Securing the 'Net -- at what price?
- 03 July, 2015 00:46
Is it possible to secure the Internet? And if so, what would it cost?
According to Jim Manico, a global board member of the Open Web Application Security Project (OWASP) Foundation, it is. And he suggested a price of $4 billion, in a recent "open letter" to President Obama.
Manico said if the government would write him a check for that amount, "or any amount of money deemed enough to instigate change -- I'll show you how to help secure the software that drives modern businesses and the Internet at large."
If he's right, it could be one of the best, most efficient investments the federal government has made.
Yes, $4 billion is a lot of money -- enough to put an individual at the top end of the 1% income bracket. But it is less than 1% of the money lost to cybercrime annually, which according to a report by Intel Security and the Center for Strategic and International Studies, is $575 billion, with $100 billion of that coming from the U.S.
And compared to the 2015 federal budget of nearly $3.2 trillion, it is not even a rounding error -- barely more than one-tenth of 1%. The government spends that much in less than half a day.
But, of course, the question is whether it is even possible. Experts are generally unanimous that there is no such thing as 100% security. What makes anybody think the World Wide Web can be secured for $4 billion, or even $400 billion?
That depends on the definition of secure. In an interview, Manico agreed that there is no such thing as perfect security, but said it is indeed possible to make vast improvements.
He said $4 billion "is an arbitrarily large number," but would be enough to, "actually get it done, or get the long-term processes rolling and hiring the right teams to work on the problem described."
He noted in his letter to Obama that the route to security is to improve the, "standards, frameworks and languages developers use to build complex software," adding that the technology already exists to protect software from threats like SQL injection and cross-site scripting (XSS).
He gets general agreement on that from Mark Stanislav, security evangelist at Rapid7, who said perfect is not possible, but perfect is not the point either. The point, he said, is that the Internet could be a lot more secure than it is, for a lot less than $4 billion.
"The information security industry has known how to effectively reduce common application security bugs like SQL injection, cross-site scripting, and buffer overflows for decades," he said, "but while frameworks and standards implore and enable developers to prevent those issues, there are still mistakes that can lead to their introduction."
Stanislav said frameworks still have bugs, which he said could be fixed for a price in the hundreds of thousands of dollars, not billions, but he contends the bigger problem is a lack of education of developers, who, "need to know why what they are doing is wrong and not just be expected to implement code a certain way, by rote."
Indeed, he said if every organization correctly implemented, "two-factor authentication, at-rest and in-transit data encryption, and rigid network segmentation, you'd see data-loss statistics drop like a ton of bricks."
Anthony Di Bello, director of strategic partnerships at Guidance Software, also agreed that significant improvement is possible, but said it is not just a matter of money. "The willpower and cooperation across major software players will be more difficult to secure than any dollar figure," he said, adding that if he had $4 billion to address Internet security, he would not spend it on improving software.
"I would focus spending on secure infrastructure, re-architecting the Internet to support secure yet rapid communications, and building security minded controls into networking hardware, firmware, routers, switches etc.," he said.
John Pirc, chief strategy officer at Bricata, agreed with Di Bello that it is more than a matter of money. But he believes it is about more than will power as well. "It will take changing the mindset of individuals, and education in secure software development," he said, noting that in the hyper-competitive world of software development, the pressure is to get products to the market.
Developers and vendors, "don't intentionally let insecure code go to market, but the main focus is time to market which equates to money, and it's a fine balance of risk to the consumer or enterprise that is decided upon from the software vendor," he said.
Manico said he thinks it is possible to improve security even without a wholesale change of mindset from developers, because they tend to use the building blocks that are available. If the most popular open-source software frameworks and web languages become more secure, that will make applications more secure, he said.
He conceded that such an effort would take years to accomplish, comparing it to a government initiative to improve building construction that would, "create both better building code standards and, more importantly, provide technical support, including hands-on support, to manufacturers who are building construction equipment and other construction building blocks to achieve those standards."
"Much of software is built by grabbing a free copy of a 'pre-fab home,' and then modifying it," he said. "If we got pre-fab manufacturers to integrate better security into their offerings, it would go a long way to providing better security across many industries."
His fellow experts agree that would help, but note that the problem is complicated, and made more so by enemies that, to continue his analogy, would be constantly coming up with new ways to knock down improved buildings.
Stanislav said it is not the Internet itself that is always the problem. "In many cases, it's the systems connected to it. Sure, there could be better Internet data security, like updated protocols and more inherent encryption, but ultimately most breaches are not occurring because of an ISP or even a networking vendor, but due to weak passwords or buggy piece of code."
Di Bello noted, as others have, that technology cannot always trump the human factor. "Phishing is a great example of an attack vector that would be undisturbed by a change in application security," he said. "However if a newly architected communication protocol could validate the source of an email, phishing could be cut down significantly."
And Pirc, while he said he is, "a huge proponent of secure coding, I don't think that is going to fix the issues."
While it might temporarily slow attackers down, "they will soon learn other ways to exploit the software. With standards and frameworks, until people stop treating them like checkboxes, they will be ineffective," he said.
The reality, Stanislav said, is that, "the ecosystem is always changing and frameworks/standards alone cannot prevent bugs inherent to an accidental coding mistake by a software engineer writing a compiler or updating an interpreter.
"One typo in a million lines of code can be the difference between a breach and data security."