justmee
Posts: 505
Joined: 14.May2007
Status: offline
|
Oh my, here we go again... If you don't do the math and argument, the words have no value . The VPN site-to-site connections, basic IPsec tunnel mode, versus the use of tunnels protected by IPsec is an old discussion. Also the discussion of protecting those tunnels with ESP in transport or tunnel mode. Fine engineers have been debated various solutions over the time, take a look here for an example if you want. It made it even in a nobody's RFC. *Everybody* knows that is hard to beat L2TP/IPsec when it comes to the overhead introduced. This is a reason why some people don't use it even with remote access VPN. People were unhappy for example using GRE over IPsec, which adds only 4 bytes of overhead, much less than L2TP/IPsec. I'm pretty sure that it says nowhere in the Microsoft documents simply that IPsec tunnel mode "is a less-efficient protocol with a higher overhead". In those documents very likely, it is only mentioned Microsoft's implementation of IPsec tunnel mode vs Microsoft's L2TP/IPsec implementation. And any features comparation, advantages/disadvantages are only in that area. If you just take out of context and quote things, then you just say so. "Microsoft's VPN site-to-site world" is just a small world, so very likely people need to interoperate with other vendors in that area. With Microsoft's implementation of IPsec tunnel mode it can be easily entered in some two cents limitations that a karate kid vendor can handle. I do not see growing this world, while these limitations are still there. I do not see growing this world by telling people that now they are in MS world, so they need to use L2TP/IPsec, 'cause IPsec tunnel mode is no good anyway, so just replace their boxes. If a forward thinking discussion, open debate is not used, to acknowledge the limitations of the current implementations, and people just gather and wonder how beautiful is L2TP/IPsec, there will be no pressure on Microsoft to up their game in the VPN site-to-site area. Personally I would like to see that world expanding, day after day, and I would like to recommend people to use ISA for site-to-site connections without the need to debate the match theory. I've used and implemented traditional IPsec tunnel mode connections, L2TP/IPsec connections, GRE over IPsec, IPIP over IPsec, Cisco's DMVPN and SVTIs, Juniper's Dynamic VPNs or Check Point's DVPNs. So I don't need to take or quote anybody. I can argument myself every bit from what I'm saying, including with packet captures(not just draws) for every mentioned above technology, if I have to. I do care how they work, how they were engineered, what they do and what they don't do.
< Message edited by justmee -- 30.Aug.2008 4:23:52 PM >
|