| Updated: 23.6.2002; 12:30:41 GMT. |
| Security weblog Network World Survey: Security is important, but we won't pay for it
This illustrates a disconnect that I have often seen in security field. People frequently say something different than they do. 10:02:47 PMSecurity blogging comunity is growing. No nonsense post mainly about network side of the things from Mathew Tanase. 4:02:32 PMThere are been about 30 web services security standards under development. Some think that this is overkill:
I don't think this is going to happen. Distributed computing utilising chain of services over unreliable and insecure infrastructure to get relatively reliable SLAs is a problem that is not easy to resolve. Combine this with the need to communicate with unknown subjects with unknown reputation and you wonder if its feasible at all. Still, it will be important for security standards not to stay in the way of bootstrapping. This means to have modular security specifications that can be added to the basic protocols as the complexity of use scenarios increase. Mandatory requirement to implement all the XACMLs, WS-Policies, WS-Trusts for simple and straightforward SOAP implementation between two mutually trusting applications with known semantics would be nonsense.
Yes, for enterprise use of web services, security is a showstopper. Corporations can't benefit from the web services without controlling access. Temporarily workarounds, like the use of SSL can be used just for point-point integration and the user identities cannot be propagated through the hole transaction chain. Then there is a host of would be standards from SAML to Kerberos to WS-Security to Liberty Alliance and who would like to deploy specification that could soon become obsolete. Everybody is therefore waiting for what the standard churn mill will spill out. 12:19:50 PM
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||