infrageeks

The cool

I've been trawling through all of the online marketing material concerning the new Sun™ Storage 7000 Series Family and overall I'm liking what I see from a general architecture standpoint. They appear to have found a simple, effective means towards mixing different tiers of storage (price/performance/capacity) that spans SSHD, SAS 10K and SATA 7K. You put create pools of SSHD devices for Logging (Write) and Pre-fetch (Read) operations in front of your primary physical disk storage. This is simplistic when compared with a true multi-tier solution like Compellent, but has the advantage of being a simpler, pragmatic solution to managing the gap between IOps and available surface.

Functionally you are adding a very large tier of cache between the very volatile DRAM cache and the much slower disks. A practical two tier storage system based on the ZFS components. In fact, there's nothing really new here at all (see: Logging and Cache devices), except that with the advent of 'affordable' high performance SSHD devices, we can now imagine pulling more than acceptable performance out of a 100% SATA based array.

The upshot is that by spreading your CapEx investment out between low-cost, high volume SATA drives and a few expensive high-speed SSHD, you can probably come out a lot cheaper than the required number of 15K FC disks to supply the same number of IOps. This has a direct impact on the hard OpEx figures like power cost since you're running SSHD (low power devices) and fewer, slower disks that also consume less (especially with the latest generation of 'green' drives). This contrasted with the high number of power hungry 15K disks that needed to be able to supply an equivalent volume of IOps.

The software

Where the solution really shines is in the management software. Go download the VMware simulator and take it for a spin. A really well done storage management interface that handles everything in a clear and understandable manner. (Hint to Sun: not everyone uses a QWERTY keyboard).

Live, real time instrumentation of everything that's happening in the system: want to see a breakdown of NFS traffic by latency? No problem. Want to correlate CPU use with CIFS traffic? Done. Is my write traffic saturating my available cache devices? Right here.

This is one of the biggest value adds here since virtualization and consolidation projects often run into issues with the shared storage and being able to easily look inside the storage environment and identify bottlenecks and performance issues is an absolute godsend.

One of the major attractions of the architecture is that it's based on standard AMD servers with a standard Solaris OS behind the scenes, so you can add whatever supported communications cards you want, up to 10GE for going head to head with some Fibre Channel solutions.

The not so great

The solutions are unfortunately (from what I can gather) restricted to very specific hardware combinations. The management toolkits are directly related to implementations based on each physical "appliance" and a specific set of expansion options. From what I can gather, you can't buy the entry level server, add your own HBA and hook up a JBOD to extend the available capacity1.

The related software cost of a solution goes up with the size of the storage bay purchased. A quick comparison of the equivalent J4540 server which is exactly the same thing, but in a DIY version where you can use ZFS to create the exact same performance profile, but you don't get the management console.

Now the following price equivalents are not exact since the appliance version replaced a few of the SATA disks with SSHD drives, but it gives you an idea of what the software will cost you over top of the basic hardware implementation:

	7210	X4540	Software cost
12Tb	$35,000	$22,000	$12,000
24Tb	$72,000	$42,000	$30,000
44Tb	$118,000	$62,000	$56,000

My take

This is an awesome solution that is a direct shot across the bow of Netapp and all of the iSCSI appliances out there and takes aim at a lot of the midrange storage solutions out there. I haven't done the pricing of equivalent Netapp solutions but in my experience, they're a whole lot more expensive and don't offer the kind of easily accessed analytics capabilites. They've got a few other arrows in their quiver (Metrocluster), but the bulk of their sales story just got a whole less compelling.

I'm disappointed that it's not more open with the ability to attach whatever storage you feel appropriate. The appliance version takes over the OS and drops you directly into an appliance shell, so you don't access to the raw command line to manage your own ZPools manually (at least that's the behavior in the simulator). I imagine that this could be hacked relatively easily, but would undoubtedly break the support agreement for the appliance.

What's also missing is the ability to easily expand your current model or move from model to model as your requirements grow. We're back in forklift upgrade mode if you started with the entry level and need to move to the next model up. The step between models is also pretty extreme - 2Tb to 12Tb makes for a big gap. From a VMware perspective this is less of pain than it used to be since the advent of SVMotion, but it's still a pain point.

I was hoping that with the entry level model you could add a SAS HBA and add in a couple of J4500 storage bays to permit a more practical growth curve. Also, the clustering option isn't available until you step up to the top of the line model which in my opinion is a mistake. The availability of storage is just as important for smaller volumes as it is for the bigger ones.

I'd really love to see Sun offer the FishWorks management software as a separate product that you can use to manipulate any available storage. This would permit you to use a standard Sun box as a SAN front end, optimized with local SSHD drives to give an instant performance boost.

But for anyone that's seriously considering a Netapp solution (without a MetroCluster) or Equalogic or EMC AX or CX these new products offer a really compelling value proposition.

1. I could be wrong here, but the simulator UI doesn't seem to offer any kind of free form pool management options. I tried adding some additional drives but haven't got them recognized yet.↑

This pretty much sums up the experience of a colleague of mine:

http://www.penny-arcade.com/2008/11/14/more-buttflo/

The hardware from HTC just keeps getting better and better, but it's hobbled by an awful user experience. It's all about the software these days, but there simply aren't enough really good developers that understand how to build a great experience. Those who are trulyfanaticalaboutthiskind of thingarealready developing for Apple, OS X and the iPhone because their fanaticism and perfectionism is recognized as a good thing in these communities.

Here's hoping that Android will step up to the plate and keep Apple on their toes but from what I've seen so far Apple's lead is still significant enough that they should be able to stay ahead of the pack.

It's been on back order forever, but my FastMac iV finally showed up last week at the office. Of course that's the week I was running all over the place so I had to wait until this week to pick it up. But now it's here and I'm quite happy with it.

I'd originally pre-ordered the model for the 1G iPhone, but since I upgraded, I had to change the order and wait for them to ramp up the manufacturing for the latest model.

What is it?

It's a heavy duty external battery that doubles as a case. The internal battery on the iPhone is about 1000mW so with the pack you have four times the autonomy of the iPhone by itself.

It's a slide in design, similar to the Mophie Juice Pack in that there's a pass through iPod connector so you don't have to pull it out of the case to charge it or sync. Very convenient. The bottom part looks a little odd because it's designed to fit into the docking station directly as well.

How much power?

Here's where the iV stands out - at 3100mW it's a substantially bigger battery than the Mophie or the new Richard Solo at 1800mW. Quoted capacity brings it up to 24 hours of talk time - 24 hours on the phone without stopping. Pretty crazy stuff. However, we all know that nobody uses their iPhone to call people - it's all about burning up the 3G radio waves and that sucks the juice down like nothing else.

In my informal first day test, I left the iPhone plugged in overnight docked in the iV and unplugged it with the alarm at 6:30 and went out for the day. I had an intense usage day (partly because I had the new battery), with the wifi and Push Mail activated, and stressing the wifi while testing out the newly acquired AirSharing while copying over about 2Gb of data in all sorts of formats to see how it performed. A few YouTube videos, some Fring VOIP communications, the usual compulsive NetNewsWire updates, a couple of chapters of Topper in Stanza, etc.

I also decided to really push my luck and disabled the autolock feature and kept the iPhone awake all day at 75% brightness.

The iV finally stopped feeding power to the iPhone around 6PM. And the iPhone internal battery is still fully charged and ready to keep going.

I'm going to try it out on a more normal cycle the rest of the week, but I think that I can safely say that this will satisfy the power needs of the hardiest road warriors.

Extras

There are a couple of additional nifty features that I quite appreciate. There's a built-in LED on the back that you can use as a flashlight - or better yet, for lighting up stuff so that you can take a picture when the conditions are less than perfect. It's not tied to the software in the iPhone, you need to toggle a manual switch on the side to turn it on and off.

There's also a standard USB port on the back. Not for charging the iV - you need to use the iPod connector, but for charging other accessories off of the battery, like your bluetooth headset

The catch

Well - nothing comes for free and more battery life means a bigger battery. This is not a slim, elegant case. If you bought your iPhone for the style, this is not the battery for you. If you like to keep your iPhone in your jeans pocket - that's probably not going to work either.

That said, it's not quite outside the realms of the reasonable, and for some odd reason it reminds me of my old Motorola MicroTAC in terms of weight and in hand feel (which is quite nice) with a soft rubberized surface that is much less slippery than the iPhone itself. It still fits in my suit jacket pockets, but does add a bit of sag on the lighter weight ones.

I suppose you could also glue one of those universal belt clips to the back, but then it wouldn't lie flat.

How much?

Currently offered at $99 for the 3G and $79 for the first generation iPhone, it represents the maximum autonomy available for those willing to live with the size.

Buy it again?

The other option is to carry around a few Novodio or Richard Solo type batteries and charge up as the battery gets low. But one thing I like about the really powerful case approach is that once the internal battery is charged up, the iPhone starts drawing power from the dock connection, thus minimizing the number of charge cycles imposed on the internal battery which should extend its useful life. Using my old Novodio approach I was going through multiple charges in a day, which is less than ideal

So would I buy it again? Yes. it fits my needs pretty well, and will be an absolute requirement if Apple finally offers some kind of tethering application...

First test: Google Mobile App for iPhone: "Google Mobile App released, but doesn't play well with British accents"

(Via Macworld UK.)

And for all those people that have been clamouring for voice commands on the iPhone remember this: Apple is a global company and anything they put out for the iPhone must work equally well in the US, India, France, Germany, Italy, ...

Given that the Newton handwriting recognition fiasco is probably still smarting, Apple isn't going to release a voice recognition or command tool as part of the basic package.

That said, it leaves open the market for third parties to propose localized applications that take into account regional characteristics - and if they don't work, Apple's not on the hook.

Witness Vocalia.

Re: Al Shipp, Apple's former Sr. V.P. of Enterprise Sales at Apple is leaving

What he said : http://www.macobserver.com/columns/hiddendimensions/2008/11/11.1.shtml

My take
This is one of the clearest explanations of Apple's approach to sales in general, not just the enterprise. Apple's clearly stated goal is to build the best possible products and that these products will sell based on their merits and applicability to a given customer, be they in the enterprise, a school or a home.

I think that Apple's current product line-up is exceedingly well adapted to many enterprises if they would stop and take the time to properly evaluate them. And it's only going to get better.

OS X Server has been evolving at least as rapidly as the desktop environment and offers a very good value many companies. There are a number of key features that will change the attractiveness of OS X Server for many Microsoft based companies, some are here today, and others are under development.

Active Directory Integration
Just like the desktop, OS X Server can be directly integrated into an Active Directory, but this integration goes even further, offering a Centrify-like ability to take any server application and base the authentication and authorization on Active Directory. This means that you can offer a standard compliant IMAP and SMTP server combination using your Active Directory accounts. No additional account management or synchronization required. Out of the box in the same amount of time that it takes you to join a workstation to an Active Directoryu domain. Oh - and without the reboot.

Mail and Calendaring
Today, the missing piece is the level of integration with the Outlook client on the Windows desktop, but there are few things that have changed recently: Apple has licenced the ActiveSync protocol for the iPhone and Exchange/Mail.app integration is promised in the Snow Leopard release. On the server side, the Snow Leopard Server mail and calendaring services described look an awful lot like the Exchange feature set.

Full Outlook client compatibility hasn't been announced by Apple, but it's a relatively small step to be able to offer the Exchange experience with these functions in the server. Either by using the ActiveSync protocol from the server side (I wonder if the license they have from Microsoft permits this) or by developing an Outlook plugin. Wouldn't that be the same thing that was developed for MobileMe and Outlook synchronization?

Collaboration
The first generation blog server included in Tiger was a bit of a hack based on the Blojsom project and worked pretty well as a blog server, but was complicated if you wanted to extend its capabilities. The Leopard release included a total rewrite of a general purpose collaboration platform that included both wiki and blog services. What's really important here is that the platform is entirely extensible using javascript and will be positioned in Snow Leopard as a viable alternative to Sharepoint. I've only just started looking at what you can do here and even in the current version, the options are astouding and I just wish I had more time to delve into the possibilities.

Going forward
Apple's goal is to build the best possible products and OS X Server is going to be even more attractive to enterprises going forward in difficult economic times. OS X Server unlimited version goes for $999 USD with no after sales gotchas, client access licences (CALs in Microsoft speak) or additional products like a database server. Again, Apple's goal is to win by building the best products and it's marketing that will bring people to take a look - pure sales has only limited value to Apple's market development.

The economic argument
For a quick example, I'm going to take a small company of 50 people to see what my Microsoft solution will cost for messaging and collaboration services. Of course, making this viable does depend on Apple stepping up to the plate and delivering with Snow Leopard Server. But even without 100% transparency, the relative value equation is still pretty impressive.

OS X Server vs Sharepoint (source)

	Apple	Microsoft
Server platform	OS X Server: $999	Windows Server 2008 Standard: $999
Collaboration platform	included	Office SharePoint Server 2007: $4,424
User licences	included	Office SharePoint Server 2007 Standard CAL $94*50=$4,700
IDE, development tools	included	Office SharePoint Designer 2007: $187
Database engine	included	SQL Server 2005, Workgroup Edition $3,899/processor = $7,798
Totals	$999	$18,108

I've only chosen the entry level versions, even though the Snow Leopard collaboration services offer deep search capabilities that require
more expensive versions for Sharepoint and you have no additional costs for putting up your server on an internet facing connection.
A dual processor server seems the minimum reasonable investment today.

OS X Server vs Exchange (source)
Now this is a more complicated question in the enterprise, since in larger environments the messaging infrastructure can be quite complicated, and Exchange does have a serious edge here in being able to manage user accounts distributed across multiple servers and stuff like that. But for the fledgling enterprise, a single server is more than capable of handling the task up to about 2000 user accounts per server on Exchange. There are of course additional complexities to look into like the fact that you will require a separate mail relay or proxy for handling incoming mail since you can't filter on the server that handles the mailboxes, but I'll assume you do this on the cheap with a Linux box rather that a second Exchange server.

	Apple	Microsoft
Server Platform	OS X Server: $999	Windows Server 2008 Standard: $999
Messaging Server	included	Microsoft Exchange Standard: $699
User licences	included	Microsoft Exchange Standard CAL $67*50=$3,350
Totals	$999	$5,048

I've heard a lot people argue that they don't want to use OS X Server since they don't have access to people that know how to run it. But given the price differences, you can easily afford to train your administrators from the savings and still have money left over.

Waiting for Snow Leopard
The decision points for the enterprise purchase will depend greatly on the eventual extensibility of the collaboration services and the level of integration with Outlook. But it seems that Apple has a very well developed product strategy for offering value to the enterprise, not just a sales strategy...

VMware Wants to Bring Virtualization to Your Smart Phone:

"VMware is looking to bring its virtualization technology into the smart phone market in 2009, with a small hypervisor that will allow nearly any mobile device to run multiple virtual machines.

On Nov. 10, VMware will announce that it will release a new virtualization platform called the Mobile Virtualization Platform, or MVP, which consists of a small, bare-metal hypervisor—20KB to 30KB—that will work with a number of mobile devices based on an ARM processors"

(Via eWeek.)

Obviously someone thinks that there's a market in here somewhere, but I have to start wondering about the new CEO's time at Microsoft and their obsession with all things new and shiny.

What's the market for a mobile virtualization platform? Since you can only run one of the currently available mobile OS solutions compiled for ARM, your OS list is pretty limited.

This also flies directly in the face of why x86 system virtualization actually works: the x86 server is a commodity item that is generally overpowered and underused. There's really very little difference between various servers other than the form factor and expansion capabilities. The UI components can be subsumed into KVM emulation and you're all set. Virtual machines don't pretend to be a DELL 1950 or an HP DL380 - they're just a generic motherboard with a given set of standard components.

In the cell phone world, you're confronted with massive variety - it might seem like a commodity environment since there are so bloody many of them, but the differing form factors (screen size), UI ergonomics (physical keyboard, 5-way button, extra buttons, ...), advanced features (pressure based touch screen, capacitive touch screen, multitouch ability, ...) mean that the OS you load had better be the one optimized for the hardware you're running it on.

Why do I want to switch operating systems on my phone? To play Windows Mobile games on my Blackberry? To test the latest version of Android without blowing up my current version? (OK - that one might have legs), Run Android on my iPhone? Oops - no physical keyboard.

And on top of that, we're a long way from having phones that are overpower and underused.

Honestly, I just don't get it. Comments or mail, someone please explain to me the market.

Or is this just a strategic decision to distract Microsoft and send them chasing down a rathole?

Up today: General Motors.

This one actually would make a whole lot more sense that buying Sprint. GM is a product company, although arguably it's two companies - a concept company and a production company and never the twain shall meet.

GM, like so many automakers has come out with some astounding and beautiful creations, but somehow, the trip between the concept group and the production side of the house seems to involve a round trip to Mars, followed by passing the plans through a shredder, taping them back together, losing them, finding them in the compost heap, peeling off the tape and deducing from the traces left on the tape the original plans.

From a strict asset point of view, GM is a great deal for someone who wants to buy it and sell off the assets at the current market valuation of $2B. But GM needs a lot of things right now first and foremost, cash. That's something that Apple could certainly help with, but given GM's current burn rate (close to $7B), I'm not sure that the purchase would be a good thing given the rest of the economic climate.

Automobiles are a necessity for many americans, but are also a luxury purchase in many ways. Given the economic slump, new cars are going to be pushed to the bottom of people's shopping lists. From an investment point of view, putting money into a local garage is probably a better bet. People are going to be doing everything they can to keep their current car running longer rather than get tied into a loan for a new car.

However, there are a lot of similarities between Apple just before Jobs arrived, and GM today. Both bleeding cash left and right, with no real sense of direction and a ridiculously scattered product lineup that doesn't match consumer expectations or needs. Or more precisely, is trying to be everything to everyone. 95 individual models in the US market alone! That's insane. Plus there's Opel and Vauxhall in Europe which add another 32 models.

GM also desperately needs help in making the migration to newer environmentally friendly technologies, similar to Apple's migration away from OS9 to a modern multi-tasking OS.

I don't know if GM would be a good investment for Apple, but I would definitely say that GM desperately needs to do exactly what Jobs did at Apple upon retaking the reins:

• Pick a core fleet of products and get rid of everything else. Sell them or kill them, but focus on something
• Identify a growing market as the key to moving forward. That would be hybrids or all electric, but pick one and run with it. Distraction is the killer.
• Learn the mantra "real artists ship". Concepts should be built with an eye towards production.
• Get an identity. 11 brands and over 100 models is just too much.

GM also has a whole slew of other issues to deal with involving their current union structures and pension plans, of which I know very little, but would definitely have an impact on any purchase or investment.

I think that Jobs has the intellectual capacity and management rigor necessary to put GM back in the black, but it seems to me that this would be an unnecessary distraction for him while trying to navigate some very difficult economic times. On the other hand, this just might be the perfect opportunity to see how Apple's management team handles things while Jobs is busy elsewhere, but not out of the loop.

Answer - not sure if it's a good idea, but it would certainly be interesting...

Now I'm curious here - according to this article it seems that there are a number of people having email issues on their iPhones, but I have yet to run across any in my immediate entourage of iPhone using colleagues and friends. Here's the summary of the problems:

So here’s the follow-up post to explain what went completely pear-shaped with the process of moving to an all-Apple device and network tech deployment at home….

This particular posts refers to one of the services I hold most important - that of email handling, and delivery across multiple devices… so where do I start….?

Well, it’s important to first of all say this is the iPhone 3G, Windows XP, and Outlook 2003 in the mix here. Outlook and the iPhone are acting as pop3 email clients.

Some of the issues noted are:

1) iPhone would state “connecting” when picking up email, but hang at that point and not collect any mail

2) iPhone could not find incoming (pop) server

3) iPhone could not find outgoing (smtp) server

4) iPhone would not collect mail on schedule (e.g. if hourly collection setup, actual collection is at iPhone’s discretion!)

A couple of important points:

• Do not, repeat, do not use POP3 on a mail account accessed by multiple devices. This is 2008 going on 2009 and there's absolutely no reason to be using POP3 any more. This is what IMAP was designed for (among other things).
• POP3 doesn't allow multiple concurrent accesses to a mailbox. If your Outlook client has opened a connection, the iPhone is going to block trying to access the mailbox. Use IMAP.
• You talk a lot about the mail clients involved here, but nothing about the mail server or network configuration. It's kind of important to include these details when trying to figure out why things aren't working the way they ought to.
• The mail collection schedule is overridden when you open the Mail application on the iPhone.
• In most cases, the issue of not being able to find a server is almost always a DNS or network issue and nothing to do with the mail client in question. You have to remember that the iPhone is happily changing network environments every time that you switch between Edge, 3G and Wifi and you need to be aware of the context. If your mail server is located on your home or office network, but is available over the internet, it's a given that you have two different IP addresses for the server - one for when you're on the local network via Wifi, and another while you're coming in over the internet. In that case, you have to look out for caching issues. If you established a connection from your Wifi network and then tried to connect from a 3G connection, you may be looking for the cached address that you can't access from the internet.

The brute force fix for this kind of connectivity problem is simple: Active Airplane mode, wait for the operator name to go away, and then disable Airplane mode and you have a fresh network connection with caches cleared and ready to go in the appropriate context.

Now this a way cool new application on the App Store. I haven't tried it out yet, but this may be exactly what I've been waiting for.

Snatch is an iPhone application that talks to a client on your Mac (no windows version that I'm aware of yet) that lets you use the iPhone as a remote trackpad. It's not a VNC remote screen sharing application - just the remote trackpad and keyboard.

I use a Mac Mini connected to the TV as my main multimedia station in the living room and it works great, but there are some times that I butt up against the limitations of the Apple Remote, especially when I'm using other media applications like EyeTV (which rocks in this configuration).

Snatch will let me go beyond just the iTunes Remote application and just drive the whole machine from the couch as required.

More details here:
http://theappleblog.com/2008/11/06/using-your-iphone-as-a-trackpad/

and more info once I get a chance to play with it...

I've been looking for a decent flexible in car mounting solution for my iPhone and my wife's iPod Nano 2G for a long time now. I used to use a generic windshield suction mount, but it was ugly, tended to move a fair bit and was finicky to align stuff just right for different sizes iPods and the like. Since we just upgraded to a Cooper Mini I've also discovered that the windshield is a little far away to conveniently mount the old generic mount without adding some kind of extension to bring it within easy reach, so we've being bumbling along with the iPhone in the cupholder - a far from ideal arrangement.

I've been tripping over references to the Brodit Proclip line of mounting products for a few years now, and last week a colleague showed me the one he had mounted in his car and that closed the deal. The mounts are custom made for individual model cars, taking into account the specific ergonomics and design of the dash, and then you buy a separate clip that's also custom designed for your model of iPod/iPhone/TomTom/GPS/you name it.

The difference between a custom made clip and a generic one is astounding. The iPhone fits perfectly into the felted clip, where it's held without any kind of stray movement. This is really important for the touch interface of the iPhone since if it's jiggling around while you're trying to push a button it's exceedingly frustrating. With the iPhone firmly in place, it feels like it's part of the car - not the slightest bit of movement anywhere.

All of the parts are made from high quality ABS plastic and it comes with all of the necessary mounting screws (although a couple of extras would be a good idea - if you drop one and don't spot it on the first bounce you'll spend some time on your hands and knees looking for it), and very clear instructions for the mounting of each individual component. Note that the screws are only for attaching the separate pieces together - no screws in the dash. The actual in car mounting is designed to fit perfectly with the dash in your car so that you don't do any damage to car dash. Although the mounts are so specific, you'll probably leave it with the car when you sell it.

The reason I'd gone with the generic mount before is that I had an iPhone and my wife has an iPod Nano so we needed something to use with both. Obviously, the custom route seemed counterproductive since we'd need two and here's where the Brodit solution shines again. They offer a set of optional matching attachment plates that attach to the car mount and the custom clip so that you can quickly and easily swap them around. Again, the quality of the construction is such that once the two plates are mated, it's as solid as a bolted on connection but still easy to pop in and out as required.

The biggest problem with the Brodit (and the main reason that I hadn't already bought one) is that their distribution network is a little hit and miss. I'd tried to find a local distributor here in France without much luck and you can't buy directly from the Swedish site.

Oddly enough, I can order from the Proclip USA site and items are shipped from Illinois. Which strikes me as being decidedly strange since the packaging clearly states "Made in Sweden". I'm feeling very unecological about that part since the items had to make a round trip from Europe to the US and back, but they get full marks for rapid (and reasonably priced) shipping via FedEx, including all duty and taxes so there were no surprises on receipt.

Combined with the Kensington LiquidAUX iPod connector, it's the perfect solution for any car with an AUX Line in, and even though it's more expensive than many of the alternatives, it's still a lot less expensive that some of the optional iPod integration kits.

It wasn't cheap at 130 € for the car mount, multi-mount plates, iPod and iPhone rotating mounts and shipping, but it's well worth it and a steal compared to the optional Cooper Mini iPod package at 400 € that doesn't even include a mounting kit. The pure iPod kits are great for the complete stereo integration, but if you want to use the iPhone as a GPS as well, putting it in glove compartment really isn't terribly practical.

I can't recommend the Proclip enough.

An often overlooked feature of the Airport Extreme is it's ability to work as a Network Attached Storage (NAS) device. That's fancy speak for a simple file server on your local network.

It's not the most sophisticated solution as far as these things go, but it's certainly all that's required for most people, especially for basic tasks like sharing a disk for backups, whether using Time Machine or SuperDuper to a disk image. You can manage sharing disks with a password for access to all disks or setup individual user accounts, each with their own password.

One issue with the Airport Extreme is that there's only one USB port, so if you want to share multiple disks or a printer at the same time you need to get a USB hub which means yet another power plug taken up and more unsightly cables messing things up.

I've been happily using the Minipartner disks for my IDE drives for a number of years, but these disks only work over Firewire, even though they integrate a USB hub for simplifying the connection of multiple USB devices. The latest generation are more flexible and have the option of presenting the internal SATA disks over Firewire or USB, and they still have the integrated USB hub. This makes them absolutely ideal for use with the Airport Extreme since you can easily daisy chain several disks using the USB connections and they're exactly the same shape as the base station so it's an elegant stack of devices, especially when you use the very short USB cables to attach each disk to the next.

With these disks you have all of the advantages of simple daisy-chaining like you used to do with Firewire over a USB connection and you can share multiple drives from the Airport Extreme without an extra external USB hub in the mix.

Now if only I could arrange the disks to format them into RAID groups directly from the Airport drive so that it could manage larger volumes than a single disk at a time and it would be the perfect inexpensive home NAS solution.

Of course, a Drobo hooked up to the Airport Extreme is the ideal solution, but the entry price is a little steep, and I still want to see one that handles more than 4 disks - maybe an 8 drive system that uses 2.5" disks?

That way I'd lose less to the parity calculations on the RAID so eight 500Gb drives would give back 3.5Tb of useful space instead of 3Tb from four 1Tb disks. Not to mention that I seem to be collecting 2.5" SATA disks faster than 3.5" ones.

Update [4-nov-2008]: Just noticed that the Airport Extreme is smart enough to put attached disks to sleep if they're not being solicited across the network.  

Now that the announcement has been made and we all know what these machines offer there has been the usual wave of reactions, both positive and negative. I think that there are a number of things that should be clarified with the disappointment evoked by various folks in the Mac and IT blogosphere.

Matte or glossy

The first one is the lack of a matte screen option. Personally, I'm more comfortable with a matte screen, but the latest generation of LED backlighting appears to be capable of pushing enough lumens such that even in bright environments, the screen is still more than readable. I'm not too concerned about that part.

I'm more concerned about glare and reflection but again, this is usually manageable by repositioning the screen.

For the folks that are concerned about color accuracy I think that argument needs to be taken with a big grain of salt. Anyone who's truly serious about this works with a color calibrated screen in a work environment that takes this requirement into account. Controlled lighting, up to and including screen shades etc. You're never going to get color accuracy on a notebook screen in multiple environments with varying types of lighting.

I'm sorry to see this option go away, but there's already a flourishing market in anti-glare films for all sorts of monitors and with the more powerful LED backlighting, the reduction in luminosity should be manageable.

Where'd my Firewire go?

This is the one thing that I will miss the most. Now, remember that it's only the MacBook that no longer has a Firewire port - the Pro model comes with a Firewire 800 port that's backwards compatible to the 400 standard.

It's worth spending a quick moment here to identify the market segments that Apple is targeting with each model. The MacBook is their entry level machine aimed at the affluent1 mass market. In this market, USB is the predominant interconnection protocol. You have to look hard to find portable disks with a Firewire interface, and the cost of the controller chip and licencing is such that these drives are notably more expensive than the equivalent USB models.

Most of the blogosphere pundits are, at the very least, prosumers. They don't buy entry level models and we're hoping to have their cake and eat it too with a cheap MacBook Pro.

Those that are complaining that Firewire is the preferred video transfer protocol are correct - in the professional and prosumer marketplace. But in the general consumer space, USB is making massive inroads with cameras like the Flip Mino.

What will be sorely missed is the Firewire target mode. But I suspect that this loss will be mitigated by the ease of access to the hard disk. Swapping a hard disk is now the work of opening a latch and unscrewing one screw. Well within the skills of almost anyone. Popping the drive into a cheap external USB box doesn't demand any huge technical skills either. I will miss the Firewire target mode more for troubleshooting - without removing the drive, I can plug the machine into another and run my disk utilities against it.

DisplayPort vs. HDMI

Another recurring point is the lack of direct HDMI. Apple is doing what they always do and pushing the newest, most open standard technologies out into the marketplace. HDMI has a number of technical limitations that make it less than ideal for offering a maximum of flexibility to a maximum of users with varying requirements. Currently, the DisplayPort can be used with a direct connection, or a DVI or VGA adaptor, and it's easy enough to connect DVI to HDMI2. To the best of my knowledge there are no portables on the market that offer onboard HDMI out - there are a few desktop motherboards that do, but nothing in the portable market.

I'm of mixed feelings on the fact that you no longer get all of the adaptors out of the box, but this reflects the reality of usage. The majority of users connect their portable to one screen via one technology and will buy the appropriate adaptor. It's the road warriors like me that use VGA with projectors on client sites and DVI at the office who get the worst deal. On the other hand, I have a drawer full of adaptors from all of my old machines that I've never used.

Summary

I think Apple has a fairly good idea of the markets they're targeting and needed to find a way to differentiate the two models. I also know that every time you send a crash report back to Apple it includes the System Profiler dump which includes a list of everything that's connected. I suspect that they have some some statistical analysis on this information, plus user interviews to determine that the majority of current MacBook owners use few, if any Firewire based peripherals.

It's only people like me who have a foot in both the "personal consumer on a budget" and "demanding power user" camps that are now forced to make a decision on just how much that additional port and screen real estate mean to me.

Recently, I've been evaluating how I use my current MacBook Pro, and I've discovered that for mobile use, I only use USB peripherals3. Now I have the choice between a MacBook Air and a MacBook for a powerful, lightweight computer.

My desktop machines however, have multiple daisy chained firewire drives and there I do want and need the additional performance and flexibility.

A MacBook for the road, coupled with an iMac or Mac Pro seems to be the ideal way to go right now.

1. It's worth recapping the points made by Tim Cook regarding where Apple positions itself in the market. They don't aim at the ultra cheap bottom scrapers.↑
2. Although I'm told there is some minor signal degradation.↑
3. I currently don't do any video outside of the odd recording with the built-in iSight. I'd probably feel this issue more if I had a SONY camcorder with Firewire.↑

From the Compellent press release

Delivering Performance of SSD for Virtual Data Centers

For enterprises requiring the fastest storage performance for data management, migration and processing, Compellent’s integrated suite of virtualized storage applications will also support solid state drives (SSDs). The Compellent SAN will reserve frequently accessed, active blocks of data for “tier 0” storage for applications like transactional databases that can take advantage of the significant performance gains of SSDs, and dynamically move inactive data blocks to lower storage tiers. Because the Compellent SAN is expected to be the first to automate tiered storage for SSD, customers will be able to accurately plan SSD purchases along with other drive technologies to significantly reduce total costs while maintaining optimal storage utilization and performance.

This was an obvious next step for Compellent and an exceedingly attractive one for the enterprise buyer. The biggest problem that all storage vendors have is that while they offer so called "tiered storage", you have practically no useful control over the use of the different tiers.

You're stuck with granularity at the LUN or file level, neither of which makes the SSD investment particularly attractive, especially if you're using VMware with the VMFS where your smallest file object is a VMDK file.

Introducing SSD at the dedicated LUN level is staggeringly expensive, and horribly wasteful except for very well targeted applications. But when you combine SSD as your new "Tier 0" storage linked with the integrated ILM or Automatic tiered storage functionality of Compellent's bays, a reasonable investment is SSD storage can be shared by multiple servers and applications transparently, but only for the subset of their data blocks that actually require this kind of performance.

Less active data block will be migrated automatically down to your "slow" 15K FC disks and eventually down to SATA.

I think that this is a really big deal that further justifies the value of the Compellent solutions. Now the other question that begs asking is power consumption on SSDs. There was quite the kerfuffle about that a little while ago for SSDs targetted for portable use, but I suspect that an SSD will consume a lot less power than a 15K FC disk that's spinning a full speed all of the time.

VMware acquires Blue Lane: "virtualization.info has just leaned that VMware acquired the security vendor Blue Lane Technologies.

The company, popular for its inline patching technology, entered the virtualization market at the beginning of 2007 and completely refocused its effort around the VMware Infrastructure over the last two years."

(Via virtualization.info.)

Wow. This is a really big deal with the upcoming VMSafe APIs from VMware. This should get rid of the current issues around this kind of appliance that requires that the box be inline on your network. I haven't used it directly myself, but I have a few customers that swear by it.

Since the virtual switches and virtual networking components are nothing but memory constructs inside your ESX server, you could patch your entire environment by deploying a single signature update to your Blue Lane instance inside the ESX Servers - and equally important, pull it out if it causes problems without touching any of your virtual machines.

This is extremely important since much of the critical patching on Windows servers are security patches. If I can externalize that, and stop patching my servers directly this eases a major pain point for Windows administration.

This spills over into other stuff like Template management that you can stop patching manually or keeping multiple generations of templates based on application compatibility issues. This obviously ties in nicely with the cluster-wide vSwitch management so that the same rules are applied globally across a cluster.

Now the big question is going to be about the granularity of control - will you/can you specify patches by individual VMs, by port groups or vSwitches? Oh - and pricing.

iPhone the most popular ebook reader: "Forbes is reporting that the iPhone has become the country's top eBook reader. The claim is based upon the number of downloads of Stanza [App Store link], an ebook reader for the iPhone and iPod touch (395,000 and counting), versus the number of projected US Kindle sales (380,000). "

(Via The Unofficial Apple Weblog (TUAW).)

I can add that everyone I know with an iPhone has a copy of Stanza installed and that it's a major winner with the public transport commuters here in France. And that's only Stanza - there are a number of other eBook readers for the iPhone, but none as popular (or as good, in my opinion)

The big question is whether this will make Steve Jobs reconsider the statement that "nobody reads any more". You must consider what he actually said which was quoting statistics that "Forty percent of the people in the U.S. read one book or less last year" which is a sad indictment of the modern state of affairs, but it leaves out a couple of interesting factors.

People are reading fewer books, but getting more and more of their reading material from electronic online sources and reading on electronic devices

The people that do read, read a lot

Audiobooks

Apple has recognized the potential in some specific markets with the inclusion of audiobook in the iTunes store via Audible (although I seem to recall seeing something about Amazon buying them earlier this year so it's up in the air if that relationship will continue). Audiobooks are the ideal "reading" method for an automobile based commuter where your hands and eyes are busy. It's also the obvious extension of iTunes since it's all the same kind of media.

But as people start tightening their belts and dealing with rising gas prices and various other economic woes, public transport will (or should, anyway) start rising again. Then you have your hands and eyes free again to flip through a book or magazine on the bus or train (except those unfortunate enough to suffer from motion sickness while reading in a vehicle).

Who's reading what?

The change in habits towards more electronic reading as opposed to paper is becoming ingrained in the younger generations - they may read fewer books, but they often have pretty impressive RSS feed lists. There's also a cultural shift going on - in my generation and older, you read for pleasure on paper and what you saw on a screen was more likely to be work. The younger generation doesn't make that distinction and if they do, it's in the opposite direction - paper books equals school and electronic reading is entertainment.

The people that do read books read more than their fair share when you consider that less than half the US population read a book last year. That subset of the population is keeping the publishing industry in business. That looks like a market to me - and I suspect that there's a serious intersection in the market of iPhone owner who are generally affluent, educated and likely to be readers.

Where to go, what to do?

Just like music before iTunes, the current state of online eBook purchasing is a pretty scattershot affair with little consistency of formats and rights management, and just like music, with a surprisingly robust peer to peer sharing community.

After trying out Stanza on the iPhone, I started seriously thinking about designing an eBook library management tool so let me clean up some of scanning errors and metadata issues as well as add in additional information like series and universe tags with ordinal tags so you know what order to read them in. But the more I started mocking it up, the more it started looking like iTunes. As soon as I see that I start to wonder if it's worth heading in that direction when it's a very likely next step for Apple.

I think right now that there's a market waiting to be pulled up out of the fringes and into the mainstream. Apple's got all of the necessary pieces in place with the iTunes Store infrastructure - all that needs to be added is content and a reader. They can placate the powers that be in the publishing industry with Fairplay protected material.

iPhone vs Kindle

Now a big question that needs to be addressed is the Kindle as competition. But it has a few failings that I've outlined in other articles concerning mini notebooks and UMPCs. It's just a little too large to fit conveniently in a pocket and it's (despite the browser) basically a one trick pony. The iPhone's convergence of all of my mobile media consumption into one device is a game changer. Each morning in the train I have the choice of listening to music or podcasts, catching a TV show recorded via EyeTV, checking up the news via my RSS feeds, reading a book, playing a game...

Compromises

Now one point that keeps coming back up is the size of the screen. Yes the iPhone's screen is relatively small, but with a reasonably sized font, the reading experience is such that you forget very rapidly that you're reading on a screen. There's a compromise here and that's size. Remember when everyone first saw the MacBook Air and said it was horribly overpriced and underpowered? Then the second wave of reviews started coming back in noting that thinness and lightness really were important features that shouldn't be underestimated. I feel that way about the iPhone. Would I like a larger screen? Sure, but not if it means that I can't pop it into a pocket. If someone comes up with a flexible OLED touchscreen that you can fold or roll out with no seams or perhaps an iPhone DS with just a hairline hinge it would be practical, but given the current technology it's the best compromise we have right now.

Next steps

My suggestion would be for Apple to buy eReader to acquire their distribution channel and Stanza for the openness of handling multiple formats. Those two assets with some relatively minor tweaks could be integrated very quickly into the iTunes media universe and we'd find out just how many people like to read.

But even if they don't, I'm more than happy with my current toolkit for the moment.

Epilogue

There's still no replacing a good paper book in front on the fire on a winter evening, or loafing in a hammock in the summer with an appropriate beverage at hand, but for all those times I'm stuck in a lobby, a train, a bus or just have a few minutes to kill between meetings, Stanza on the iPhone is just about perfect.

Ideas for Apple: AirPort Mobile Tethering: "The previous segment looked at how Apple is uniquely situated to deploy ubiquitous WiFi by offering iTunes affiliate commissions to encourage casual AirPort sharing. A second issue Apple is poised to solve relates to iPhone tethering. Here’s what the current problem is and how it could be solved"

(Via RoughlyDrafted Magazine.)

An interesting article from Daniel Eran Dilger that bears some reflection. I think that there's definitely some legs to this idea, especially with the integration of a microphone with the latest iPod Touch. Some interesting variations on this theme are already starting to crop up here in France.

We've seen that with some of hacked Linksys firmware options that you can publish more than one Wifi network from the same base station, permitting you to open a public channel to the internet that coexists securely with your own private network. This approach has become a standard procedure here with the French ISPs, notably Orange with the UNIK service and Free.fr with the Freephonie. Basically, these services offer a standard VOIP service for all subscribers as long as they are in range of another client's Wifi network. The Free service has already been beta tested with the publication of the Siphon iPhone application, and they have promised a version on the App Store this fall.

UNIK goes a step further and integrates this with their 3G telephony service so that you can switch between 3G and Wifi communications and are reachable on your 3G phone number, whatever the connection.

The promise of ubiquitous Wifi is unfortunately far from the reality that I (and others) hoped for. Most providers have horribly expensive offers and most are protected by pretty lame web authentication front ends, making seamless use impossible.

The Fon service is an interesting foray into this world, but not the sort of thing that your average consumer is likely to jump onto and is hobbled by the same authentication issue.

If we want to see ubiquitous Wifi, we need to bring the average consumer into the fold, in a simple manner. Apple is perfectly positioned to push out a firmware update to enable this kind of option. The question of payment or lack thereof is a huge one. I'd like to see an option to publish a freely accessible network, with an option to either do some QoS or bandwidth throttling in order to make it less interesting for P2P freeloaders, but sufficient to do VOIP.

I think that it makes the most sense to see this tied to the MobileMe service than to Apple Store user accounts and then it becomes yet another reason to justify buying into the service.

From a publishing standpoint, I would be more than happy to to set aside a little bandwidth for anyone nearby that needs the connection, and I would be doing so right now if the tools didn't conflict with my current rather complex Wifi setup.

As a consumer, I want a seamless connection from my portable, iPhone or iPod Touch. I don't want to be nickelled and dimed to death with usage fees and if there's a flat fee to pay, it had better be reasonable.

Thinking it through, it seems to me that the simplest method would be a tit-for-tat approach for MobileMe account holders. You configure your Airport Base station with your account, which advises Apple that you are a participant, thus giving you access to other enabled Airport Base stations.

This would be an excellent method for Apple to add some serious value to the MobileMe service with minimal infrastructure additions. They already have the back end LDAP authentication infrastructure1 which serves the me.com web services, the App Store, the iTunes Store, the developer site, the discussions forums... and the Airport firmware already supports RADIUS authentication - it's just a question of ensuring that the servers can handle the additional load.

Once the service is in place, that really opens the door for VOIP services from your iPhone or iPod touch. Apple could arrange to put in place a mapping system so that people could "call" your MobileMe account which you link to your phone number as a connection of last resort, checking if you're online on a Wifi connection first. This may be why we still haven't seen iChat on the iPhone...

What really works with this idea is that Apple is the only company capable of offering this service since they own the whole stack: hardware, software and service.

1. I don't know for sure that they're using LDAP, but it would be the logical technology for them to be using. ↑

How Many IOPS?: "A question I get asked occasionally is; 'How many IOPS can my RAID group sustain?' in relation to Enterprise class arrays."

(Via The Storage Architect.)

An excellent little reminder about how to estimate your storage bay's performance capabilities. It also makes me wonder if we aren't approaching the question of shared storage the wrong way. Instead of asking each vendor what the best performance they can get, how about we ask them the worst we can expect?

Everyone knows that you can tune a storage system and the IO test profile to pull out massive IOps or MBps (but not both at the same time). Wouldn't a more realistic measure of the value of the storage system be based on how it performs when being bombarded with a completely random IO profile?

This approach (while somewhat brutal) looks to carry more value as an analytic tool for sizing storage used for virtualisation projects where all of your different IO profiles turn into a smear. If you buy your storage based on the worst case, rather than an optimistic evaluation of the maximum capacity, you'll know that in real world use you'll come out ahead.

We know that we can calculate with some precision the worst case IOps per disk based on the measured latency - what we need to know is how much the value is being added by the array with the various tiers of cache and the algorithms used.

On paper, a standard RAID group of 5 disks (based on the FC disks noted in the Storage Architect article) we can count on a minimum of 925 IOps. Obviously we rarely see performance this bad since the disks themselves also carry 32Mb of local cache, and then there is the cache on the controller and so on up the chain (not to mention that real-world IO is never totally random, even with the smear factor taken into account). I think that the question we should be asking storage vendors is: How much value do you add to the equation? (1)

I think that instead of trying to tune our bays and tests to determine the maximum values, we should develop a multi client test where the bay is bombarded by random IO of varying sizes over a representative surface in order to see the delta that is achieved by the intelligence of the storage solution.

Ideally, I'd like to see storage array's performance rated as a factor over the baseline capacity of the physical disks. As Mr. Evans notes in the article, IOps are a function of speed, not capacity. The buying decision then comes down to two criteria :

Speed factor

Using a performance factor of two, this means that under maximum stress the storage array should be able to double the number of raw disk IO operations. For a 5 disk RAID, the 925 IOps guaranteed by the disks are multiplied by 2 for a baseline of 1850 IOps.

Storage volume

Once I have my IOps requirements, I simply choose the size of disk required to cover my storage needs.

I've already done some basic benchmarking based on this approach using multiple virtual machines (5-20) each using a different IOMeter profile with a mix of read/write operations, plus wildly variable different block sizes. On a name brand bay the results were pretty painful and the performance factor was actually a negative value (less than one). I did the same tests on a storage virtualisation solution that gave a performance factor of 2.7. (2)

Now we need to learn what an averaged IO profile looks like for a virtual server implementation so that we can add in a reasonable "real-world" usage factor. My currect guesstimate is that this will be on the order of 3-5, but I'm being conservative here.

Other issues that come into play in order to build a useful benchmark structure are details like standardizing a number of RAID groups in order to validate different possible optimizations, the use of multiple LUNs based on the same RAID group, IO going to one LUN or shared out over multiple LUNs, number of active paths to the storage array, number of concurrent IO sources, both physical servers and virtual machines, ... Obviously there are a lot of factors to consider, but if we can normalize a bunch of them, this could be a very useful metric.

Comments anyone?

FOOTNOTE 1 ↑Noting of course that raw performance is only one factor in the purchase decision. Price, maturity of management tools, replication options, support, etc. also need to be considered.

FOOTNOTE 2 ↑ Names not given to protect the innocent/guilty.

This is an perfect example of a well designed little utility that does one thing well, without extras or confusion. Anyone who's used synchronization software has run into either one of two situations : You end up with an empty calendar, or you end up with duplicated of everything.

It's been a while since I've had any serious problems with either of these problems (a testament to the maturity of the tools on OS X), but I still have a number of archive calendars that are chock full of duplicate entries from older transactions (mixing google sync, Palm Desktop, Exchange, Entourage and .mac).

I can't remember which web site pointed me to this utility, but it's an absolute must have if you want to clean up your iCal calendars without starting from scratch or passing hours paging through screens and deleting everything that you see.

iCal Dupe Deleter to the rescue. A one trick pony that checks your calendars for duplicates and one click cleans them up for you.

Highly recommended.