Slashdot Mirror
GGF and Grid Security
An anonymous reader writes "Things are changing fast in the grid community. Our communication networks connect millions of systems and billions of individuals on the planet. These myriad systems, and the data they contain, present juicy targets for those who want to steal, damage, corrupt, or otherwise gain unlawful access to those systems."
There are ways to protect sensitive data, such as using VPN's rather than the internet for e.g. Doctors accessing hospital records, grid computing etc. Doing everything on the open internet is neither necessary nor desirable.
I think our software deployment capability exceeds our network architecture design capability.
And if you thought that was boring you obviously havn't read my Journal ;-)
The most secure system int he world won't protect you if your employees aren't trained on how to prevent social engineers from bypassing their security systems anyway. Why spend countless hours trying to hack passwords when you can pretend to be an employee and ask for the info outright? Just take a look at The Art of Deception by Kevin Mitnick. What a great book...
didn't an ISP in NY or something have a room compromised, and 3 T1/T3 cards or something stolen?
What exactly are they, and why is breaking into their systems any worse that breaking into a normal system?
EVERYDAY IS CATURDAY
yea... how many have relevant infrastructure...
First there are resource allocation problems. The OS has to provide a sandbox with strict limits on all resources: memory, filesystem, and networking, as well as CPU time. It's fine with me if the "background compute demon" takes 25% of my processor but I don't want to take more than 10% of my memory.
Then there's the security issue.
But I see another problem which is even harder to solve: the tragedy of the commons. Consider a university campus, and suppose that anyone on campus can submit jobs to the Campus Grid. You come in the next morning and see that there are 10000 jobs in your grid queue, and 9800 of them are encoding random people's MP3's.
The problem is that if you give free resources to a large anonymous community, it takes only a few of those people to suck up all the resources. So you need some way of identifying everyone who submits a job, and some way of charging for the jobs.
The government has actually taken a proactive role in network security with the implementation of the HIPPA act. This has been a blessing in disguise for network admins who have stessed security on their local grids. This act put into law guidelines for securing electronic transmission patient information. Going more indepth with how the information is actually retained within the system (not just the output). For the network admins this act also gave them the flexibility for instating secuity measures that the management may have deemed not revelant earlier.
A bunch of Tech Stuff
In such a vast network of billions upon billions of bits, all interconnected, would we see an AI emerge such as Jane in Orson Scott Card's Ender Series?
I wonder what that AI would do upon emerging? Lurk around in silence? Help or harm the human race? Would it develop its own set of laws?
Or maybe it'll end up being another ELIZA chatbot.
"What about clueless make you want beer drown?"
-Cyc
/.'s 10 Millionth
When you look at case studies of commercial "grid computing", what they're really talking about are dedicated clusters of machines. This is just clustering.
If "grid computing" were saleable, ISPs would be offering off-peak compute time on their server farms, and people would be buying it. They're not.
It's time sharing, people. And time sharing is dead.
How, in this day and age, does Cisco leave sensitive information like their network OS source code on a computer/grid that is accessible from the outside internet?
There can't be real security if people openly allow access to data on their devices.
Poor GUI design, insecure appliction defaults and lack of awareness by users all contribute to poor security.
For example just do a search for boot.ini or inbox.dbx on any p2p program to get an idea of just how many open boxes are out there.
Because this is the same company that sold hardware comprising the backbone of the Internet but was full of H.323, BGP, SNMP, and TCP flaws. Such flawed implementations led to dozens of different exploits being circulated. But they were able to stay ahead of the 8 ball and release patches JIT every time around.
Plan 9 is a great OS to use for gridding and provides extensive security.
All this time I've been saying that the GGF (AuthZ-WG, OGSA-SEC [WS-SEC], CAOPS-WG [CP/CPS with CA], OGSA-AUTHZ [PERMIS, CAS, VOMS...], SA3-RG, ARRG-RG [X.509, SAML...]) needs to address OGSA, OGSI, and WSRF problems with PKI-based security!
Yup, you know it!
It seems to be a commonly held misconception that making your node part of a computational grid implies making all of its resources available in the absence of "physical" layers of security (e.g. VPN, virtual memory protection). This is not true - in the "client pull" model, a node on the grid can choose exactly what it cares to run - and if this is a selection from a small set of trusted programs which operate within certain limits, regardless of parameters, then the client is still secure.
As Bruce Schneier might say, there's no absolute level at which you can judge a system "secure". It's all about establishing a level of security that is acceptable to you.
For one example, the grid compute-farm s/w I wrote allows you to do this.
Grid infrastructure is not just about compute time. It will also attempt to deal with the predicted "data deluge" in the various sciences (chiefly high energy physics, but genetics are also a big producer of data). Storage requirements will increase much faster than the media technology, meaning that new distributed systems will have to be developed to store and access this in a useful way.
Anyway, you can't expect this to leap straight from research papers into commercially viable systems right away. Remember that the Internet started off as a solution in a very specialised field (defence) before commerce started making use of it. Also like the Internet, the main benefits of a global Grid will only become apparent once some significant interoperable installations have been made.
The reason there are few buyers for grid computing services is because it's not ready yet. It is being designed for tomorrow's problems, and, when the time is right, it will certainly have plenty of interested parties.
My primer on distriuted computing includes a shed load of tips for dealing with parasites, spoiler attacks and innocent errors.
In the case of the military, it is possible to enforce security due to the inherent hierarchical nature of the relationships being modeled: witness the structure of X500 and subsequently LDAP based directory services which are derived from X500.
The situation is somewhat more complex with globalized Grid Computing because of the lack of a universally trusted authority. If the ideal of a ubiquitous "on tap" computing resource is to be realised then a model which unifies hierarchical and peer based trust models is needed.
The original post identifies a high level introduction to some of the consortium and forum based research in this area (which everyone is encouraged to contribute to). "It has", as you say, "been some time", however, it's going take significantly more time before the scientists view of grid computing (as opposed to the marketing hype) can be realised.
boakes.org
To answer your first question: the grid community exists through several forums and consortiums.
Now, question 2: The machines in today's grid testbeds are typically just cloned machines so if they get compromised they're easy enough to purge.
The risks are many. Should such a powerful system become compromised at a high enough level (through a social or technical attack), then the potential for a brute force attack on other cryptographically secured systems is high.
Also, due to the "webs of trust" that have to exist between machines within these grids, breaking into a grid which exists across multiple organisations may provide strategic attack points from within those organisatiosn firewalls.
Add to this the fact that the early adopters of grid computing are likely to include pharmecutical corporations, oil companies & the defence industry and it may be apparent that the cost of breaking into such a system may be higher than if John Doe's machine gets owned.
boakes.org
Don't worry - I meta-moderated appropriately.