1. Two factor authentication is not the answer
Why are most websites insecure? Many years ago, I heard that the best “bang
for your buck” was to adopt two-factor authentication. Why is two-factor
authentication not the answer? And does implementing two factor
authentication actually have a high ROI?
I certainly agree that weak authentication costs money. Compromised
passwords can result in serious financial damages for both the person and the
organization. In addition to the direct resource cost of mitigating a security
breach, one has to calculate the opportunity cost of the work not done, and the
potential damage to your personal or organizational reputation.
I also agree that password-only authentication is a big part of the reason
websites are insecure. Hackers have a long list of tactics to find passwords. Some
programs don’t encrypt passwords stored on the disk. Hackers setup fake wifi
access points to capture passwords. People get tricked to login to phony web
sites after reading increasingly sophisticated phishing emails. The list goes on.
Combine this with a person’s tendency to re-use passwords. If a hacker gets
access a person’s email account, all sorts of havoc may ensue.
2. So its seems logical that implementing two-factor authentication would be a good place
to start. However, implementing strong authentication hasn’t been the silver bullet for
organizational security. Why not? The reason is simple… the breadth of applications that
can use the strong authentication has been limited. Although your VPN and company
portal might be protected with strong authentication, today people use so many websites
and SaaS services that they use (and re-use) passwords anyway.
So maybe this is splitting hairs, but I would posit that the “answer” to website insecurity is
Internet standards for authentication. This was something that no one company could
solve alone. Happily, that standard is almost here: OpenID Connect. This standard will
enable any website, even one previously unknown to your company, to use that strong
authentication service you spent so much time and effort launching, and enable twofactor authentication to become the “best bang for your buck” that it was rumored to be
more than a decade ago.
Article Resourse: http://thegluu.weebly.com/1/post/2013/10/two-factor-authentication-isnot-the-answer.html
3. Think about the front door
Businesses are advised to invest in the part of their facility that the customer sees. With access management
systems, this is the login experience, and the authorization experience. Frequently I remind Gluu customers to
consider the authentication triangle, the vertices are (1) security, (2) price and (3) usability. Each authentication
mechanism has its own unique triangle. Much attention lately has been focused on security. But many of the
advancements have been to enable stronger security, while at the same time improving usability. The best kind of
authentication is the one you never see! Consumer IDPs are looking at many contextual indicators to figure out if an
interactive authentication is needed. Organizations should follow suit.
Try your best, but be flexible.
If a certain application can’t use OAuth2, its ok to fallback. There might be an old version of IIS you need to support. Or
the SaaS provider just supports SAML… its ok! Don’t worry. You want to guide applications to use open standards.
SAML or even SiteMinder is a lot better than for the website to store credentials for the person.
Is SiteMinder “Dead”
Granted… “SiteMinder is Dead” is sensationalist. Old SSO protocols hang around until you disconnect the last site.
That can be some time, which is why we want the standards to be well tested. That’s why the title of the previous blog
said “Decline”, not “Dead”. If you have a sizable organization, and are looking at a green field, are you installing a
commercial IAM Suite, an IDaaS, or open source? The last two didn’t even exist until a few years ago. No matter how
you slice it, monolithic IAM Suites like CA SiteMinder are going to get a smaller percentage of the market, and
reducing prices to get a small number of new customers might not be offset by revenue loss from existing customers.
In rapidly growing markets, the price goes down, the total size of the market increases, and the initial suppliers are
challenged to make a very difficult pivot.
4. In any case, at Gluu, we think there is a bigger opportunity to provide service to the
market that doesn’t yet have a “SiteMinder”, than disrupting current monolithic IAM
customers. Most current solutions are hub and spoke: usually a big IDP and lots of
internal websites, some external SaaS services, and partner sites. How many inbound
SAML connections does your average organization support? The answer is frequently
“not many.” Big companies can afford commercial Access Management / Federation
software, but their partners usually cannot. Net-net, this means the cost of “extranet”
user management is either too high or even worse, its insecure. Organizations want
open source because there is a benefit if their partners can cost effectively upgrade their
IAM.
You can substitute “SiteMinder” with the IAM product of your choice, for example Oracle
Access Manager (OAM), RSA Cleartrust, or IBM Tivoli Access Manager (TAM). Although
some IAM products also use HTTP reverse proxies, the idea is generally the same: align
with the old until you migrate existing apps. Notice in this diagram, there are two OAuth2
Authorization Servers. OAuth2 enables federated authorization… sometimes many
parent organizations make different policies, and application developers need to ensure
all the policies are considered.
Article Source - http://www.gluu.org/blog/how-to-move-away-from-ca-siteminder-toopen-source-authn-authz/