Desire to Focus on IPS

I like stuff which is well designed (minimalist), well engineered, high quality, with well thought out usability.

My latest gadget arrived on Friday - a HTC Desire HD and I'm tickled pink with it.

It replaces my Nokia 5800 XpressMusic whose screen I cracked in the gym a couple of weeks ago.  For those who know me, they'll read that last sentence with incredulity.  But yes, I attend the gym once a quarter whether I need to or not. 

I put the Nokia in the rectangular holder on the exercise bike and it fell straight on the floor.  Thinking I'd simply missed the holder, I picked it up and put it in again, only for it to fall straight back to the floor, hitting the leg of the exercise bike and cracking the screen.  Looking into the holder, I saw it's "designed" with a big hole in the bottom of it - presumably sponsored by the local phone repair company.  Who designs a holder for iPods/phones with a ruddy great hole in the bottom of it ?  Arrgh.

Anyway, rather than pay EURO 80 for another new screen, I was due an upgrade so it was time to get a "smarter" phone.  The Nokia XpressMusic is great for playing music, but I found it useless for web browsing due to it's glacial connect times (which may be at least partially due to my local provider), slow loading, and small screen.  But the Nokia's web capabilities did come in handy in an emergency when I got stuck in Spain last year due to the Icelandic volanic ash cloud.

I spent a couple of evenings googling smartphone reviews and reading everything I could find. 

Since I purchased an iMac last year, and Santa Claus brought iTouches for the kids, I was leaning towards the iPhone4, perhaps partially due to Apple's slick marketing.  But one of the reasons I bought the iMac (apart from my natural UNIX affiliation to MacOS, the 27" screen, cool design, and my engineers telling me how cool everything Apple is), was exasperation with Microsoft for continually changing how to change/fix settings on its various releases.  Being a 25 year UNIX veteran, I'm pretty amateurish at finding my way around Windows, but having grasped the basics of XP, I found it really frustrating having to re-learn how to do the same functions on Vista and my father-in-law's Windows 7 laptop.

While I'm very happy with my iMac, the Apple iTunes lock-in and Apple marketing machine gives me a slightly uneasy sense of deja-vu with Microsoft from 10-15 years ago.  Granted, Apple designs Operating Systems and hardware a hell of a lot better and I don't think any company will ever again get away with that sort of monopolization tactics.

Anyway, the consensus amongst most of the smartphone reviews was that the HTC Desire / Desire HD is superior to the iPhone4 and Samsung Galaxy S.  And being Android based, it satisfies my UNIX principles too.  And so far, so good.  It's intuitive, fast, well designed, with good apps, and all very well integrated.  Pity about the appalling battery life, but there's always a socket or USB connection nearby.  I carry the charging cable with me in my coat pocket.

So my HTC Desire HD joins my favorite gadget collection alongside the superb Panasonic Lumix DMC-ZS7 camera I bought at Christmas, the SunRay which I've used as my main computer at work for the last 10 years and which is truly excellent (allowing me to bring my desktop session with me anywhere in the world), my Pansonic G10 FullHD TV, the kids' PS3, and my all time favorite, my Rado watch.

The Rado fulfills the pinnacle of design IMHO.  Beautifully minimalist, utterly unscratchable (and believe me, I'm a Philistine - I do gardening and DIY with it on), and perfectly functional apart from the date field which is so small that it's very hard to read even in good light.  Four years on, and it genuinely looks like I bought it yesterday.  Truly a product which exceeds even my most demanding expectations.

This circuitous product review brings me to Image Packaging System (IPS), with which I'm currently working in preparation for the release of Solaris 11.

As I've mentioned in previous postings, Image Packaging System (IPS) is a single-tier packaging architecture which in Solaris 11 replaces the old System V (five), Release 4 (SVR4) based 2-tier package and patching architecture in Solaris 10 and earlier releases. 

IPS architects, Bart Smaalders and David Comay, spent a lot of time with me around the Solaris 10 Update 3 timeframe to understand the deficiencies in the SVR4-based patch architecture, and helped fix the issues around patching Zones and applying arbitrary change to a live boot environment.

Bart and David have used that deep understanding of the deficiencies of the SVR4-based patch architecture when designing IPS to ensure their design addresses these and other issues.  The result is a highly flexible IPS architecture.  Feature and process development is continuing as the target audience moves from developers in OpenSolaris, to ISVs and evaluators in the currently available Solaris 11 Express release, to meeting the needs of Enterprise production customers in Solaris 11.

You can learn lots more about IPS at http://www.oracle.com/technetwork/server-storage/solaris11/technologies/ips-323421.html and by trying in out in the current Solaris 11 Express release.

I, for one, will not be sorry to see the back of patches.  While my team and I have done our best to improve our customers' patching experience over the last decade, it's very difficult to make a silk purse out of a sow's ear.

Much of the work to be done between now and the Solaris 11 release revolves around defining and communicating the processes and best practices which we recommend customers adopt around maintaining Solaris 11. 

While we still have a lot of work to do, I look forward to adding IPS to my most favorite technology list.

Comments:

Gerry, Gerry...

http://www.faqs.org/faqs/unix-faq/shell/csh-whynot/

:-)

Looking forward to IPS as well...thanks...

Posted by Pete Hartman on March 14, 2011 at 08:26 PM GMT #

Thanks Pete,

I knew that new fangled shell wouldn't catch on. :)

Best Wishes,

Gerry.

Posted by Gerry Haskins on March 15, 2011 at 04:28 AM GMT #

"The result is a highly flexible IPS architecture."

I disagree, vehemently, and here is why:

we have hundreds of configuration packages. Configuration packages do most of their work within the installation context, that means preinstall, postinstall, preremove, and postremove. For those times where changes must be made on a system which is not yet running, and we are installing a configuration package via JumpStart(TM), sysconfig(1M) facility is used to register a transient program to do any post-configuration on a running system.

These packages differ from your run of the mill packages in that they have little to no payload as files, since most of the work is preformed in the installation context.

IPS, on the other hand, EXPLICITLY DISALLOWS such installation mechanism - by not supporting it AT ALL!

Which in and of itself would not be a problem, if IPS were tied into SMF completely, and if SMF had the capability to perform the tasks normally performed in package preremove and postremove.

But it does not, at least not publicly documented.

And because of these factors, it is currently impossible to create configuration packages with IPS.

You (Sun now Oracle) might have alleviated the patching pain (which was unnecessary to begin with, since new package revisions could have been used instead of separate patches), but by explicitly not having IPS have installation scripts, you've created another, far bigger issue: no configuration packages. Customers like ourselves with hundreds of those would be in big trouble if they go to IPS-based Solaris 11.

Here is an example of a typical configuration package:

imagine two firewalls. A customer wants to automatically create an IPsec VPN between them. A network configuration package configures the interfaces. A VPN foundation package configures /etc/inet/ipsecinit.conf and /etc/inet/ike/config with common settings for all VPNs. Finally, a VPN-specific configuration package configures IKE, IPsec, /etc/hostname.ip.tun0, routing for that tunnel, and restarts ipsec/policy and ipsec/ike (via sysconfig(1M) if the system is not live).

This is possible to do with SMF and IPS. But it is not possible to execute the reverse of these operations in SMF, because there is no way to detect when an IPS package is being removed!

We are in so much trouble because of lack of functionality in IPS. In so much trouble, that we might be forced to go to Linux. And we are not the only customer with this problem: a lot of financial institutions produce and run the same type of packages.

If IPS does not integrate with SMF better, or failing that, does not allow installation scripts, Linux will be in the cards.

Posted by UX-admin on March 15, 2011 at 07:09 AM GMT #

UX-admin, we've had this discussion before. Please note the following:

1) Actions tagged with restart-frmi cause the named service to be restarted when the action is added, modified or removed. Thus, removing a package will cause a properly designed service to restore the previous environment.

2) The SMF Profiles project is integrating shortly. This allows configuration information to be delivered to systems as xml manifests, and allows this information to be kept distinct from service manifest and administrative operations.

There are many reasons for the elimination of scripting; please review my and Stephen Hahn's blogs on the subject.

If you'd like to continue our previous discussion, please join me on pkg-discuss@opensolaris.org.

- Bart

Posted by Bart Smaalders on March 18, 2011 at 08:33 AM GMT #

Post a Comment:
  • HTML Syntax: NOT allowed
About

This blog is to inform customers about patching best practice, feature enhancements, and key issues. The views expressed on this blog are my own and do not necessarily reflect the views of Oracle. The Documents contained within this site may include statements about Oracle's product development plans. Many factors can materially affect these plans and the nature and timing of future product releases. Accordingly, this Information is provided to you solely for information only, is not a commitment to deliver any material code, or functionality, and SHOULD NOT BE RELIED UPON IN MAKING PURCHASING DECISIONS. The development, release, and timing of any features or functionality described remains at the sole discretion of Oracle. THIS INFORMATION MAY NOT BE INCORPORATED INTO ANY CONTRACTUAL AGREEMENT WITH ORACLE OR ITS SUBSIDIARIES OR AFFILIATES. ORACLE SPECIFICALLY DISCLAIMS ANY LIABILITY WITH RESPECT TO THIS INFORMATION. ~~~~~~~~~~~~ Gerry Haskins, Director, Software Lifecycle Engineer

Search

Categories
Archives
« April 2014
MonTueWedThuFriSatSun
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
    
       
Today