January 24, 2018 at 9:05 pm
Comments posted to this topic are about the item Are You Patched?
January 25, 2018 at 1:40 am
Heterogenity may sound nice, but if I have x different systems, it multiplies the number of vulnerabilities too. And of course I would need staff, which knows this systems very good and is able to test / patch them.
I guess there a a lot of small and medium shops out there, who will patch their Windows systems regularly, but runs a 5 year old unpatched Linux server which will be never be touched, because it is too risky to break it (when applying all outstanding patches) or because nobody in this shop knows Linux well enough and because it just runs.
God is real, unless declared integer.
January 25, 2018 at 3:21 am
If your SQL server is physically hosted then doing a test conversion to a virtual machine has quite limited practical test value. It will test that your application still works with the amended OS (Operating System) but with another layer of indirection between the OS and the hardware the overall usefulness of this testing is low as you are likely to be using a different chipset with different configuration. Unfortunately the only really practical test is an identical physical clone of the original system, configured identically but with the patched OS on it but this is an arrangement that is quite hard to arrange unless spare systems were ordered with the original order. If your physically hosted SQL server is well enough secured that users cannot run arbitrary executables on it, and the general principle is that access should be very effectively locked down on such a system, and the few privileged users that do have access do not use it to browse the Internet then as a system it should be pretty much safe and you can relatively safely avoid the OS patches (software workarounds to hardware errors). The key point here is that the server should be effectively isolated, and if you're not effectively isolating your SQL server and running just SQL server on it then you really should be.
If your SQL server is virtually hosted then unless the host OS (VMWare, etc) is patched then there is nothing much that you can do to make it safe because a running process on one guest (hosted) system would be able to read the physical memory of the host OS, which would naturally include the physical memory of all guest (hosted) OSes. Assuming that you have multiple identical host systems then the test process becomes upgrading one of them and either moving or cloning a guest system onto the patched system and testing that.
January 25, 2018 at 6:40 am
Do not rush in to patch without doing your homework and testing!
Dell has rolled back patches for many of their models, people that jumped on the patch train early are having to manually undo patches on hundreds of laptops with a good percentage bricking, and the patches Intel submitted to Linux are poorly written and flawed. Plus there's no firmware patches for older devices.
This situation is a mess, but with no true competition in the laptop, workstation and server world, Intel brought it on through their hubris.Knowledgeable voices in the outskirts of the tech community have warned about this possible situation with Intel processors for years.
It's sad that my two RaspberryPi's are the only semi-useful computers in the house not affected.
January 25, 2018 at 6:46 am
t.franz - Thursday, January 25, 2018 1:40 AMI guess there a a lot of small and medium shops out there, who will patch their Windows systems regularly, but runs a 5 year old unpatched Linux server which will be never be touched, because it is too risky to break it (when applying all outstanding patches) or because nobody in this shop knows Linux well enough and because it just runs.
Caveat emptor!
Anything you buy or use that needs maintenance is your ultimate responsibility no matter the OS. Linux consultants are available, it's not rocket-surgery or the domain of grey-bearded priests. It's no more difficult than Windows, just different.
January 25, 2018 at 7:57 am
So, who knew about this flaw, and when did they know it?
The flaw in chip design has been there since the late '90s, and then the news breaks overnight, 20 years later. If some insiders (ie: tech manufacturers, NSA, Wall Street) knew about it long before it went public, that information could have been used for insider trading or hacking. When big stuff like this happens in the financial, healthcare, or automotive industries, there are typically Congressional hearings and FBI investigations. Perhaps the IT industry should be held to the same standard.
https://www.wired.com/story/meltdown-spectre-bug-collision-intel-chip-flaw-discovery/
https://arstechnica.com/information-technology/2018/01/intel-ceos-sale-of-stock-just-before-security-bug-reveal-raises-questions/
https://www.eteknix.com/expert-says-nsa-have-backdoors-built-into-intel-and-amd-processors/
"Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho
January 25, 2018 at 8:28 am
One of the implications of Steve's advice is that if you are a Windows shop, your firewall and external security solution should be based on Linux, and you should invest as necessary to maintain it. If your are a Linux shop, your security should be Windows. Exploiters tend to have expertise in one or the other, rarely both. This one measure will knock out virtually all of the opportunistic exploiters, leaving you to deal only with those who are targeting you. As for having systems around the operation unpatched because nobody knows and everyone is afraid to touch it: this is just unprofessional. Raise your bar.
January 25, 2018 at 8:51 am
GeorgeCopeland - Thursday, January 25, 2018 8:28 AMOne of the implications of Steve's advice is that if you are a Windows shop, your firewall and external security solution should be based on Linux, and you should invest as necessary to maintain it. If your are a Linux shop, your security should be Windows. Exploiters tend to have expertise in one or the other, rarely both. This one measure will knock out virtually all of the opportunistic exploiters, leaving you to deal only with those who are targeting you. As for having systems around the operation unpatched because nobody knows and everyone is afraid to touch it: this is just unprofessional. Raise your bar.
Frankly, unless you run a gateway packet inspection service (i.e. something that tries to determine malware intent in protocols) your firewall brand really doesn't matter as long as it is adequately secure. This isn't to say that firewalls aren't important, they are an important layer in security, just that they don't protect most systems from attacks because most attacks will pass through intentionally opened routes in the firewall. Why should an attacker overly care to hack a firewall when you've already opened a route? It's usually not worth the trouble and often requires a lot of specialist knowledge and the attack is much more likely to be noticed.
January 25, 2018 at 9:38 am
I wasn't specifying hardware, just a principle. I think that Steve's advice about heterogeneous systems is good, and you are looking for an application in your operation, security is a good one.
January 25, 2018 at 9:53 am
GeorgeCopeland - Thursday, January 25, 2018 9:38 AMI wasn't specifying hardware, just a principle. I think that Steve's advice about heterogeneous systems is good, and you are looking for an application in your operation, security is a good one.
OK, yes - heterogeneous systems are good. In fact, I nearly gave an example that with a, usually Windows, MS-SQL server sitting behind a hardware firewall, which is often Cisco but could be anything. The fact that it's not worth an attacker hacking the firewall device because it's too hard and not worth the effort does kind of imply that having different systems is a good thing. It doesn't help protect the open route but it does help protected the firewall 🙂 In a big IT shop IT staff should be encouraged to have a mixture of systems, partly for support reasons (hard to support a system you have no access to) but also because an attack against one shouldn't affect the others.
January 25, 2018 at 10:46 am
Never mind the issues stemming from shared physical architecture approaches to things like memory and CPU access, what about everything else that sits on top? All virtualization and all abstraction (Java, .NET) could be problematic at these higher levels as well. What about compatibility and portability? Imagine AMD if they hadn't made their platform compatible with intel procs back in the day?
I'm just not sure I understand the meaning of the second to last paragraph, particularly in relation to relational databases and other "types of systems used for data storage;" it makes it seem as if these vulnerabilities uniquely affect [them] or Microsoft products which is not the case. No one has ever argued that competition was bad except for monopoly proponents. Lack-of competition is not the root cause for these vulnerabilities.
This is the type of fundamental architectural bug that's is very worrisome. The race to be efficient, to copy what works from others, this leads to less innovation not more. I hope that this is a bit of a lesson that we do need separate architectures and approaches to computing problems, both in hardware and software. I love relational databases, but I'm glad that there are other types of systems being used for data storage. I think Windows works really well, but I like competition and think it's good that we have MacOS, Linux, and more.
January 25, 2018 at 8:53 pm
t.franz - Thursday, January 25, 2018 1:40 AMHeterogenity may sound nice, but if I have x different systems, it multiplies the number of vulnerabilities too. And of course I would need staff, which knows this systems very good and is able to test / patch them.
Not necessarily. The vulnerabilities are not evenly shared across systems or architectures.
January 25, 2018 at 8:56 pm
n.ryan - Thursday, January 25, 2018 9:53 AMGeorgeCopeland - Thursday, January 25, 2018 9:38 AMI wasn't specifying hardware, just a principle. I think that Steve's advice about heterogeneous systems is good, and you are looking for an application in your operation, security is a good one.OK, yes - heterogeneous systems are good. In fact, I nearly gave an example that with a, usually Windows, MS-SQL server sitting behind a hardware firewall, which is often Cisco but could be anything. The fact that it's not worth an attacker hacking the firewall device because it's too hard and not worth the effort does kind of imply that having different systems is a good thing. It doesn't help protect the open route but it does help protected the firewall 🙂 In a big IT shop IT staff should be encouraged to have a mixture of systems, partly for support reasons (hard to support a system you have no access to) but also because an attack against one shouldn't affect the others.
To be clear, I'm not saying that you should implement different platforms/architectures in your shop, though it may not be an issue. I was saying that as an industry, we need different approaches to be tried, tested and worked on. Having everyone in the world run Linux, or Windows, or OSX, would be bad. I want some choice and options, though some interoperability for data exchange or application APIs.
January 25, 2018 at 9:00 pm
Ray Ayyelos - Thursday, January 25, 2018 10:46 AMLack-of competition is not the root cause for these vulnerabilities.
Not my intention, and if you read that, perhaps I poorly explained things. We do have different architectures and OSes that aren't compatible across chips, but allow choice. I mourn Unix falling away and not competing with Linux. I am glad we have different architectures like RaspberryPi, ARM, etc. that compete with Intel/AMD. I would hope there are different approaches to solving problems tried that will give us options if we find issues.
I didn't say, nor do I think I implied that using the same approach was the cause of a vulnerability.
January 29, 2018 at 11:17 am
From what I've read about Spectre/Meltdown, the problem is in the chip architecture; yes, Windows is impacted, but Linux and Mac OS could also be impacted.
Viewing 15 posts - 1 through 15 (of 16 total)
You must be logged in to reply to this topic. Login to reply