|
Sender Address Validation and Authentication (SAVA) is the silver bullet. It will send to Cyberia all dark forces that make us shiver when we make a purchase on the internet, pose a threat to our very identities and have made DDoS a feared acronym.
Some of you will remember the heated debates when Calling Line Identification (CLID) was first introduced in telephony. Libertarians of all stripes called passionately to ban such an evil tool threatening our most precious civil liberties like the impunity of calling home from the bar, pretending to be still at work or with a customer. Today everybody welcomes the decline of crank and obscene calls even if telemarketers can continue to be a nuisance. Will SAVA be for the internet what CLID was for telephony?
One of the beauties and at the same time a source of potential vulnerability of the internet design is that it forwards packets connectionless, hop by hop, based on the destination address. This has proven a cornerstone of the amazing resiliency and scalability of the internet. The flip side is that this makes the blue box offspring, address spoofing more prevalent. From making occasional free calls in the ‘telephony era’, internet address spoofing now substitutes legitimate source addresses to fraudulently obtain personal information from unsuspecting end-users or wreak havoc flooding network hosts, DNS systems and even networks with DDoS attacks. So much so that a number of ISP’s now offer ‘scrubbing services’ to their customers. Zacks Investment sees Cyber Security firms as a major investment opportunity. This is surely a growing and lucrative market segment; I might follow their advise.
SAVA was first presented at an IEEE conference in 2007 and subsequently proposed as a RFC to the IETF in 2008 with Tsinghua University of Beijing as lead author. The paper addressed the need for source address verification on the access network, intra-AS within a network, and inter-AS between networks across BGP boundaries. This led to the creation of a quite active IETF working group called SAVI to tackle the subject. An informational draft issued this February provides a good overview of a variety of ‘attack vectors’ and threats. How fast some of these RFC will be completed and approved and, more importantly, implemented remains however an open question.
China has reported that it is experimenting with a SAVA implementation in its CNGI (China Next Generation Internet) IPv6 only based R&E network, in no less than the United Kingdom’s prestigious Philosophical Transactions of the Royal Society. This has in turn triggered some activity in the blogosphere ranging from more factual to a bit more alarming. Concluding yet again that China is light years ahead of the United States in IPv6 deployment remains questionable however. While CNGI has without question been the benchmark for native IPv6 deployment for many years in a Research and Education Networking environment, China has been really lagging so far in the commercial deployment of IPv6. They obviously bide their time.
While some will argue that SAVA would undermine their civil liberties and individual freedom especially when they prefer anonymity in whatever they are doing on the internet and others will see it as another step to big brother watching us, the need for better security is undeniable and even more urgent as we accelerate towards a mobile broadband data environment. IDC predicts that, this year, smartphone sales will for the first time surpass feature phones. Mobile operators enjoy usage based services and billing; to correctly identify the source will always remain essential to revenue generation and corporate wellbeing. And what would the impact be of a DDoS attack choking a major LTE network?
Major ISP’s and mobile operators might want to track SAVA more closely; ça va ou ça va pas?
Sponsored byDNIB.com
Sponsored byIPv4.Global
Sponsored byWhoisXML API
Sponsored byRadix
Sponsored byCSC
Sponsored byVerisign
Sponsored byVerisign
I wrote about the benefits of IP address verification (briefly) in section 7.3 of my PhD thesis. The thesis is an analysis of techniques which can be used to limit the impact of hostile participants in network protocols, and an example of design using the techniques which are found to be most effective. Source address verification has real benefits, but it’s only a “silver bullet” against a very narrow class of attacks, and DDoS isn’t one of them.
Can we please cut the hype, and pretend to be engineers?