While iCloud is as expected, I am still curious to see it in action. Will the terrible file management experience of iOS devices disappear with the iCloud?
Actually, MobileMe mail service is running on the Oracle Communications Suite formerly known as iPlanet/Java Messaging Server: Mac Web Server:Apache/1.3.33 (Darwin) Solaris or Linux IMAP Server: OK [CAPABILITY IMAP4 IMAP4rev1 ACL QUOTA LITERAL+ NAMESPACE UIDPLUS CHILDREN BINARY UNSELECT SORT CATENATE URLAUTH LANGUAGE ESEARCH ESORT THREAD=ORDEREDSUBJECT THREAD=REFERENCES CONDSTORE ENABLE CONTEXT=SEARCH CONTEXT=SORT WITHIN SASL-IR SEARCHRES XSENDER X-NETSCAPE XSERVERINFO X-SUN-SORT ANNOTATE-EXPERIMENT-1 X-UNAUTHENTICATE X-SUN-IMAP X-ANNOTATEMORE XUM1 IDLE AUTH=PLAIN AUTH=PLAIN] Messaging Multiplexor (Oracle Communications Messaging Exchange Server 7.4-20.01 (built Nov 21 2010)) Solaris SMTP Server: asmtpout027.mac.com -- Server ESMTP (Oracle Communications Messaging Exchange Server 7u4-18.01 64bit (built Jul 15 2010))
I am guessing that we are talking about SPARC Solaris, at least for the IMAP server. On Solaris SPARC, UltraSPARC is a requirement. I am going on a limb and guessing that they are not using Solaris on intel. While they might even run RHEL, I would be very surprised.
That's because parent doesn't have his own Solaris based NAS exporting his high performance disks over iSCSI or NFS. If you have your own real NAS (not the 500MHz ARM kind), you don't need eSATA, FW, USB or Thunderbolt for storage.
Technically speaking Firewire also tops at 3200Mbps. Unfortunately most products don't actually implement that. Exceptionally, the Mac Pro supports 3200Mbps and has done so for quite some time. However, the rest of the Apple products offer only 800Mbps. That is OK because I haven't yet seen any consumer products that actually use a 3200Mbps link over FW. With Thunderbolt it's not about external disks. Except for a few users (think movie editing), most people will end up having a reasonably fast NAS at home. It is however about Thunderbolt based port replicators/docking stations, since it extends the PCI bus, thus being able to add USB controllers, NICs, FW cards and other devices physically to the computer by a single cable. They missed a great opportunity by not including also power over Thunderbolt. It could of been the single cable required to charge, dock and extend the screen of a Mac. It's not the speed of Thunderbolt that matters, it's the PCI-E part that matters. Being able to extend the PCI-E bus has a lot of applications. Imagine an ultra-high density mac mini tray that extends the mac minis to add a second NIC (for redundancy), display and a LOM. That would make the Mac Mini the best server out there for hosting websites. In the width of a rack you can put 12 mac minis. In 800mm of depth plus 200 for cable routing you can put 4 rows of Mac Minis. That means 48 mac minis in 5u. That equals 384 mac minis in 40u. Simple math tells you that you can get at under $700/core a total of 768 cores with 4GB of RAM/core (including UPS and switches) in under $500k and under 33W/core in 2 racks (one with UPSs and one with the actual Mac Minis. No blade solution out there can beat that and even the cheapest ones are still at over $1000/core.
The filesystem part, almost matches the ZFS features, but not quite. The rest: NO! ZFS does a lot more than just being a filesystem, the most important non-filesystem part being the storage pool management in a simple and sane way. Some of that might be achievable with LVM, but most of it NO!
Returning to the filesystem part: When do you think that we'll get NFSv4 ACLs in Linux? Regardless of the filesystem. How about iSCSI integration (ZFS style)? How about transparent encryption? How about transparent compression? How about inheritable ACLs (posix or NFSv4)? How about data deduplication? How about something similar to L2ARC? It gives a ZFS with two SSDs and 10 cheap 1TB Samsung 4200RPM drives (a total of about $2 with all the hardware), the same capacity and performance as 20+ 15k RPM fibre-channel drives ($20k just the drives). How about incremental dumps? Btrfs is a COW filesystem that allows snapshots after all. Solaris had that, even in UFS for a long time.
You know how you should think about ZFS? As a $30.000 NetApp FAS2020 filler for free in any $1.000 computer.
I like and use Linux, but comparing ZFS with anything on the Linux side is like comparing an excavator with a shovel. Right now, anything on the Linux side is at best at NTFS feature level with arguably better performance, though investing in faster hardware might alleviate that.
Without DTrace and ZFS Solaris wouldn't really be irreplaceable today, but those two components make it the perfect OS for a little while longer.
ZFS doesn't really suffer from licensing problems. The OpenSolaris derived implementations (Solaris/FreeBSD/OpenIndiana), do however suffer from that problem. I would like to point out that ZFS has been available under GPL for a long time as GRUB patches to Solaris/OpenSolaris/Solaris_Express, and just recently integrated into the mainstream GRUB. The only milestone to adopting it in Linux is the insane amount of work. The GRUB implementation is GPL and it clearly exposes the internals of ZFS enough to implement it, it is not suitable for a direct import. It is most probably suitable for documenting the filesystem layout for a brand new implementation. The reason is quite simple: the boot loader implementation has only a single goal: read any requested file on a filesystem (ignoring access rights, write support, checksumming, snapshots, properties, or other fancy stuff).
Blu-ray is a method of distribution, H.264 is a video codec. Most blu-ray releases actually use H.264 as a codec. Furthermore, nobody is stopping you from buying a USB blu-ray disk. Regarding the integrated drive, while on the laptops it's not feasible because there are no 9.5mm slot loading SATA blu-ray drives, on the iMac and (possibly) on the mac mini you can upgrade to blu-ray because you can fit a 12.7mm drive in them.
If Windows wasn't a good OS, why are everyone using it?
Who's everyone? While Windows market share is still huge, I'm not considering it dominant anymore. Don't look at MacOS when you look at the ratios, look also at iPad, iPhone and whatever else. The tablets will eventually become at least as common as the PCs and Apple kicks arse for now, smart-phones are taking care of more of our needs than ever before and the iPhone is doing well and Android is doing excellently. Slowly we are getting to a point where the PC is not required anymore. For average users a NAS, a PS3/Apple_TV/XBox360 and a tablet are everything that they need at home. PCs are slowly fading from their dominant position and we don't have Windows on any of the devices that I mentioned before.
Actually Facetime is an open spec, anyone can implement a device that supports it...
Now how you find them from an iOS to non-IOS device, that part I'm not sure how easy it is to implement.
Care to give me a link to the spec? I know it uses industry standard protocols, but the spec?
I like FaceTime for being the first simple, grandma' proof, button free SIP client, but it's not available on other platforms that Apple's own. An open spec SIP based platform would be usable with other video enabled SIP clients. Facetime AFAIK is not.
That's a matter of debate. OS X is as good as Ubuntu, Fedora, as a desktop but where it really shines is the apps. OSS apps are great, but unusable unless the complete workflow of the department allows for them. In ours, we HAVE to use MS Office, Sharepoint, Exchange, regardless of our personal preference.
Regardless of that, you have VMWare, Parallels and VirtualBox in Mac OS X and I am running RHEL, Solaris, NetApp Simulator, Windows Server and others in it.
Whether OS X is better than Linux at being a UNIX workstation is of little relevance. Having MS Office with Outlook, as required by policy is relevant though.
The 2010 MacBook Pro is a very powerful beast, under OS X. With the SSD my Sandbox infrastructure boots up in 90 seconds while the physical servers take about 10 minutes to boot.
These Ultra Secure Environment things are usually governed by a clear set of policies created for each such project uniquely. The keys are usually not allowed by policy to be in the same geographical area unless so requested by the policy creating authority. For such an op (strictly speaking it's a ceremony), you should schedule the arrival in advance and securely send all the documentation detailing the process to all participants with time to spare. If the information is valuable enough, the participants should be housed at different hotels and should travel on different dates with different airlines through different hubs. In some scenarios, no more than 3 people are allowed to be in the same country at the same time for any reason, personal or else, unless so requested by the policy creating authority or an approved operations committee.
What actually surprises me at this article is the fact that the list of countries as well as the identity of at least one of the persons is public. In some environments, even this is a security breach.
After kidnapping this guy, you are 3 persons away from being able to recreate the root DNS signing key. Imagine that in some small countries such as CZ and BE it's quite easy to identify the persons with a bit of detective work (IT guys 200000, IT members of the local TLD 100, IT workers trustworthy of such a secret: 10/country). Considering that the scope of the global DNSSEC private key is worth a hefty budget to any hostile party, this disclosure should be treated as a security risk or a policy failure.
Getting the necessary 4 keys would take at most USD 1M for the detective work, easily achievable for a hostile entity and with an identical operation at a trusted CA or a signed enterprise CA you can easily create valid EV certs for paypal, amazon and ebay and transparently replace DNS for a given geography.
Nope. It's common practice in the PKI world to use an HSM which calculates the private key upon startup. The key is not stored anywhere. It's calculated when you start the HSM. It's a function with 7 intersection points with the X axis. Knowing any 4 of the 7 intersection points is enough to calculate the function parameter. That in turn is the actual private key.
RAID has nothing to do with this. The HSMs operate under the presumption that the safest guard for the private key is not to have it at all, encrypted or not. You calculate it only when needed. If the HSM goes down you need a new key migration ceremony in a worst case scenario, and in the best case scenario, just the administrator and operator smart cards to unlock the security world.
This is what is being done at any public CA installed in your browser and at any Publicly signed Enterprise CA.
My iPhone generates 3GB of data monthly. That's just my iPhone! I don't watch online TV or listen to music on it. Just occasional light browsing and email. And a very rare 64kbps AAC radio.
My prepaid German SIM card gives me 5GB of data for €25 on O2. My Romanian Vodafone card gives me 5GB for €11 added to my €18 voice plan (1000 minutes outgoing, unlimited incoming).
It's only in the States that people are overcharged like that.
My Vodafone Account Manager called me 2 days ago to ask me if I have any devices that need a micro-SIM card. The 3G coverage is almost nation-wide and they do provide 21Mbps HSPA.
For the iPad (and other devices), they also have some very nice options that can be negotiated down quite a lot:
7,99 € 7.2 Mbps 1 GB 10,99 € 7.2 Mbps 2 GB 12,99 € 21.6 Mbps 4 GB 16,99 € 21.6 Mbps 8 GB 26,99 € 21.6 Mbps 16 GB
http://www.linux-kvm.org/page/Main_Page - It says clearly that it uses "x86 hardware containing virtualization extensions (Intel VT or AMD V)". See also: http://www.linux-kvm.org/page/Status On http://en.wikipedia.org/wiki/X86_virtualization it says that the AMD V instructions came up in the second generation Opteron processors (the ones with Socket F). Furthermore, from the same Wikipedia: "All Socket 939 and only Sempron processors except Sable, Huron and Sargas do not include support for AMD-V".
If you don't believe me, do a 'cat/proc/cpuinfo' and tell me if you see any virtualization instructions over there.
Don't really care about Xen vs. KVM from a product perspective, but for the Opteron 270, Xen is the only one that works since that Opteron doesn't have hardware virtualization instructions. KVM doesn't (to my knowledge) support software based paravirtualization like Xen.
We have an environment with AMD Opteron 270 based servers where we use virtualization heavily. We either have to give up on the servers or on RHEL 6. I think that we'll stick with EL5 until we go into a server refresh cycle.
1) Zarafa - the OSS version is lacking the Outlook and ActiveSync parts => not exactly what we need. I'm not really impressed with the administration and integration with existing services.
2) Zimbra - just like above, we don't have what we need and it's really a pain in the way it's packaged. It comes with it's own MySQL, Tomcat, OpenLDAP, etc. Thus it doesn't really integrate with what you already have. If you already have you own Port389/iPlanet or OpenLDAP you're probably not looking forward for another one.
3) Scalable OpenGroupware.org. - This has potential. It comes with RHEL packages for everything, it uses your own Postgres, your own IMAP, your own SMTP and your own LDAP => it rocks. A SOGo + Mozilla Messaging solution is by far the best thing that you could get to replace Exchange. There's even a proprietary plugin for Outlook (ZideOne) if you need it. It even has integration with Funambol for SyncML and ActiveSync features. The webmail is great, but it's not yet production ready, despite the 1.0 version number.
And you find that having all your emails go through a 3rd party's servers is a good thing? Otherwise I'm pretty sure that a honest comparison between the iPhone and the Blackberry Exchange integration wouldn't show any differences feature-wise. Regarding the Push-Email thing, it's actually there in the iPhone OS 3.0 and it works.
On the other hand, trying to remove the blackberry from the hands of blackberry obsessed users is a sacrilege just as big as upgrading their Office 2003 to anything else. And that is regardless of the fact that the blackberry is actually crappy for a very important reason: having a J2ME app that runs on both any J2ME enabled phone as well as on a blackberry is a pain. Just think of the button mappings. Granted that on this one in particular, the iPhone sucks even more because it doesn't even have a J2ME implementation.
Or, if you enabled it, just tap the track pad with two fingers. I can't understand why people still use the button when they can actually tap the track pad with one or two fingers for left, respectively right clicks.
Actually the fault is split. 2D acceleration in Linux for most video drivers is shabby at best.
On the other hand, Adobe doesn't really put that much engineering force into X11 optimizations. Adobe Flash on a non-accelerated Mac OS X (hackintosh using the included Vesa 3.0 driver) is still faster than on X11/Linux.
I can't really blame Adobe for this. There are quite a lot of ways in which you can accelerate SOME drawing operations, but they are not available on all desktops. Clutter comes to mind right now, but it's not really the best option for QT/KDE users. It's hard to create an accelerated, desktop environment independent piece of software.
The irony is that hybrid diesels would be perfect, but nobody takes the concept to it's true potential. Diesel electric all the way, like in train engines. A Diesel likes to have a constant RPM at it's peak performance value. Imagine connecting an alternator directly to the engine and giving up the inefficient gear system. Imagine a Diesel engine that is always at it's peak performance RPM, even when there's barely any electrical load on it. That car would be a rocket that goes for free (or almost free). It's also pretty easy to build if you have 2 things: 1) 1x 150kW alternator (it's the right amount) that also fits under the hood along with the engine. 2) 4x 40kW electric engines that you connect directly to the drive shaft (and should also fit in there somewhere). As far as I know a 150kW alternator is very big (about as big as the engine itself) and the 40kW engines are also huge, but at least in theory this would be by far the best way to bring the top possible performance of an engine to the tarmac. Electricity is the best way to transfer energy between two points and a constant RPM diesel is the most efficient and performant diesel out there.
The only difference between the US and Romania (for example) is the cost of the circuit. In Romania you get fibre even for 2 Mbps, and you can switch to more bandwidth almost instantly, with a simple telephone. A 1Gbps circuit a your address is about $400, and the bandwidth is in anywhere between $40 (for 1-2Mbps) and $6 (for more than 1Gbps) depending on the amount that you need. An added extra is that they typically give you 5 to 10 times more bandwidth for in-country access, limited of course by the circuit that you have. Having 100Mbps nation-wide is great for torrents, using OSS ftp mirrors and even remote backups.
Feature and usability-wise is it getting close to the iPhone? I have a lot of "toys" at home, including a GTA01 and a Nokia N800. While a lot better in some technical aspects, and in most philosophical ones, they all fade in comparison to the iPhone. No SyncML, no PIM suite (GPE doesn't count as it's not really integrated to the platform).
Slashdot Mirror
While iCloud is as expected, I am still curious to see it in action. Will the terrible file management experience of iOS devices disappear with the iCloud?
Actually, MobileMe mail service is running on the Oracle Communications Suite formerly known as iPlanet/Java Messaging Server:
Mac Web Server:Apache/1.3.33 (Darwin)
Solaris or Linux IMAP Server: OK [CAPABILITY IMAP4 IMAP4rev1 ACL QUOTA LITERAL+ NAMESPACE UIDPLUS CHILDREN BINARY UNSELECT SORT CATENATE URLAUTH LANGUAGE ESEARCH ESORT THREAD=ORDEREDSUBJECT THREAD=REFERENCES CONDSTORE ENABLE CONTEXT=SEARCH CONTEXT=SORT WITHIN SASL-IR SEARCHRES XSENDER X-NETSCAPE XSERVERINFO X-SUN-SORT ANNOTATE-EXPERIMENT-1 X-UNAUTHENTICATE X-SUN-IMAP X-ANNOTATEMORE XUM1 IDLE AUTH=PLAIN AUTH=PLAIN] Messaging Multiplexor (Oracle Communications Messaging Exchange Server 7.4-20.01 (built Nov 21 2010))
Solaris SMTP Server: asmtpout027.mac.com -- Server ESMTP (Oracle Communications Messaging Exchange Server 7u4-18.01 64bit (built Jul 15 2010))
I am guessing that we are talking about SPARC Solaris, at least for the IMAP server. On Solaris SPARC, UltraSPARC is a requirement. I am going on a limb and guessing that they are not using Solaris on intel. While they might even run RHEL, I would be very surprised.
Most probably a Thunderbolt FW adapter like the Sonnet Allegro FireWire 800 adapter. It's a Thunderbolt to PCI-E bridge and a PCI-E Firewire adapter.
That's because parent doesn't have his own Solaris based NAS exporting his high performance disks over iSCSI or NFS. If you have your own real NAS (not the 500MHz ARM kind), you don't need eSATA, FW, USB or Thunderbolt for storage.
Technically speaking Firewire also tops at 3200Mbps. Unfortunately most products don't actually implement that. Exceptionally, the Mac Pro supports 3200Mbps and has done so for quite some time. However, the rest of the Apple products offer only 800Mbps. That is OK because I haven't yet seen any consumer products that actually use a 3200Mbps link over FW.
With Thunderbolt it's not about external disks. Except for a few users (think movie editing), most people will end up having a reasonably fast NAS at home. It is however about Thunderbolt based port replicators/docking stations, since it extends the PCI bus, thus being able to add USB controllers, NICs, FW cards and other devices physically to the computer by a single cable. They missed a great opportunity by not including also power over Thunderbolt. It could of been the single cable required to charge, dock and extend the screen of a Mac.
It's not the speed of Thunderbolt that matters, it's the PCI-E part that matters. Being able to extend the PCI-E bus has a lot of applications. Imagine an ultra-high density mac mini tray that extends the mac minis to add a second NIC (for redundancy), display and a LOM. That would make the Mac Mini the best server out there for hosting websites. In the width of a rack you can put 12 mac minis. In 800mm of depth plus 200 for cable routing you can put 4 rows of Mac Minis. That means 48 mac minis in 5u. That equals 384 mac minis in 40u. Simple math tells you that you can get at under $700/core a total of 768 cores with 4GB of RAM/core (including UPS and switches) in under $500k and under 33W/core in 2 racks (one with UPSs and one with the actual Mac Minis. No blade solution out there can beat that and even the cheapest ones are still at over $1000/core.
The filesystem part, almost matches the ZFS features, but not quite.
The rest: NO!
ZFS does a lot more than just being a filesystem, the most important non-filesystem part being the storage pool management in a simple and sane way.
Some of that might be achievable with LVM, but most of it NO!
Returning to the filesystem part:
When do you think that we'll get NFSv4 ACLs in Linux? Regardless of the filesystem.
How about iSCSI integration (ZFS style)?
How about transparent encryption?
How about transparent compression?
How about inheritable ACLs (posix or NFSv4)?
How about data deduplication?
How about something similar to L2ARC? It gives a ZFS with two SSDs and 10 cheap 1TB Samsung 4200RPM drives (a total of about $2 with all the hardware), the same capacity and performance as 20+ 15k RPM fibre-channel drives ($20k just the drives).
How about incremental dumps? Btrfs is a COW filesystem that allows snapshots after all. Solaris had that, even in UFS for a long time.
You know how you should think about ZFS? As a $30.000 NetApp FAS2020 filler for free in any $1.000 computer.
I like and use Linux, but comparing ZFS with anything on the Linux side is like comparing an excavator with a shovel. Right now, anything on the Linux side is at best at NTFS feature level with arguably better performance, though investing in faster hardware might alleviate that.
Without DTrace and ZFS Solaris wouldn't really be irreplaceable today, but those two components make it the perfect OS for a little while longer.
ZFS doesn't really suffer from licensing problems. The OpenSolaris derived implementations (Solaris/FreeBSD/OpenIndiana), do however suffer from that problem.
I would like to point out that ZFS has been available under GPL for a long time as GRUB patches to Solaris/OpenSolaris/Solaris_Express, and just recently integrated into the mainstream GRUB.
The only milestone to adopting it in Linux is the insane amount of work. The GRUB implementation is GPL and it clearly exposes the internals of ZFS enough to implement it, it is not suitable for a direct import. It is most probably suitable for documenting the filesystem layout for a brand new implementation. The reason is quite simple: the boot loader implementation has only a single goal: read any requested file on a filesystem (ignoring access rights, write support, checksumming, snapshots, properties, or other fancy stuff).
Blu-ray is a method of distribution, H.264 is a video codec. Most blu-ray releases actually use H.264 as a codec. Furthermore, nobody is stopping you from buying a USB blu-ray disk. Regarding the integrated drive, while on the laptops it's not feasible because there are no 9.5mm slot loading SATA blu-ray drives, on the iMac and (possibly) on the mac mini you can upgrade to blu-ray because you can fit a 12.7mm drive in them.
If Windows wasn't a good OS, why are everyone using it?
Who's everyone? While Windows market share is still huge, I'm not considering it dominant anymore. Don't look at MacOS when you look at the ratios, look also at iPad, iPhone and whatever else. The tablets will eventually become at least as common as the PCs and Apple kicks arse for now, smart-phones are taking care of more of our needs than ever before and the iPhone is doing well and Android is doing excellently. Slowly we are getting to a point where the PC is not required anymore. For average users a NAS, a PS3/Apple_TV/XBox360 and a tablet are everything that they need at home. PCs are slowly fading from their dominant position and we don't have Windows on any of the devices that I mentioned before.
Actually Facetime is an open spec, anyone can implement a device that supports it...
Now how you find them from an iOS to non-IOS device, that part I'm not sure how easy it is to implement.
Care to give me a link to the spec? I know it uses industry standard protocols, but the spec?
I like FaceTime for being the first simple, grandma' proof, button free SIP client, but it's not available on other platforms that Apple's own. An open spec SIP based platform would be usable with other video enabled SIP clients. Facetime AFAIK is not.
That's a matter of debate. OS X is as good as Ubuntu, Fedora, as a desktop but where it really shines is the apps. OSS apps are great, but unusable unless the complete workflow of the department allows for them. In ours, we HAVE to use MS Office, Sharepoint, Exchange, regardless of our personal preference.
Regardless of that, you have VMWare, Parallels and VirtualBox in Mac OS X and I am running RHEL, Solaris, NetApp Simulator, Windows Server and others in it.
Whether OS X is better than Linux at being a UNIX workstation is of little relevance. Having MS Office with Outlook, as required by policy is relevant though.
The 2010 MacBook Pro is a very powerful beast, under OS X. With the SSD my Sandbox infrastructure boots up in 90 seconds while the physical servers take about 10 minutes to boot.
These Ultra Secure Environment things are usually governed by a clear set of policies created for each such project uniquely. The keys are usually not allowed by policy to be in the same geographical area unless so requested by the policy creating authority.
For such an op (strictly speaking it's a ceremony), you should schedule the arrival in advance and securely send all the documentation detailing the process to all participants with time to spare. If the information is valuable enough, the participants should be housed at different hotels and should travel on different dates with different airlines through different hubs.
In some scenarios, no more than 3 people are allowed to be in the same country at the same time for any reason, personal or else, unless so requested by the policy creating authority or an approved operations committee.
What actually surprises me at this article is the fact that the list of countries as well as the identity of at least one of the persons is public. In some environments, even this is a security breach.
After kidnapping this guy, you are 3 persons away from being able to recreate the root DNS signing key. Imagine that in some small countries such as CZ and BE it's quite easy to identify the persons with a bit of detective work (IT guys 200000, IT members of the local TLD 100, IT workers trustworthy of such a secret: 10/country). Considering that the scope of the global DNSSEC private key is worth a hefty budget to any hostile party, this disclosure should be treated as a security risk or a policy failure.
Getting the necessary 4 keys would take at most USD 1M for the detective work, easily achievable for a hostile entity and with an identical operation at a trusted CA or a signed enterprise CA you can easily create valid EV certs for paypal, amazon and ebay and transparently replace DNS for a given geography.
Nope. It's common practice in the PKI world to use an HSM which calculates the private key upon startup. The key is not stored anywhere. It's calculated when you start the HSM. It's a function with 7 intersection points with the X axis. Knowing any 4 of the 7 intersection points is enough to calculate the function parameter. That in turn is the actual private key.
RAID has nothing to do with this. The HSMs operate under the presumption that the safest guard for the private key is not to have it at all, encrypted or not. You calculate it only when needed. If the HSM goes down you need a new key migration ceremony in a worst case scenario, and in the best case scenario, just the administrator and operator smart cards to unlock the security world.
This is what is being done at any public CA installed in your browser and at any Publicly signed Enterprise CA.
My iPhone generates 3GB of data monthly. That's just my iPhone! I don't watch online TV or listen to music on it. Just occasional light browsing and email. And a very rare 64kbps AAC radio.
My prepaid German SIM card gives me 5GB of data for €25 on O2. My Romanian Vodafone card gives me 5GB for €11 added to my €18 voice plan (1000 minutes outgoing, unlimited incoming).
It's only in the States that people are overcharged like that.
My Vodafone Account Manager called me 2 days ago to ask me if I have any devices that need a micro-SIM card. The 3G coverage is almost nation-wide and they do provide 21Mbps HSPA.
For the iPad (and other devices), they also have some very nice options that can be negotiated down quite a lot:
7,99 € 7.2 Mbps 1 GB
10,99 € 7.2 Mbps 2 GB
12,99 € 21.6 Mbps 4 GB
16,99 € 21.6 Mbps 8 GB
26,99 € 21.6 Mbps 16 GB
http://www.linux-kvm.org/page/Main_Page - It says clearly that it uses "x86 hardware containing virtualization extensions (Intel VT or AMD V)".
See also: http://www.linux-kvm.org/page/Status
On http://en.wikipedia.org/wiki/X86_virtualization it says that the AMD V instructions came up in the second generation Opteron processors (the ones with Socket F).
Furthermore, from the same Wikipedia: "All Socket 939 and only Sempron processors except Sable, Huron and Sargas do not include support for AMD-V".
If you don't believe me, do a 'cat /proc/cpuinfo' and tell me if you see any virtualization instructions over there.
Don't really care about Xen vs. KVM from a product perspective, but for the Opteron 270, Xen is the only one that works since that Opteron doesn't have hardware virtualization instructions. KVM doesn't (to my knowledge) support software based paravirtualization like Xen.
We have an environment with AMD Opteron 270 based servers where we use virtualization heavily. We either have to give up on the servers or on RHEL 6. I think that we'll stick with EL5 until we go into a server refresh cycle.
Let's make a small analisys of the landscape:
1) Zarafa - the OSS version is lacking the Outlook and ActiveSync parts => not exactly what we need. I'm not really impressed with the administration and integration with existing services.
2) Zimbra - just like above, we don't have what we need and it's really a pain in the way it's packaged. It comes with it's own MySQL, Tomcat, OpenLDAP, etc. Thus it doesn't really integrate with what you already have. If you already have you own Port389/iPlanet or OpenLDAP you're probably not looking forward for another one.
3) Scalable OpenGroupware.org. - This has potential. It comes with RHEL packages for everything, it uses your own Postgres, your own IMAP, your own SMTP and your own LDAP => it rocks. A SOGo + Mozilla Messaging solution is by far the best thing that you could get to replace Exchange. There's even a proprietary plugin for Outlook (ZideOne) if you need it. It even has integration with Funambol for SyncML and ActiveSync features. The webmail is great, but it's not yet production ready, despite the 1.0 version number.
And you find that having all your emails go through a 3rd party's servers is a good thing? Otherwise I'm pretty sure that a honest comparison between the iPhone and the Blackberry Exchange integration wouldn't show any differences feature-wise. Regarding the Push-Email thing, it's actually there in the iPhone OS 3.0 and it works.
On the other hand, trying to remove the blackberry from the hands of blackberry obsessed users is a sacrilege just as big as upgrading their Office 2003 to anything else. And that is regardless of the fact that the blackberry is actually crappy for a very important reason: having a J2ME app that runs on both any J2ME enabled phone as well as on a blackberry is a pain. Just think of the button mappings. Granted that on this one in particular, the iPhone sucks even more because it doesn't even have a J2ME implementation.
Or, if you enabled it, just tap the track pad with two fingers. I can't understand why people still use the button when they can actually tap the track pad with one or two fingers for left, respectively right clicks.
Actually the fault is split. 2D acceleration in Linux for most video drivers is shabby at best.
On the other hand, Adobe doesn't really put that much engineering force into X11 optimizations. Adobe Flash on a non-accelerated Mac OS X (hackintosh using the included Vesa 3.0 driver) is still faster than on X11/Linux.
I can't really blame Adobe for this. There are quite a lot of ways in which you can accelerate SOME drawing operations, but they are not available on all desktops. Clutter comes to mind right now, but it's not really the best option for QT/KDE users. It's hard to create an accelerated, desktop environment independent piece of software.
The irony is that hybrid diesels would be perfect, but nobody takes the concept to it's true potential.
Diesel electric all the way, like in train engines. A Diesel likes to have a constant RPM at it's peak performance value. Imagine connecting an alternator directly to the engine and giving up the inefficient gear system. Imagine a Diesel engine that is always at it's peak performance RPM, even when there's barely any electrical load on it. That car would be a rocket that goes for free (or almost free). It's also pretty easy to build if you have 2 things:
1) 1x 150kW alternator (it's the right amount) that also fits under the hood along with the engine.
2) 4x 40kW electric engines that you connect directly to the drive shaft (and should also fit in there somewhere).
As far as I know a 150kW alternator is very big (about as big as the engine itself) and the 40kW engines are also huge, but at least in theory this would be by far the best way to bring the top possible performance of an engine to the tarmac. Electricity is the best way to transfer energy between two points and a constant RPM diesel is the most efficient and performant diesel out there.
The only difference between the US and Romania (for example) is the cost of the circuit. In Romania you get fibre even for 2 Mbps, and you can switch to more bandwidth almost instantly, with a simple telephone. A 1Gbps circuit a your address is about $400, and the bandwidth is in anywhere between $40 (for 1-2Mbps) and $6 (for more than 1Gbps) depending on the amount that you need. An added extra is that they typically give you 5 to 10 times more bandwidth for in-country access, limited of course by the circuit that you have. Having 100Mbps nation-wide is great for torrents, using OSS ftp mirrors and even remote backups.
Feature and usability-wise is it getting close to the iPhone?
I have a lot of "toys" at home, including a GTA01 and a Nokia N800. While a lot better in some technical aspects, and in most philosophical ones, they all fade in comparison to the iPhone. No SyncML, no PIM suite (GPE doesn't count as it's not really integrated to the platform).
Don't bears hibernate?