Summary:  Google gets rid of their firewalls and gives us a new model for Security in our mobile world

You have a traditional business IT infrastructure.

Everyone lives in the same building and use the same, internal network.  You know where everyone is, what they need access to, and who is on your network.  You buy firewalls,  switches, and other network infrastructure and services to support this philosophy.  You built your network to keep “Bad, unauthorized” people outside the security fence and only let “Good, authorized” users inside the security fence.  Your firewall keeps you “safe”.

Then the IT universe created mobile workers with mobile phones, tablets, and laptops.  The Work universe created workers who work at home and at Customer sites.  It created Customers who look at their data inside your systems.  All these users need to be “inside” your security fence.

So now what to do?

You have IT create “holes” in your security fence i.e. your firewalls – places in the network infrastructure where you let users outside your network, outside your building, access your internal network and applications.  You are now less safe if only because creating the holes in the firewall creates a possible vulnerability to the outside world.

If you have any type of mobile worker,

you know this problem.  And the bigger you get – the expenses add up to buy equipment, support the infrastructure, and find some place to put it.  Is there a better way ?

Google decided there had to be.

And they put their very considerable resources into creating a model, systems, and processes to change their security model and  increase its effectiveness.  They are leading in showing a new way to secure networks and enable effective workers no matter where they are.  And while this is a model that does not yet translate to smaller businesses (think cost) the principles do.

At a security conference held mid-February 2017,

Google’s Heather Adkins, Director of Security shared their multi-year (5+ so far) journey with their security initiative/system called BeyondCorp.  Google posted an abstract with an overview of their philosophy and approach.  And they are still not done yet.

It’s pretty cool. 

Their basic premise:  “Walls don’t work,” said Heather Adkins. Adkins said the goal was to de-emphasize firewalls and other perimeter defenses and to move to a “zero trust” model.  In other words, until you prove you are an authorized user, you get no access to any network.  Their goal is to allow any user to work from any untrusted network and access services in such a way that they are authenticated, authorized, and their data encrypted.

Here’s how BeyondCorp authenticates users:

 Securely identifying the Device

  1. Using a Device Inventory Database – any device that is going to access the Google network has to be managed via this database.
  2. Device Identity – devices get unique identifiers that tie them to the Inventory database using a device certificate specific to each device; certificates do not guarantee access but are used as keys to information about the device.

Securely Identifying the User

  1. User and Group Database – Users are tracked and managed here and tightly integrated with Google HR processes to identify a user’s job, job category, usernames, group memberships.  As roles and responsibilities change, this information is updated to give proper enterprise access.
  2. Single Sign-On – is a centralized user authentication portal. It validates primary and secondary factory credentials.  Users validate against the User and Group database and then receive short-lived tokens used in the processes that give specific access to specific resources.

In my next blog post, I will summarize how they have and are moving all their applications to the Internet.