer to your question about IPs or domains used in email.
RESOURCE CONSUMPTION
Next, let's look at resource consumption. For an inbound email server
to do a large number of DNS lookups for each piece of mail, WHOIS
lookup and parsing, possibly grey-listing, local lookups of RHS
white or black rules, SPF checking, Habeus, Bonded Sender, IADB,
etc is far more resource intensive than doing one Server Index Query
protocol query to the Outbound Index.
(The SIQ protocol by the way is unencumbered and in the public
domain:
clients already exist for it for Sendmail, Exchange, postfix
policy daemon, plus command line clients for both Linux and Windows.
Founders and contributors to the Outbound Index project created the SIQ
protocol - it has an RFC at the IETF and we placed it in the public
domain immediamente, just as we placed VARA and .mxout in the public domain.)
UPDATING & STABILITY
If you are going to duplicate, by yourself, on your own inbound server,
the useful stuff that the Outbound Index does - I'm thinking you are a
single workaholic, get little sleep and have no responsibilities other
than this one task of tuning your inbound mail server. You're obsessed
with anti-spam and avidly read mountains of material daily and spend
feverish nights dreaming up new ways to stop spammers. You're a highly
competent bit level programmer who knows all the protocols inside and
out and can actualize any kind of anti-spam test you or the SpamTools
gang can dream up. Bravo - let me know what mailing lists you are on so
that I can learn from your trials and triumphs.
Most email admins don't fit the above description. They don't have time
to be an expert in this niche area. They can't write a milter or an
SMTP interfaced C or python program every day. Smaller operations
install Spam Assassin and periodically update it or mess with it a bit,
or find another such solution or combination of solutions. If the
solution requires manual attention, updating or upgrading, it may or
may not get done. It may or may not be optimally configured. After all,
the people creating these solutions are themselves just learning what
might be optimal before the scene shifts again.
On the downside, stability of email service is at risk every time the
inbound email server is modified - thus with every required update,
upgrade or "new kind of test." In an ideal world, I'd like to leave my
email server untouched and let it run reliably forever just as it does
today. Customers get really pissy when something goes wrong with email
- more so than even their website or their workstation.
The Outbound Index was conceived as a resource-offloading, offsite,
outsourced powerhouse of databases about email servers. Domains and IPs
used in email. Equivalent for inbound servers - to what a render farm
is for animated film makers.
ECONOMY OF ACTION
If you are operating an inbound server alone, you may not care about
the 100,000 other operators like you, or the economy of action provided
by adding a new kind of anti-spam test in ONE location (the Outbound
Index) - automatically benefiting 100,000 guys who didn't need to risk
their own server stability by installing an update or spend the time to
do so. Not to mention if they each had to dream it up, test it over a
wide variety of scenarios, write it and interface it to their various
systems, and maintain the additional servers required to keep the data
behind it up to date.
Some tests done or data the Outbound Index uses - just wouldn't be
practical for you as a single server operator to carry out. But when we
are serving a wide spectrum of inbound servers, it's very practical.
SUBJECTIVE vs OBJECTIVE "FACTS"
This all ties together that we conceived the Outbound Index as a
community-beneficial repository of facts about email servers. What do I
mean by facts?
- NO "samples of spam" - we actually never get any headers or bodies of mail and have no clue of the content
- NO subjective judgements or judgements applied to servers with the same statistical profile differently
Facts are things like "date the domain of the envelope-from was
registered" or "churn rate of the name server" or "correlated identity
data" or "number of gosh darn it had another affiliate fool us" public
statements.
Leave a comment here or at the bottom of any http://wiki.OutboundIndex.net/ page and give me your feedback on the question this article title poses.
|
|
|||||
|
Search
Categories
Recent Photos
Recent Visitors
Cristian - Thu 19 Feb 2009 01:10 AM PST
April Lorenzen - Fri 11 Apr 2008 08:31 AM PDT
kitchen - Wed 23 May 2007 11:00 PM PDT
Al Turtle - Fri 24 Mar 2006 04:52 PM PST
Lelain - Sun 19 Feb 2006 10:17 PM PST
Recent Entries
Recent Comments
This Month
Month Archive
Login
|
Should the Outbound Index be Expanded to Serve a Global User Base?
Comments
No comments found.
Trackbacks
TrackBack URL: |
||||
|
|
|||||