Even Akamai screws up their SSL certs

https://www.akamai.com/

by staunch • 12 years ago

91 33 12 years ago

agwa• 12 years ago

Actually, they probably don't intend for people to use www.akamai.com with HTTPS. If you add an override for the bad cert, you end up getting redirected to a non-HTTPS site anyways. I don't see anything like a login link on www.akamai.com so this is probably OK. (Of course it would be nice if everything were HTTPS...)

pipeep• 12 years ago

Of course it would be nice if everything were HTTPS...

Nice try, Certificate Authorities!

fragmede• 12 years ago

> Of course it would be nice if everything were HTTPS...
Nice try, Certificate Authorities!

Nice try, NSA.

Shish2k• 12 years ago

It would be nice if everything were HTTPS, but the default behaviour for a self-signed cert was "accept it the first time, but don't show the lock icon or anything, and then warn if it changes" - ie self-signed HTTPS is the new HTTP

slipangel• 12 years ago

The a248 cert is used for all SSL requests for properties delivered over Akamai's non-ssl network. It's useful to be able to serve SSL requests even if the cert doesn't have a the hostname in question on it.

Regarding the cipher sneering below: Akamai's default settings probably err on the side of acceptance. Not all of their customers are Amazon.coms - mom & pop ecommerce are unlikely to do anything that would cause a visitor's SSL connection to fail, and changing the ciphers back to the high-acceptance settings would increase their cost of implementation per cert. I'm guessing they manage more than a few of those.

Mindless2112• 12 years ago

Only partly related: Most websites don't get proper certificates for their FQDN -- even Google [1]. That, to me, is screwed up.

[1] https://www.google.com./

theGimp• 12 years ago

That works just fine for me.

cynix• 12 years ago

Works in Chrome 32, does not work in Firefox 26.

chrismorgan2• 12 years ago

It shouldn't; if it does, it is most likely that your browser is removing the trailing dot which makes it a FQDN.

Firefox's verdict:

www.google.com. uses an invalid security certificate. The certificate is only valid for www.google.com

The fact of the matter is that FQDNs are probably only used in three ways:

1. By accident;

2. As a curio;

3. In complaints about how they don't work properly—badly configured server configurations, invalid HTTPS certificates, &c.

Buge• 12 years ago

I don't think it's possible to buy a certificate with a dot at the end, so I don't see how this is Google's problem.

It seems to be either a problem with Firefox or with the domain name system in general.

It looks like the problem should be fixed in 2016 because then all non-FQDN certificates will have expired. Then browsers can go ahead and assume that google.com. in a url is the same as google.com in a certificate.

chrismorgan2• 12 years ago

I don't think it's possible to buy a certificate with a dot at the end,

Citation needed.

breser• 12 years ago

I think Firefox is actually wrong. They do have long standing open bug on this with a lot of, in my opinion, nonsensical arguments. [1]

RFC 6125 [2] says that DNS-IDs should be fully qualified domain names and points you at RFC 1034 [3] for a definition (which doesn't have a definition sadly).

RFC 1034 does specify the difference between complete (or absolute) domain name and a incomplete (or relative) domain name, with complete domain names including the trailing dot. As best as I can tell the first use of Fully Qualified Domain Names (FQDN) came from RFC 1535 [4]. It doesn't really define the term either but proceeds to use it and mentions "absolute 'rooted' FQDN" in such a way that it implies that absolute is adjective. Finally, we get to RFC 1594 section 5.2 [5] which tries to clear up the matter and describes a FQDN in a way that would seem to agree with your usage (requiring the trailing dot), while giving examples that omit the trailing dot. The Wikipedia article on FQDN [6] says otherwise that the trailing dot is necessary. So at best the definition of FQDN is ambiguous in my opinion which means the RFCs don't help us much here with determining the proper way of handling the trailing dot with respect to PKIX certificate validation.

However, I think that there are some conclusions we can draw using some common sense and thinking about the implications of the behavior.

1) Hostnames are commonly used without the trailing dot. As long as they only are missing the trailing dot they are still fully qualified. They may not be absolute and are still relative to the root.

2) Nobody issues certificates with the trailing dot in the common name. It doesn't make much sense for the RFC to be defined in a way that is entirely incompatible with the existing certificates. Especially since it went to such an effort to define how to deal with the old Common Name sytle and not only the newer SubjectAltName style.

Maybe they should have been using absolute hostnames in order to allow everyone to distinguish between certificates that were issued for relative domains but they don't. In general I don't think CAs should be issuing certificates for relative domains.

Mozilla's own CA Certificate Inclusion Policy [7] disallows certificates being issued that are not absolute domains. It does so by saying "the CA takes reasonable measures to verify that the entity submitting the certificate signing request has registered the domain(s) referenced in the certificate." I can't fathom how issuing a certificate based on a relative domain say "www" could ever be compliant with that.

3) What's the worst that could happen if you allow a match on "www.example.com." to the certificate with "www.example.com" in the CN or SNA? Well not much because the trailing dot forces you to the absolute hostname which happens to be what the CA issued the certificate for.

You can't (assuming a search domain of "example.org") ever end up at "www.example.com.example.org" and therefor be matching the wrong domain because of the trailing dot tells the resolver not to do that. If a browser does do that then they have a separate bug.

[1] https://bugzilla.mozilla.org/show_bug.cgi?id=134402 [2] http://tools.ietf.org/html/rfc6125 [3] http://tools.ietf.org/html/rfc1034 [4] http://tools.ietf.org/html/rfc1535 [5] https://tools.ietf.org/html/rfc1594#section-5 [6] http://en.wikipedia.org/wiki/Fully_qualified_domain_name [7] http://www.mozilla.org/en-US/about/governance/policies/secur...

patcheudor4• 12 years ago

Better yet, where they absolutely mean to use HTTPS they sometimes use weak keys and ciphers and get an "F" from the Qualys SSL Labs tool. Blogs.akamai.com isn't the only place this happens:

https://www.ssllabs.com/ssltest/analyze.html?d=blogs.akamai....

patcheudor4• 12 years ago

Someone's parents are going to be really mad about this report card:

https://www.ssllabs.com/ssltest/analyze.html?d=developer.aka...

https://www.ssllabs.com/ssltest/analyze.html?d=a248.e.akamai...

https://www.ssllabs.com/ssltest/analyze.html?d=network.akama...

The thing is, the worst part is knowing a child is capable of A's:

https://www.ssllabs.com/ssltest/analyze.html?d=control.akama...

brians2• 12 years ago

That's my report card.

The grades are up to Cs now and should be As within a day.

I'm glad Ivan and Qualys are helping show us where we really are.

mnordhoff• 12 years ago

Not the biggest crisis. Okay, they have several awful cipher suites enabled, but no sane client would ever use them. The client report shows that almost every client uses AES; a couple crappy ones use RC4 or 3DES.

They don't have PFS, either. That's bad, though unfortunately still common. As far as I know Akamai's position is that the (small) performance cost of PFS is unacceptable. They would be delighted if it was faster (which people are working on).

I think the F is because of the 1024-bit, MD5 CA. That seems to be more of an argument for clients to disable that CA certificate, especially since there's another, good trust path, but maybe I'm missing something.

patcheudor4• 12 years ago

All great points. The biggest issue I see with their apparent lack of consideration towards a passable SSL/TLS configuration is that it extends outside of their own corporate web space to impact their customers. As just one example:

https://www.ssllabs.com/ssltest/analyze.html?d=ozsale.com.au

You'll see in the Qualys report that the Akamai accelerated Ozsales gets an "F" & from the Qualys report you'll note that results in a PCI compliance failure (near the bottom of this report):

https://www.ssllabs.com/ssltest/analyze.html?d=www.ozsale.co...

tmx• 12 years ago

no sane client would ever use them

Unfortunately, that's not the case. From RFC 5246:

  Note: some server implementations are known to implement version
  negotiation incorrectly.  For example, there are buggy TLS 1.0
  servers that simply close the connection when the client offers a
  version newer than TLS 1.0.  Also, it is known that some servers will
  refuse the connection if any TLS extensions are included in
  ClientHello.  Interoperability with such buggy servers is a complex
  topic beyond the scope of this document, and may require multiple
  connection attempts by the client.

An attacker that can insert themselves between client and server can cause a negotiation failure - causing the client or server to negotiate a weaker protocol/cipher combination - so its still important to disable weak cipher suites on the server, especially with browsers as old as IE 6 or 7 still kicking around.

brians2• 12 years ago

Indeed, it looks like that's a bug in the SSL labs rating scheme: given two trust paths, it takes the longer one.

The Baltimore root is trusted, but also signed by the old GTE 1024-bit root. It's not clear to me what harm it does to have an appendix of old roots above a well-managed, trustworthy trusted root.

dredmorbius• 12 years ago

This issue has been hitting APM's MarketPlace website (https://www.marketplace.org/) for at least a year now.

Mucks up their CSS as well, best I can tell.

dsl2• 12 years ago

This happens with almost any Akamai'zed domain, if you drop https:// in front of it you get the shared Akamai cert. Customers that pay for SSL get their own pool of IPs that respond with only their cert.

Same thing happens with any shared web host that happens to listen on SSL.

tantalor2• 12 years ago

So Akamai couldn't afford to pay themselves for their own IP pool and SSL cert? Poor guys.

dsl2• 12 years ago

The customer portal is at https://control.akamai.com/

www is literally just a sales site.

jdubs• 12 years ago

I blame the keystore and the dreaded keytool.

hercynium• 12 years ago

chuckles

This isn't a SSL cert screw-up, but certainly a silly misconfiguration. They should certainly have known better than to let that happen.

To be fair... when I worked there, I thought about different ways to transparently enable SSL across all domains using a CDN that would work with all existing SSL-enabled browsers and it's a freakin' hard problem. There are potential solutions, but they're not particularly cheap or simple given the IPv4 address crunch and for Akamai even more so since the edge servers are so widely distributed.

Bottom line - I never pitched the idea anyway because while I was interested in making the internet a better place, I knew that it was DOA at Akamai because they're much more interested in making it a more profitable place, and I couldn't think of a strong-enough business case...

tantalor2• 12 years ago

It might compete with their current, which bundles SSL with dedicated, static IPs. Although IPv6 should solve that scarcity.

ggoddard• 12 years ago

https://twitter.com/csoghoian/status/428228534675308545

acdha• 12 years ago

The main site is basically marketing - their customer portal, which actually has sensitive content uses HTTPS sanely:

https://www.ssllabs.com/ssltest/analyze.html?d=control.akama...

Even Akamai screws up their SSL certs

Related Stories