I live in Japan and work as an engineer and I can tell you, engineering work pays crap here. Starting pay out of university is barely over 200,000 yen/mo. You'd only do it in Japan if you really enjoyed the work.
And the quality of japanese engineers you do find is generally crap too. Every engineer in my department is a local-hire foreigner.
I've rescued a hijacked/16 a while back, cleaned it up and now have it as a souvenir (with the blessing of the original owner who is not using it globally). Fortunately the spammer who grabbed it wasn't very smart about it (contact info changed to obviously non-japanese info and the block was allocated through JPNIC).
Had to engage in a bit of a BGP war (deaggregated the block and announced it as 4 seperate/18s) but eventually got AT&T (AS7018), who the spammer had gotten to illegally announce the block, got the message and stopped.
One thing you always have to do is ALWAYS announce every block you have control over, even if you are not using it. There's a highly technical term for an unannounced block: TARGET.
For historical IP hijacking info, see completewhois.com (IP hijacking is by no means a new thing)
I'm working on adding IPv6 support to a server package written in java. Part of what it does is use a subnet mask to discriminate LAN and WAN clients. Can I still do this if everything is IPv6? Would I just have a different subnet mask?
Are there still LAN ranges like 192.168 or 10. in IPv6? There are link-local and site-local address ranges
Do I still have a DHCP server on my LAN? v6 has an autoconfigure mechanism that assigns addressing within a subnet based on an encoding of the mac address of the device (in a predictable manner to boot)
Actually, IPV6 in it's current form could make things a lot worse. At least NAT forces organizations to manage their internal address space and keeps some of the routing burden off our backbone. It also provides some extra security by keeping all those soft targets (client workstations) off the big bad Internet, even when people make a mess of their firewall.
Now think about that fact that IPV6 bumps up the address space 2^96 times. Imagine the burden that will place on routing tables. Imagine how many more nodes will pop up when you consider that people don't feel the need to hassle with NAT. Think about the routing overhead and the security nightmares that could result. ipv6 should actually be able to reduce routing table size by reducing the number of prefixes that need to be advertised. For instance, my employer advertises 2/17s, a/16 and a bunch of legacy/24s and we will probably be at least doubling that before the v4 address space runs out. Our single ipv6/32 allocation, even if it doesn't last until the heat-death of the universe, should at least suffice until the great disaster of 2038.
Fortunately the US considers renunciation of citizenship under these circumstances as being made under duress and routinely issues new passports to those who have done so.
Or is it they don't want to lose a source of tax revenue?
All the CCIE requires is one qualifier test to take the practical. You can get a CCIE without even holding a CCNA.
How well you passed the CCNP can give you a rough idea of how ready you were to pass the CCIE qualifier (which is by no means any kind of assurance you're even really prepared to pass the practical).
IMO, CCIE-level certifications are the only Cisco certs that mean anything because the practical is the only thing that comes close to testing a real-world crisis.
Actually, in Japan all 3 of the major carriers use SIM cards in their phones now. All the phones you buy through the carrier are locked though. You can buy WCDMA/GSM handsets direct from Nokia Japan that will work on both DoCoMo and Softbank however. (au is the odd man out here since they use CDMA2000 on their network instead of WCDMA)
Coincidentally, I have a W41CA myself. It was the smallest/lightest phone in the au lineup when I got it (the replacement model may also still be the lightest one available). All the reviews seem to miss the biggest feature - the penguin animations when you have the phone open.
Felica terminals that are not directly tied to the register. You ring up the correct value, it pops up on the register display. The cashier enters a higher value debit into the felica terminal. You plop your card down and are distracted by the cashier and miss the deduction which only shows up on the tiny terminal display for about a second.
The king of handling files for video editing was XFS/XLV on SGI iron. You could do guaranteed I/O throughput reservation from a filesystem/lv and the system would steal every bit of cpu/io possible to assure you meet the reservation requirements.
It was great technology, but was killed in the marketplace by cheap fast commodity hardware being "fast enough" that it wasn't needed any more.
At present ZFS will not let you add or subtract disks to a single raidz component. This is one of the big shortcomings to ZFS flexability at present.
What you can do to add capacity is to add another component, either raidz or mirror to your pool.
You can also replace individual drives in the raidz with larger ones, and after all drives are replaced, exporting and importing your zpool will make the additional space available.
Even at 2GB there are lots of issues with existing embedded devices, because that's where the change from 512-byte to 1k blocks takes place. As an example, most current mobile phones won't handle anything over 1GB because their hardware assumes 512-byte blocks and can't deal with anything different.
Actually caching for ISPs or any carrier where you don't know what your customers are doing is problematic. In the real world, people push all kinds of non-HTTP traffic over port 80 because it is the path of least resistance in getting by filters, and in many cases they do not encapsulate it inside an HTTP wrapper.
There's a case to be made for online application suites like this in an intranet context to enforce centralized document storage for compliance reasons (oh god, a SOX clone is coming to Japan - maybe its time to move to China).
Assume: Bandwidth costs $50/mbit/sec (a reasonable ballpark figure if you're buying transit in bulk) Client size is 2GB (Never played L2, but another guesstimate) Peak 95th percentile bandwidth usage of 5 times average The 500,000 users you just quoted
Conversely, a shift to IPv6 will also reduce the effectiveness of worms because the sparseness of address distribution makes portscanning orders of magnitude less effective. It's all a wash really.
We already use a bonded disposal company to send our media to its final repose. They come to our office with a metal case, which is locked after filling with DLT tapes, then tracked all the way to the disposal site where it is run through a crusher that turns it into bits no larger than 0.75cm on a side.
They've been really hard-pressed to keep up their service availability since the japanese personal information protection law came into effect a year and a half or so back.
WWV and other similar radio clocks (I use JJY in japan) are only 1 second resolution, which makes those sort of clocks not really suitable for seriously mission critical timing needs, even though they technically qualify as stratum 1.
At home I use a little JJY-clock built from a kit that cost me about US$40 and connects via RS-232 and is supported in the reference ntp implementation. Has a PIC, some RS-232 glue and not much more other than the antenna and VCO.
Slashdot Mirror
I call BS on this unless it changed since 1997.
1/2 is an either-or, never even heard of 3.
#4 is on the mark though.
I call BS on that.
I live in Japan and work as an engineer and I can tell you, engineering work pays crap here. Starting pay out of university is barely over 200,000 yen/mo. You'd only do it in Japan if you really enjoyed the work.
And the quality of japanese engineers you do find is generally crap too. Every engineer in my department is a local-hire foreigner.
I've rescued a hijacked /16 a while back, cleaned it up and now have it as a souvenir (with the blessing of the original owner who is not using it globally). Fortunately the spammer who grabbed it wasn't very smart about it (contact info changed to obviously non-japanese info and the block was allocated through JPNIC).
/18s) but eventually got AT&T (AS7018), who the spammer had gotten to illegally announce the block, got the message and stopped.
Had to engage in a bit of a BGP war (deaggregated the block and announced it as 4 seperate
One thing you always have to do is ALWAYS announce every block you have control over, even if you are not using it. There's a highly technical term for an unannounced block: TARGET.
For historical IP hijacking info, see completewhois.com (IP hijacking is by no means a new thing)
Among other reasons, because the iPhone is useless in non-GSM countries like Japan.
Anyone who has a public v4 address can route v6 using 6to4 at least as far as a proof-of-concept (using the anycast relay routers is rather slow).
http://6to4.version6.net/?lang=en_GB
Yamaha consumer routers have supported IPv6 for at least 5 years.
Are there still LAN ranges like 192.168 or 10. in IPv6? There are link-local and site-local address ranges Do I still have a DHCP server on my LAN? v6 has an autoconfigure mechanism that assigns addressing within a subnet based on an encoding of the mac address of the device (in a predictable manner to boot)
Now think about that fact that IPV6 bumps up the address space 2^96 times. Imagine the burden that will place on routing tables. Imagine how many more nodes will pop up when you consider that people don't feel the need to hassle with NAT. Think about the routing overhead and the security nightmares that could result. ipv6 should actually be able to reduce routing table size by reducing the number of prefixes that need to be advertised. For instance, my employer advertises 2
Fortunately the US considers renunciation of citizenship under these circumstances as being made under duress and routinely issues new passports to those who have done so.
Or is it they don't want to lose a source of tax revenue?
All the CCIE requires is one qualifier test to take the practical. You can get a CCIE without even holding a CCNA.
How well you passed the CCNP can give you a rough idea of how ready you were to pass the CCIE qualifier (which is by no means any kind of assurance you're even really prepared to pass the practical).
IMO, CCIE-level certifications are the only Cisco certs that mean anything because the practical is the only thing that comes close to testing a real-world crisis.
Actually, in Japan all 3 of the major carriers use SIM cards in their phones now. All the phones you buy through the carrier are locked though. You can buy WCDMA/GSM handsets direct from Nokia Japan that will work on both DoCoMo and Softbank however. (au is the odd man out here since they use CDMA2000 on their network instead of WCDMA)
Coincidentally, I have a W41CA myself. It was the smallest/lightest phone in the au lineup when I got it (the replacement model may also still be the lightest one available). All the reviews seem to miss the biggest feature - the penguin animations when you have the phone open.
Felica terminals that are not directly tied to the register. You ring up the correct value, it pops up on the register display. The cashier enters a higher value debit into the felica terminal. You plop your card down and are distracted by the cashier and miss the deduction which only shows up on the tiny terminal display for about a second.
I caught a cashier in Japan doing this to me.
The king of handling files for video editing was XFS/XLV on SGI iron. You could do guaranteed I/O throughput reservation from a filesystem/lv and the system would steal every bit of cpu/io possible to assure you meet the reservation requirements.
It was great technology, but was killed in the marketplace by cheap fast commodity hardware being "fast enough" that it wasn't needed any more.
At present ZFS will not let you add or subtract disks to a single raidz component. This is one of the big shortcomings to ZFS flexability at present.
What you can do to add capacity is to add another component, either raidz or mirror to your pool.
You can also replace individual drives in the raidz with larger ones, and after all drives are replaced, exporting and importing your zpool will make the additional space available.
Even at 2GB there are lots of issues with existing embedded devices, because that's where the change from 512-byte to 1k blocks takes place. As an example, most current mobile phones won't handle anything over 1GB because their hardware assumes 512-byte blocks and can't deal with anything different.
Actually caching for ISPs or any carrier where you don't know what your customers are doing is problematic. In the real world, people push all kinds of non-HTTP traffic over port 80 because it is the path of least resistance in getting by filters, and in many cases they do not encapsulate it inside an HTTP wrapper.
Startforce http://www.startforce.jp/ lets you do the same thing. I've even run YouOS inside Startforce.
There's a case to be made for online application suites like this in an intranet context to enforce centralized document storage for compliance reasons (oh god, a SOX clone is coming to Japan - maybe its time to move to China).
Mass transit stops from about half past midnight to 5am or so. Anyone that needs to get anywhere during this time needs to either:
1. Wait
2. Pay a very expensive (US$100-150 for me to get home if I work after midnight) taxi fare
3. Have a car
I don't have any stats either, however after living here for 9 years I know it to be typical behaviour.
You can also google for the Steven Herman vs. Asahi Bank case, or numerous cases of credit denials explicitly for being non-japanese.
OK, coctail-napkin math time:
Assume:
Bandwidth costs $50/mbit/sec (a reasonable ballpark figure if you're buying transit in bulk)
Client size is 2GB (Never played L2, but another guesstimate)
Peak 95th percentile bandwidth usage of 5 times average
The 500,000 users you just quoted
16,000,000,000 bits/(30*24*3600)=6172 bits/sec*$50/mbit/sec*500,000*5=$771,604
Not too far off those figures
Conversely, a shift to IPv6 will also reduce the effectiveness of worms because the sparseness of address distribution makes portscanning orders of magnitude less effective. It's all a wash really.
Playing devil's advocate: this has everything you just asked for that is not a Cisco proprietary protocol.
We already use a bonded disposal company to send our media to its final repose. They come to our office with a metal case, which is locked after filling with DLT tapes, then tracked all the way to the disposal site where it is run through a crusher that turns it into bits no larger than 0.75cm on a side.
They've been really hard-pressed to keep up their service availability since the japanese personal information protection law came into effect a year and a half or so back.
WWV and other similar radio clocks (I use JJY in japan) are only 1 second resolution, which makes those sort of clocks not really suitable for seriously mission critical timing needs, even though they technically qualify as stratum 1.
At home I use a little JJY-clock built from a kit that cost me about US$40 and connects via RS-232 and is supported in the reference ntp implementation. Has a PIC, some RS-232 glue and not much more other than the antenna and VCO.