Thank you to the
Real World Crypto steering committee
and to Max Levchin for this
recognition. I couldn’t be more proud of what our team has accomplished
since we started working on Let’s Encrypt back in 2013.
My first temptation is to name some names, but there are so many people who
have given a significant portion of their lives to this work over the years
that the list would be too long. You know who you are. I hope you’re as
proud as I am at this moment.
Let’s Encrypt is currently used by more than
280 million websites,
issuing between two and three million certificates per day. I often think about how we got here, looking for some nugget of wisdom
that might be useful to others. I’m not sure I’ve really come up with
anything particularly profound, but I’m going to give you my thoughts
anyway. Generally speaking: we started with a pretty good idea, built a
strong team, stayed focused on what’s important, and kept ease of use in
mind every step of the way.
Let’s Encrypt ultimately came from a group of people thinking about a pretty
daunting challenge. The billions of people living increasingly large
portions of their lives online deserved better privacy and security, but in
order to do that we needed to convince hundreds of millions of websites to
switch to HTTPS. Not only did we want them to make that change, we wanted
most of them to make the change within the next three to five years.
We thought through a lot of options but in the end we just didn’t see any
other way than to build what became Let’s Encrypt. In hindsight building
Let’s Encrypt seems like it was a good and rewarding idea, but at the time
it was a frustrating conclusion in many ways. It’s not an easy solution to
commit to. It meant standing up a new organization, hiring at least a dozen
people, understanding a lot of details about how to operate a CA, building
some fairly intense technical systems, and setting all of it up to operate
for decades. Many of us wanted to work on this interesting problem for a
bit, solve it or at least put a big dent in it, and then move on to other
interesting problems. I don’t know about you, but I certainly didn’t dream
about building and operating a CA when I was younger.
It needed to be done though, so we got to work. We built a great team that
initially consisted of mostly volunteers and very few staff. Over time that
ratio reversed itself such that most people working on Let’s Encrypt on a
daily basis are staff, but we’re fortunate to continue to have a vibrant
community of volunteers who do work ranging from translating our website and
providing assistance on our community forums, to maintaining the dozens
(maybe hundreds?) of client software options out there.
Today there are just 11 engineers working on Let’s Encrypt, as well as a
small team handling fundraising, communication, and administrative tasks.
That’s not a lot of people for an organization serving hundreds of millions
of websites in every country on the globe, subject to a fairly intense set
of industry rules, audits, and high expectations for security and
reliability. The team is preparing to serve as many as 1 billion websites.
When that day comes to pass the team will be larger, but probably not much
larger. Efficiency is important to us, for a couple of reasons. The first is
principle - we believe it’s our obligation to do the most good we can with
every dollar entrusted to us. The second reason is necessity - it’s not easy
to raise money, and we need to do our best to accomplish our mission with
what’s available to us.
It probably doesn’t come as a surprise to anyone here at Real World Crypto
that ease of use was critical to any success we’ve had in applying
cryptography more widely. Let’s Encrypt has a fair amount of internal
complexity, but we expose users to as little of that as possible. Ideally
it’s a fully automated and forgettable background task even to the people
running servers.
The fact that Let’s Encrypt is free is a huge factor in ease of use. It
isn’t even about how much money people might be willing or able to pay, but
any financial transaction requirement would make it impossible to fully
automate our service. At some point someone would have to get a credit card
and manage payment information. That task ranges in complexity from finding
your wallet to obtaining corporate approval. The existence of a payment in
any amount would also greatly limit our geographic availability because of
sanctions and financial logistics.
All of these factors led to the decision to form
ISRG, a nonprofit entity to
support Let’s Encrypt. Our ability to provide this global, reliable service
is all thanks to the people and companies who believe in TLS everywhere and
have supported us financially. I’m so grateful to all of our contributors
for helping us.
Our service is pretty easy to use under normal circumstances, but we’re not
done yet. We can be better about handling exceptional circumstances such as
large revocation events. Resiliency is good. Automated, smooth resiliency is
even better. That’s why I’m so excited about the
ACME Renewal Info
work we’re doing in the IETF now, which will go into production over the
next year.
Everyone here has heard it before, but I’ll say it again because we can’t
afford to let it slip our minds. Ease of use is critical for widespread
adoption of real world cryptography. As we look toward the future of ISRG,
our new projects will have ease of use at their core. In fact, you can learn
about our newest project related to privacy-preserving measurement at two of
this afternoon’s sessions! Getting ease of use right is not just about the
software though. It’s a sort of pas de trois, a dance for three, between
software, legal, and finance, in order to achieve a great outcome.
Thank you again. This recognition means so much to us.