Agreed. This seems to be more of an official non-classified download repository than anything else. If I were in a small business and called their samba share that had the install images of Office, Acrobat, and other licensed packages for internal use an "app store", I'd be looked at by their IT people like I was some troll or pirate.
To me, a true "app store" is something like Apple's offering, Handango, Digital River, or a place where one looks through a catalog and either downloads a demo, or pays a license fee, then gets an executable to download.
There are some things I'd like to see the USG do though, if they are offering a large repository like this for internal use. The first thing is to PGP or gpg sign everything on the store so if it gets tampered with, one can find the app that has no or an invalid signature. (I'd also like to see Authenticode signing on Windows installs, and gpg package signing on BSD/RedHat/debian as another method that is transparent to the user, but will alert them if something is not right.)
I have not heard of any catastrophic data losses firsthand, but I don't like my data stored in a vendor specific format I couldn't dig out by plugging the component drives into another machine.
If you are a homebrew type, you might consider your favorite OS of choice [1] that can do software RAID, building yourself a generic server level PC, and use that for your backups. This way, when you need more drives, you can go to external SATA frames.
[1]: Almost all UNIX variants support RAID 5, Linux supports RAID 6 (two drives as parity), and of course, some BSDs and Solaris support ZFS for RAID-Z. Windows Server 2000, Windows Server 2003, Windows Server 2008, and Windows Server 2008R2 support RAID 5.
I almost see a market niche for an archival cloud provider with a lot less of a price premium than Amazon's cloud. The caveat would be that the customer is not getting instant access to data that is archived off. The archives are restorable in nowhere near real time (as the server has to retrieve the customer's info from tape and copy it to a drive array for transfer back).
This would be similar to Mozy or Carbonite, but the data would persist indefinitely once copied, instead of having old versions of it evaporate in a number of days.
This way, someone can play a cloud provider "X" amount to store some data permanently, copy it up, and essentially forget about it until its needed for an audit. Of course, pricing will need to be done right because there wouldn't be a constant income stream coming in. Perhaps a yearly maintainence fee would cover this. Pricing for volume of data would have to be included too, so people don't just shove the contents of every single computer up every several weeks for a constant price.
Backups for UNIX, backups for Windows, and backups all across the board almost require different solutions.
For an enterprise "catch all" solution, I'd go with TSM, Backup Exec, or Networker. These programs can pretty much back up anything that has a CPU, although you will be paying for that privilege.
If I were in an AIX environment, I'd use sysback for local machine backups and backups to a remote server.
If I were in a general UNIX environment, I'd use bru (it used to be licensed with IRIX, and has been around so long, it works without issue with any UNIX variant.) Of course, there are other solutions that work just as well, both freeware, and commercial.
If I were in a solidly Windows environment, I'd use Retrospect, or Backup Exec. Both are good utilities and support synthetic full backups so you don't need to worry about a full/differential/incremental schedule.
If I were in a completely mixed environment, I'd consider Retrospect (it can back up a few UNIX variants as well as Macs), Backup Exec, or an enterprise level utility that can back up virtually anything.
Please note, these are all commercial solutions. Bacula, Amanda, tar over ssh, rsync, and many others can work just as well, and likely will be a lot lighter on the pocketbook. However, for a business, some enterprise features like copying media sets, or backing up a database while it is online to tape or other media for offsite storage may be something to consider for maximum protection.
The key is figuring out what you need for restores. A backup system that is ideal for a bare metal restore may be a bit clunky if you have a machine with a stock Ubuntu config and just a few documents in your home directory. However, having 12 terabytes on Mozy, and needing to reinstall box from scratch that has custom apps with funky license keys would be a hair puller. Best thing is to use some method of backups for "oh crap" bare metal stuff, then an offsite service just in case you lose your backups at that location.
Figure out your scenario too. Are multiple Drobos good enough, or do you need offsite storage in case the facility is flooded? Is tape an option? Tape is notoriously expensive per drive, but is very economical once you start using multiple cartridges. Can you get away with plugging in external USB/SATA/IEEE 1394 hard disks, backing to them, then plopping them in the Iron Mountain tub?
Remote storage at a provider like Backblace, Mozy, or Carbonite is a good tertiary level backup, just in case your site goes down, but you are limited by your Internet pipe. A full restore of terabytes of videos through a typical business Internet connection will take a long time, perhaps days. Of course, one could order a hard disk or several from the backup company, but then you are stuck waiting for the data to physically arrive.
Remote storage is one solution, but before that, you have to have local ones in place for a faster recovery should a disaster happen. The first line of defense against hard disk stuff is RAID. The second line of defense would be a decent tape drive, a tape rotation, and offsite capabilities. This way, if you lose everything on your RAID (malware or a blackhat formats the volume), you can stuff in a tape, sit on pins and needles for a couple hours, and get your stuff back, perhaps back a day or two.
For a number of machines, the best thing to have would be a backup server with a large array and D2D2T (disk to disk to tape) capabilities so you can do fast backups through the network (or perhaps through a dedicate backup fabric), then when you can, copy them to the tapes for offline storage and the tub to Iron Mountain.
Of course, virtually all distributed backup utilities support encryption. Use it. Even if it is just movies.
On the higher end of the paranoia scale, if the chip fabbing equipment is from offshore, it can introduce its own items. There is always the exercise for the reader about having hidden code in a compiler introducing compromised code into anything it compiles, the ultimate rootkit (in theory) scenario.
There are places where the cost to reach a light bulb to change it is prohibitive. It could be theater marquee lights, lights atop a vaulted ceiling, or places behind a recessed opening that takes a lot of disassembly to get to. So even though $40 might be expensive up front, not having to set up scaffolding 30-40 feet up to get to some fixtures is worth it to some.
I know Apple has on their website a security document stating that all their computers have the ability to have the camera, wireless network, or Bluetooth antenna pulled for environments where devices like that are forbidden. Pretty much it states to take the device into an Apple facility and they do the rest.
Not sure if this is available for the Nano though.
There is one feature that I wished netbooks had, but it might change the price point unfavorably:
I would like to see netbooks come with a TPM chip, and BitLocker, or some FDE security that allows for booting without requiring a volume password unless the TPM is tampered with.
Since netbooks are portable and cheap, what comes with that is the fact that they are easily stolen and resold. Password security (TrueCrypt, RedHat/Ubuntu loopback mountings, and to a lesser extent FileVault) does go a long way, but in general, people are using netbooks in a public environment (coffee shops, the pit at metal concerts, airports). This attracts shoulder surfers that might be able to see one's password when the machine is booted up or kicked out of suspend mode. With protection that isn't just relying on the security of that password, a thief may be able to format the system disk (be it a flash mass storage device or a hard drive), but will not be able to access anything other than the contents of the laptop that it had while on.
This mechanism also gives the laptops owner to have more than just a password for protection. One can require a USB flash drive be inserted and/or a PIN before the laptop will boot. This way, if it gets stolen, and the former owner of the laptop still has possession of the USB flash drive, there is no real way a thief could obtain the contents of the laptop.
Even though netbooks are small and inexpensive, their data on them is just as valuable to both the owner of the machine and to second level thieves (fences) as the data on a regular laptop. Security is just as important on a person's $300 netbook as it is on their $5000 Thinkpad.
The only real way I see this possible would be a Steam-like service run by a large, neutral organization like ISO, IEEE, or perhaps a government agency that can guarentee privacy safeguards (perhaps an extension to NIST). They would have a highly fault-tolerant system of either storing anonymous IDs, or perhaps linking them to someone's user account. Then, if someone needed to redownload/reinstall on a new comp, all it would take would be firing up the client and clicking "download".
Of course there are a lot of holes in this, from who knows what data about people, to what (if any) DRM mechanisms would be used, and so on. Pretty much the same issues that people have concerns about with Steam and GameTap, but on a larger scale, because of the concern about longevity and accessibility to licensed content. Valve can shut Steam down tomorrow and walk off. A clearinghouse like this would have to keep stuff available forever.
I am willing to predict that it is only a matter of time before releasing "late betas" will start appearing on consoles too. The main reason is that newer consoles have persistant storage, and a way to store patches in game executables. Once console game writers realize that the DVD that is stamped and sent out to their customer base doesn't have to be a perfect release version (as it used to be on previous gen consoles with not enough storage for executable diffs), the quality of releases will most likely drop marketedly.
I'm hoping that this won't happen, but time is money when it comes to a release date, and pushing a game back for even a show-stopper bug can cost a company and its publisher a lot of cash. So, I'm sure that given the choice of angering stores by a missed deadline or shipping now and patching later (if ever), companies will take the former option.
That's how ID software got on the map. A shareware game engine and one level, then for the full game, one paid for it.
I can see something like this working. One downloads the core game engine and it comes with maybe 5-10 hours of play. That is definitely enough time for someone to tell that they like the game and plunk down cash for it, or uninstall it. If they like it, they register, and get an executable which provides the ability to create content, as well as the engine for downloading the rest of the game's content.
Even better, doing a framework like this makes expansions really easy to sell.
Of course, one can buy the boxed game and install everything without requiring a download of additional content, but if someone plays a game enough and the price is reasonable, in all likelyhood they will pay the cash to have the rest of the game immediately downloaded as opposed to the relative time and trouble to dig up a non-Trojanized version of the game, uninstall the old game and install the pirated one.
If game companies are concerned about the sales figures about the first few days, then they can have their DRM for the first couple weeks.
Then push out a patch that dumps the DRM, about a month after release. This way, legitimate users can still play their game and not have deal with hostile "user experience" after a couple weeks. People who just don't like DRM on principle can buy the game after the patch comes out and not worry about if a video card change out is going to blow out their activation limit.
The only caveat is that is costs a lot of cash to license SecuROM or Rovi's DRM system, and bean counters will ask about why using it only for it to exist for a couple weeks. However by having it there for the critical first weeks, the licensed DRM system has done its job, and from there on out, it is a liability as opposed to an asset.
I have used the Thinapp version of Firefox (www.thindownload.com) myself. It can store the info from Firefox on a USB flash drive, or in the \Users\blah\AppData\Roaming\thinstall directory. The nice thing is it stores all its stuff there, Registry changes, and the like, and it never needs Administrator access. One zap of that folder and a run of cypher/w, and my preferences are completely off a system.
Caveat: Use at your untrusted executables at your own risk.
Of course, you can license EMC's product and make versions of your own (which might be a very good idea to do in a company which has to have system images that remain locked down, with multiple levels of authorization to allow any type of packages to be installed.)
I'd like to hear more about this technology myself. I've heard of this in the lab, but never at a point where it would be a sustainable, production energy generation tool.
I wouldn't say bring it on. They will be throwing big money at it, and I'm almost sure it will have some mechanism to autoupdate or perhaps ban non-compliant devices from being "authorized" to play files, similar to how modchipped Xboxes get permanently banned off of XBL. If a device gets permanently "cracked" in a way where it can't be updated, a new line of models come out, and newer music content will not be able to play on those.
I can forsee a scenario where it ends up like BD+, where its an arms race. The BD+ guys patch and invalidate all cracks, then 2-3 months later, that gets patched and the cycle starts all over again. There are even areas where it has been years and stuff has not been thoroughly cracked, such as satellite. There are probably people who have managed to crack it, but it won't be something the public will have access to.
And DPP already has sort of happened. I remember the early Sony ATRAC players that required checking in and out of music. More than three copies of a song checked out. Rerip, or rebuy. Even Sony realized that this started irritating consumes to no end and dumped OpenMG and SonicState for more generic MP3 capabilities. This also reminds me of SDMI, an initiative that tried to get all MP3 players to organize on some DRM standard with both security and watermarking. However, it got broken wide open by some researchers (even with the DMCA hammer ready to pulp them), and Apple didn't bother including any of the capabilities in the iPod, so when Apple grabbed the market, SDMI became pointless.
I doubt DPP, or "SDMI 3.0" is going to take off. First, if there is software, it will be broken. People will not buy MP3 players that add DRM to their existing libraries. Finally, if music companies ship DRM protected media that cannot be read by the CD standard (DVD-R, SD), people just won't buy it because it doesn't work with their existing stuff.
However, this could end up being reality. It would take record labels releasing their stuff on a new copy-protected format, digital audio player makers (Microsoft, Apple, Creative) to support the format, crack resistant software on PCs and Macs, and a network infrastructure to quickly ban cracked PCs/devices. It can be done, but it would be having to rip up a large established base.
True. If you forget the "seed" password, then you are cooked. However, if you have the PW and application (which can just be a script that does 'echo "mypasswordmychallengephrase"|md5sum"' (where mypassword is your core password, and mychallengephrase is your challenge question), and paste in the hash), you can pretty much enter anything for the challenge questions and it will be unguessable to an attacker.
I keep having people hit up my Gmail account with lost password queries, usually about 3-4 times a week. Even though those mails are routed to a junk mailbox designed for that, all it would take is accidently clicking on one of the recovery links to lose control of the account.
I do wish Gmail would have an option to require someone trying to obtain a gmail password to pass the challenge/response questions before it sends a link to recover. This isn't foolproof, but it will keep Joe Skiddy from being able to blanket gmail with PW requests in hopes someone clicks on a link.
One system I've thought of for security questions requires a simple app on a cellphone. App asks for a password, then when you type in what it wants for a security question, it SHA-256 hashes the question + the password [1], drops all but the first x characters, and then you use the x (10+ depending on the system, preferably 15-20) amount of characters in the result as the answer.
This way, its easy to have your answer to security questions, you can enter almost anything in for the question, but yet nobody would be able to get the answer without brute forcing your password on your cellphone app.
[1]: For additional security, the program can hash stuff a large number of times to help combat brute forcing.
I wonder if in a case like this, the ex can make up where he/she found the info, to hide the real source. For example, it could be claimed that the passwords were gleaned through a keylogger or a hidden camera. Unless the other attorney knows what questions to ask, there would not
What I'd like to see would be more ability to use a standardized keyfob (such as RSA's SecurID), a smart card that has one's client certificate, or perhaps both in one device like the Aladdin eToken NG-OTP. Combine this with some type of decentralized but usable authentication system like OpenID, and this would go a long way to making bad or guessed passwords a thing of the past.
Smart cards go a long way to ease authentication hassles, but they bring their own issues, such as card lockouts due to too many failed PIN attempts, lost/stolen/accidently microwaved cards, user training, to malware which captures the PIN on a compromised computer then if the card is still inserted, uses it for its own bad stuff.
In a perfect world, you would have 512 bit security provided you used different keys for both cyphers. However, there are a number of attacks like birthday and meet in the middle attacks which would reduce the effective amount of bits from 512 down, such as how double-DES is pointless instead of triple DES. They may not be as effective because of the use of two different algorithms, but they may be possible.
If the attacker knows anything about the text between the two algorithms, the 512 bits drops to 257 (2*256) bit security. So, it doesn't hurt to assume worst case.
The point of multiple algorithms is not to have a very large bit size. Its to ensure the data remains protected should an algorithm get broken.
All you need is a link fast enough to transmit the key for bulk encryption, then dump the rest of the data via a normal pipe encrypted via either AES, or a cascade [1]. Quantum computers are not useful against symmetric encryption such as DES, AES, Serpent, or IDEA.
[1]: If you have two algorithms each 256 bit, cascading them only gives you 257 bit security (256 + 256), but the reason one would do this is to deal with a possible algorithm breakage of either algorithm. Say some algorithm that is 256 bits is able to be brute forced in 48, and you cascade it with one that hasn't been broken. Your data is still safe.
We need to do not just one strategy, but a number. Until fusion is sustainable and usable at the gigawatt level, it will take a mixture of various energy sources. The key is to move from fossil fuels to a carbon neutral civilization as fast as we can.
Short term, technologies like solar + coal, and combining sources are important. These allow companies whose existance is on coal or fossil fuels to transition away from them, but still keep their stock listing.
Medium term, it will take infrastructure changes. If we are going to go to electric cars, we will need the grid to be able to handle the power needed. This means more wires and substations, but it means more R&D to develop better ways to time vehicle charging to prevent brownout-causing spikes. Perhaps, the vehicles charging on the grid may give a bigger benefit overall than a hiderance if one could get them to output power on the short term, should a power emergency arise. It might not be the best for the individual driver, but it would prevent an area from having a brownout or a blackout.
Long term, it will take R&D. This is something US companies are scared off now. In the 1980s and 1990s, product liability lawsuits drove companies to spend as little as they can on anything new for fear that they didn't CYA enough. These days, if a company doesn't get their quarterly results in and charges off some earnings for R&D for future stuff, shareholders may start a class action, so companies cannot spend too much on R&D, only trying to keep the sales numbers as good if not better than last quarter. R&D comes from startups which are then bought up by large companies because of this. The ever-elusive fusion would be the ultimate problem solver. This is where the lottery is. First company or organization that can come up with a sustainable reactor that can be mass produced, even if it generates energy on a small scale like kilowatts or megawatts will end up being the future IBM, GE, Aramco, or Microsoft.
If I were to place money into technologies, I'd always try to keep cash going for fusion just for the payoff. However for the medium term payout, it would be solar as the secondary source of energy, and nuclear power as a primary. Solar can help with peak surges, and nuclear is the best energy source per real estate square foot we have.
Short term, anything that slows down the rate CO2 gets belched into the air is good. Supporting not just efforts like combining solar and existing technologies, but efforts to be able to reforest large tracts of land (trees are very good carbon sinks) are stuff to do now.
My concern is that having cloud providers store companys' data means that it is a bigger target for thieves and blackhats than decentralized storage. There are a lot of eggs in that cloud provider's basket.
Even legit uses, all it would take would be a bankruptcy or sale of the cloud assets, and even the most well written privacy and TOS contract will go out the window, perhaps allowing the buyer complete and unrestricted use of the information. Rival company to you or an ally? Their trade secrets are now yours. Someone offshore wants to know the exact chemical process that is highly confidental because it gives 50% more yield of methane from cow farts than any competitors? They got it.
This doesn't mean cloud storage is useless. This just means that companies need an encryption layer before a single bit hits the cloud. This could mean using EncFS or a CFS layer for filesystem layer encryption, or a device that sits on the LAN and virtualizes the cloud storage. Local boxes pass it the data to stick on the cloud, the appliance encrypts it and does the remote read/writes.
Cloud sharing of CPU, is also risky. One never knows if someone would take a snapshot of running processes serverside, then copy that somewhere for analysis for anything they could figure out. This might be useful if someone is needing a bunch of webservers to mirror data for a planned capacity spike, to mirror public information. However for anything else, it treads on risky territory, especially if the data is something that Sarbanes Oxley or HIPAA cover.
Cloud computing is not something to be abolutely shunned, but it isn't something to be embraced completely. It is another generation of client server stuff on a higher computing layer (web layer, where with Javastations and some Xstations Java was the app layer, VT100 terminals were lower than that.)
What would make a company a killing would be a hardware appliance, or a software solution that would allow encryption and enterprise key management, with a passthrough interface between it to a commercial disk cloud.
Slashdot Mirror
Agreed. This seems to be more of an official non-classified download repository than anything else. If I were in a small business and called their samba share that had the install images of Office, Acrobat, and other licensed packages for internal use an "app store", I'd be looked at by their IT people like I was some troll or pirate.
To me, a true "app store" is something like Apple's offering, Handango, Digital River, or a place where one looks through a catalog and either downloads a demo, or pays a license fee, then gets an executable to download.
There are some things I'd like to see the USG do though, if they are offering a large repository like this for internal use. The first thing is to PGP or gpg sign everything on the store so if it gets tampered with, one can find the app that has no or an invalid signature. (I'd also like to see Authenticode signing on Windows installs, and gpg package signing on BSD/RedHat/debian as another method that is transparent to the user, but will alert them if something is not right.)
I have not heard of any catastrophic data losses firsthand, but I don't like my data stored in a vendor specific format I couldn't dig out by plugging the component drives into another machine.
If you are a homebrew type, you might consider your favorite OS of choice [1] that can do software RAID, building yourself a generic server level PC, and use that for your backups. This way, when you need more drives, you can go to external SATA frames.
[1]: Almost all UNIX variants support RAID 5, Linux supports RAID 6 (two drives as parity), and of course, some BSDs and Solaris support ZFS for RAID-Z. Windows Server 2000, Windows Server 2003, Windows Server 2008, and Windows Server 2008R2 support RAID 5.
I almost see a market niche for an archival cloud provider with a lot less of a price premium than Amazon's cloud. The caveat would be that the customer is not getting instant access to data that is archived off. The archives are restorable in nowhere near real time (as the server has to retrieve the customer's info from tape and copy it to a drive array for transfer back).
This would be similar to Mozy or Carbonite, but the data would persist indefinitely once copied, instead of having old versions of it evaporate in a number of days.
This way, someone can play a cloud provider "X" amount to store some data permanently, copy it up, and essentially forget about it until its needed for an audit. Of course, pricing will need to be done right because there wouldn't be a constant income stream coming in. Perhaps a yearly maintainence fee would cover this. Pricing for volume of data would have to be included too, so people don't just shove the contents of every single computer up every several weeks for a constant price.
Backups for UNIX, backups for Windows, and backups all across the board almost require different solutions.
For an enterprise "catch all" solution, I'd go with TSM, Backup Exec, or Networker. These programs can pretty much back up anything that has a CPU, although you will be paying for that privilege.
If I were in an AIX environment, I'd use sysback for local machine backups and backups to a remote server.
If I were in a general UNIX environment, I'd use bru (it used to be licensed with IRIX, and has been around so long, it works without issue with any UNIX variant.) Of course, there are other solutions that work just as well, both freeware, and commercial.
If I were in a solidly Windows environment, I'd use Retrospect, or Backup Exec. Both are good utilities and support synthetic full backups so you don't need to worry about a full/differential/incremental schedule.
If I were in a completely mixed environment, I'd consider Retrospect (it can back up a few UNIX variants as well as Macs), Backup Exec, or an enterprise level utility that can back up virtually anything.
Please note, these are all commercial solutions. Bacula, Amanda, tar over ssh, rsync, and many others can work just as well, and likely will be a lot lighter on the pocketbook. However, for a business, some enterprise features like copying media sets, or backing up a database while it is online to tape or other media for offsite storage may be something to consider for maximum protection.
The key is figuring out what you need for restores. A backup system that is ideal for a bare metal restore may be a bit clunky if you have a machine with a stock Ubuntu config and just a few documents in your home directory. However, having 12 terabytes on Mozy, and needing to reinstall box from scratch that has custom apps with funky license keys would be a hair puller. Best thing is to use some method of backups for "oh crap" bare metal stuff, then an offsite service just in case you lose your backups at that location.
Figure out your scenario too. Are multiple Drobos good enough, or do you need offsite storage in case the facility is flooded? Is tape an option? Tape is notoriously expensive per drive, but is very economical once you start using multiple cartridges. Can you get away with plugging in external USB/SATA/IEEE 1394 hard disks, backing to them, then plopping them in the Iron Mountain tub?
Remote storage at a provider like Backblace, Mozy, or Carbonite is a good tertiary level backup, just in case your site goes down, but you are limited by your Internet pipe. A full restore of terabytes of videos through a typical business Internet connection will take a long time, perhaps days. Of course, one could order a hard disk or several from the backup company, but then you are stuck waiting for the data to physically arrive.
Remote storage is one solution, but before that, you have to have local ones in place for a faster recovery should a disaster happen. The first line of defense against hard disk stuff is RAID. The second line of defense would be a decent tape drive, a tape rotation, and offsite capabilities. This way, if you lose everything on your RAID (malware or a blackhat formats the volume), you can stuff in a tape, sit on pins and needles for a couple hours, and get your stuff back, perhaps back a day or two.
For a number of machines, the best thing to have would be a backup server with a large array and D2D2T (disk to disk to tape) capabilities so you can do fast backups through the network (or perhaps through a dedicate backup fabric), then when you can, copy them to the tapes for offline storage and the tub to Iron Mountain.
Of course, virtually all distributed backup utilities support encryption. Use it. Even if it is just movies.
On the higher end of the paranoia scale, if the chip fabbing equipment is from offshore, it can introduce its own items. There is always the exercise for the reader about having hidden code in a compiler introducing compromised code into anything it compiles, the ultimate rootkit (in theory) scenario.
There are places where the cost to reach a light bulb to change it is prohibitive. It could be theater marquee lights, lights atop a vaulted ceiling, or places behind a recessed opening that takes a lot of disassembly to get to. So even though $40 might be expensive up front, not having to set up scaffolding 30-40 feet up to get to some fixtures is worth it to some.
I know Apple has on their website a security document stating that all their computers have the ability to have the camera, wireless network, or Bluetooth antenna pulled for environments where devices like that are forbidden. Pretty much it states to take the device into an Apple facility and they do the rest.
Not sure if this is available for the Nano though.
There is one feature that I wished netbooks had, but it might change the price point unfavorably:
I would like to see netbooks come with a TPM chip, and BitLocker, or some FDE security that allows for booting without requiring a volume password unless the TPM is tampered with.
Since netbooks are portable and cheap, what comes with that is the fact that they are easily stolen and resold. Password security (TrueCrypt, RedHat/Ubuntu loopback mountings, and to a lesser extent FileVault) does go a long way, but in general, people are using netbooks in a public environment (coffee shops, the pit at metal concerts, airports). This attracts shoulder surfers that might be able to see one's password when the machine is booted up or kicked out of suspend mode. With protection that isn't just relying on the security of that password, a thief may be able to format the system disk (be it a flash mass storage device or a hard drive), but will not be able to access anything other than the contents of the laptop that it had while on.
This mechanism also gives the laptops owner to have more than just a password for protection. One can require a USB flash drive be inserted and/or a PIN before the laptop will boot. This way, if it gets stolen, and the former owner of the laptop still has possession of the USB flash drive, there is no real way a thief could obtain the contents of the laptop.
Even though netbooks are small and inexpensive, their data on them is just as valuable to both the owner of the machine and to second level thieves (fences) as the data on a regular laptop. Security is just as important on a person's $300 netbook as it is on their $5000 Thinkpad.
The only real way I see this possible would be a Steam-like service run by a large, neutral organization like ISO, IEEE, or perhaps a government agency that can guarentee privacy safeguards (perhaps an extension to NIST). They would have a highly fault-tolerant system of either storing anonymous IDs, or perhaps linking them to someone's user account. Then, if someone needed to redownload/reinstall on a new comp, all it would take would be firing up the client and clicking "download".
Of course there are a lot of holes in this, from who knows what data about people, to what (if any) DRM mechanisms would be used, and so on. Pretty much the same issues that people have concerns about with Steam and GameTap, but on a larger scale, because of the concern about longevity and accessibility to licensed content. Valve can shut Steam down tomorrow and walk off. A clearinghouse like this would have to keep stuff available forever.
I am willing to predict that it is only a matter of time before releasing "late betas" will start appearing on consoles too. The main reason is that newer consoles have persistant storage, and a way to store patches in game executables. Once console game writers realize that the DVD that is stamped and sent out to their customer base doesn't have to be a perfect release version (as it used to be on previous gen consoles with not enough storage for executable diffs), the quality of releases will most likely drop marketedly.
I'm hoping that this won't happen, but time is money when it comes to a release date, and pushing a game back for even a show-stopper bug can cost a company and its publisher a lot of cash. So, I'm sure that given the choice of angering stores by a missed deadline or shipping now and patching later (if ever), companies will take the former option.
That's how ID software got on the map. A shareware game engine and one level, then for the full game, one paid for it.
I can see something like this working. One downloads the core game engine and it comes with maybe 5-10 hours of play. That is definitely enough time for someone to tell that they like the game and plunk down cash for it, or uninstall it. If they like it, they register, and get an executable which provides the ability to create content, as well as the engine for downloading the rest of the game's content.
Even better, doing a framework like this makes expansions really easy to sell.
Of course, one can buy the boxed game and install everything without requiring a download of additional content, but if someone plays a game enough and the price is reasonable, in all likelyhood they will pay the cash to have the rest of the game immediately downloaded as opposed to the relative time and trouble to dig up a non-Trojanized version of the game, uninstall the old game and install the pirated one.
If game companies are concerned about the sales figures about the first few days, then they can have their DRM for the first couple weeks.
Then push out a patch that dumps the DRM, about a month after release. This way, legitimate users can still play their game and not have deal with hostile "user experience" after a couple weeks. People who just don't like DRM on principle can buy the game after the patch comes out and not worry about if a video card change out is going to blow out their activation limit.
The only caveat is that is costs a lot of cash to license SecuROM or Rovi's DRM system, and bean counters will ask about why using it only for it to exist for a couple weeks. However by having it there for the critical first weeks, the licensed DRM system has done its job, and from there on out, it is a liability as opposed to an asset.
I have used the Thinapp version of Firefox (www.thindownload.com) myself. It can store the info from Firefox on a USB flash drive, or in the \Users\blah\AppData\Roaming\thinstall directory. The nice thing is it stores all its stuff there, Registry changes, and the like, and it never needs Administrator access. One zap of that folder and a run of cypher /w, and my preferences are completely off a system.
Caveat: Use at your untrusted executables at your own risk.
Of course, you can license EMC's product and make versions of your own (which might be a very good idea to do in a company which has to have system images that remain locked down, with multiple levels of authorization to allow any type of packages to be installed.)
I'd like to hear more about this technology myself. I've heard of this in the lab, but never at a point where it would be a sustainable, production energy generation tool.
I wouldn't say bring it on. They will be throwing big money at it, and I'm almost sure it will have some mechanism to autoupdate or perhaps ban non-compliant devices from being "authorized" to play files, similar to how modchipped Xboxes get permanently banned off of XBL. If a device gets permanently "cracked" in a way where it can't be updated, a new line of models come out, and newer music content will not be able to play on those.
I can forsee a scenario where it ends up like BD+, where its an arms race. The BD+ guys patch and invalidate all cracks, then 2-3 months later, that gets patched and the cycle starts all over again. There are even areas where it has been years and stuff has not been thoroughly cracked, such as satellite. There are probably people who have managed to crack it, but it won't be something the public will have access to.
And DPP already has sort of happened. I remember the early Sony ATRAC players that required checking in and out of music. More than three copies of a song checked out. Rerip, or rebuy. Even Sony realized that this started irritating consumes to no end and dumped OpenMG and SonicState for more generic MP3 capabilities. This also reminds me of SDMI, an initiative that tried to get all MP3 players to organize on some DRM standard with both security and watermarking. However, it got broken wide open by some researchers (even with the DMCA hammer ready to pulp them), and Apple didn't bother including any of the capabilities in the iPod, so when Apple grabbed the market, SDMI became pointless.
I doubt DPP, or "SDMI 3.0" is going to take off. First, if there is software, it will be broken. People will not buy MP3 players that add DRM to their existing libraries. Finally, if music companies ship DRM protected media that cannot be read by the CD standard (DVD-R, SD), people just won't buy it because it doesn't work with their existing stuff.
However, this could end up being reality. It would take record labels releasing their stuff on a new copy-protected format, digital audio player makers (Microsoft, Apple, Creative) to support the format, crack resistant software on PCs and Macs, and a network infrastructure to quickly ban cracked PCs/devices. It can be done, but it would be having to rip up a large established base.
True. If you forget the "seed" password, then you are cooked. However, if you have the PW and application (which can just be a script that does 'echo "mypasswordmychallengephrase"|md5sum"' (where mypassword is your core password, and mychallengephrase is your challenge question), and paste in the hash), you can pretty much enter anything for the challenge questions and it will be unguessable to an attacker.
I keep having people hit up my Gmail account with lost password queries, usually about 3-4 times a week. Even though those mails are routed to a junk mailbox designed for that, all it would take is accidently clicking on one of the recovery links to lose control of the account.
I do wish Gmail would have an option to require someone trying to obtain a gmail password to pass the challenge/response questions before it sends a link to recover. This isn't foolproof, but it will keep Joe Skiddy from being able to blanket gmail with PW requests in hopes someone clicks on a link.
One system I've thought of for security questions requires a simple app on a cellphone. App asks for a password, then when you type in what it wants for a security question, it SHA-256 hashes the question + the password [1], drops all but the first x characters, and then you use the x (10+ depending on the system, preferably 15-20) amount of characters in the result as the answer.
This way, its easy to have your answer to security questions, you can enter almost anything in for the question, but yet nobody would be able to get the answer without brute forcing your password on your cellphone app.
[1]: For additional security, the program can hash stuff a large number of times to help combat brute forcing.
I wonder if in a case like this, the ex can make up where he/she found the info, to hide the real source. For example, it could be claimed that the passwords were gleaned through a keylogger or a hidden camera. Unless the other attorney knows what questions to ask, there would not
What I'd like to see would be more ability to use a standardized keyfob (such as RSA's SecurID), a smart card that has one's client certificate, or perhaps both in one device like the Aladdin eToken NG-OTP. Combine this with some type of decentralized but usable authentication system like OpenID, and this would go a long way to making bad or guessed passwords a thing of the past.
Smart cards go a long way to ease authentication hassles, but they bring their own issues, such as card lockouts due to too many failed PIN attempts, lost/stolen/accidently microwaved cards, user training, to malware which captures the PIN on a compromised computer then if the card is still inserted, uses it for its own bad stuff.
In a perfect world, you would have 512 bit security provided you used different keys for both cyphers. However, there are a number of attacks like birthday and meet in the middle attacks which would reduce the effective amount of bits from 512 down, such as how double-DES is pointless instead of triple DES. They may not be as effective because of the use of two different algorithms, but they may be possible.
If the attacker knows anything about the text between the two algorithms, the 512 bits drops to 257 (2*256) bit security. So, it doesn't hurt to assume worst case.
The point of multiple algorithms is not to have a very large bit size. Its to ensure the data remains protected should an algorithm get broken.
All you need is a link fast enough to transmit the key for bulk encryption, then dump the rest of the data via a normal pipe encrypted via either AES, or a cascade [1]. Quantum computers are not useful against symmetric encryption such as DES, AES, Serpent, or IDEA.
[1]: If you have two algorithms each 256 bit, cascading them only gives you 257 bit security (256 + 256), but the reason one would do this is to deal with a possible algorithm breakage of either algorithm. Say some algorithm that is 256 bits is able to be brute forced in 48, and you cascade it with one that hasn't been broken. Your data is still safe.
We need to do not just one strategy, but a number. Until fusion is sustainable and usable at the gigawatt level, it will take a mixture of various energy sources. The key is to move from fossil fuels to a carbon neutral civilization as fast as we can.
Short term, technologies like solar + coal, and combining sources are important. These allow companies whose existance is on coal or fossil fuels to transition away from them, but still keep their stock listing.
Medium term, it will take infrastructure changes. If we are going to go to electric cars, we will need the grid to be able to handle the power needed. This means more wires and substations, but it means more R&D to develop better ways to time vehicle charging to prevent brownout-causing spikes. Perhaps, the vehicles charging on the grid may give a bigger benefit overall than a hiderance if one could get them to output power on the short term, should a power emergency arise. It might not be the best for the individual driver, but it would prevent an area from having a brownout or a blackout.
Long term, it will take R&D. This is something US companies are scared off now. In the 1980s and 1990s, product liability lawsuits drove companies to spend as little as they can on anything new for fear that they didn't CYA enough. These days, if a company doesn't get their quarterly results in and charges off some earnings for R&D for future stuff, shareholders may start a class action, so companies cannot spend too much on R&D, only trying to keep the sales numbers as good if not better than last quarter. R&D comes from startups which are then bought up by large companies because of this. The ever-elusive fusion would be the ultimate problem solver. This is where the lottery is. First company or organization that can come up with a sustainable reactor that can be mass produced, even if it generates energy on a small scale like kilowatts or megawatts will end up being the future IBM, GE, Aramco, or Microsoft.
If I were to place money into technologies, I'd always try to keep cash going for fusion just for the payoff. However for the medium term payout, it would be solar as the secondary source of energy, and nuclear power as a primary. Solar can help with peak surges, and nuclear is the best energy source per real estate square foot we have.
Short term, anything that slows down the rate CO2 gets belched into the air is good. Supporting not just efforts like combining solar and existing technologies, but efforts to be able to reforest large tracts of land (trees are very good carbon sinks) are stuff to do now.
My concern is that having cloud providers store companys' data means that it is a bigger target for thieves and blackhats than decentralized storage. There are a lot of eggs in that cloud provider's basket.
Even legit uses, all it would take would be a bankruptcy or sale of the cloud assets, and even the most well written privacy and TOS contract will go out the window, perhaps allowing the buyer complete and unrestricted use of the information. Rival company to you or an ally? Their trade secrets are now yours. Someone offshore wants to know the exact chemical process that is highly confidental because it gives 50% more yield of methane from cow farts than any competitors? They got it.
This doesn't mean cloud storage is useless. This just means that companies need an encryption layer before a single bit hits the cloud. This could mean using EncFS or a CFS layer for filesystem layer encryption, or a device that sits on the LAN and virtualizes the cloud storage. Local boxes pass it the data to stick on the cloud, the appliance encrypts it and does the remote read/writes.
Cloud sharing of CPU, is also risky. One never knows if someone would take a snapshot of running processes serverside, then copy that somewhere for analysis for anything they could figure out. This might be useful if someone is needing a bunch of webservers to mirror data for a planned capacity spike, to mirror public information. However for anything else, it treads on risky territory, especially if the data is something that Sarbanes Oxley or HIPAA cover.
Cloud computing is not something to be abolutely shunned, but it isn't something to be embraced completely. It is another generation of client server stuff on a higher computing layer (web layer, where with Javastations and some Xstations Java was the app layer, VT100 terminals were lower than that.)
What would make a company a killing would be a hardware appliance, or a software solution that would allow encryption and enterprise key management, with a passthrough interface between it to a commercial disk cloud.