Tuesday, November 10, 2009

Into the sunset?

The announcement to EOL Solaris Express Community Edition (SXCE) was telegraphed well in advance, and we're coming to the end of the road with only a handful of planned releases left to look forward to.

But, is this just the end of the road for SXCE, or is it something bigger that's at stake here?

Read the Sun marketing and you might believe this is a glorious new dawn for the Solaris/OpenSolaris world. The reality may be more like sailing off into the sunset and disappearing from view.

The fundamental difference between the old and the new is instalation and packaging, which have been ripped out wholesale and been incompatibly replaced. Even if the replacements had been perfect (and, quite frankly, they fall a huge distance short) this would have been a huge challenge. Organisations (and individuals) are under huge pressure to retrench and consolidate. Adding additional technologies that they're expected to support is an uphill battle. Adding brand new (and essentially untested) technologies that they're going to have to learn from scratch makes it doubly hard.

If the next version of Solaris had been based around SXCE, with traditional deployment technologies - traditional packaging and jumpstart - then customers would have been able to start rolling it out tomorrow. Everything a customer knew, all their existing investment in skills and tools, would be preserved. New customers would be able to leverage the skills and expertise of existing customers. All the great features and functionality present in OpenSolaris would be there to be taken advantage of.

Contrast that with the planned OpenSolaris transition. You have to retrain all your staff, replace your entire toolset, rebuild your entire systems deployment and administration infrastructure. Most environments are heterogeneous, so this means you now have an entire extra set of infrastructure to support - you aren't going to transition everything to the new scheme immediately, so you're going to have to shoulder the burden of supporting the extra scheme in parallel for years. Isn't the most likely course of action for a cash-strapped IT
department with a CIO breathing down their neck to simply reject that and migrate everything over to RHEL?

Solaris and OpenSolaris contain fantastic technologies that make them a great choice for IT departments - ZFS, Crossbow, CIFS support, zones (especially sparse root zones), Dtrace, and many more - but by making deployment such an unattractive proposition we're making it far less likely that customers will try or use these technologies, and are giving organizations and managers every excuse to ignore Solaris and OpenSolaris as an option.

The best thing Oracle could do for Solaris and OpenSolaris would be to scrap the OpenSolaris distribution (but not the rest of OpenSolaris) and redirect our energies into building a better Solaris based on SXCE. If not, then I fear that Solaris will ride off into the sunset and be consigned to the wastebasket of superior technologies that failed due to bad strategic decisions, and that's a prospect that truly saddens me.


milek said...

Oh, I can't disagree more with you! Except for some old die hard Solaris admins Solaris/SXCE is actually hard to use and far from sys-admin friendly especially when compared to Linux. Open Solaris as a distribution is much better in this regard often being better than Linux counterparts in packaging management, etc. It actually makes life much easier and simple for all these environments which already have Linux in and imho thanks to Open Solaris the Solaris Next won't be perceived as that legacy dinosaur with couple of cool technologies...

For some environments which will required backward compatibility and more years to do transition to new packaging system, etc. there are Solaris 10 zones in Open Solaris developed specifically for them.

While there a still many thing to improve in open Solaris distribution it is already much better received by many sysadmins than Solaris has ever been, especially by those people with little or no exposure to Solaris before.

So I hope SXCE is not coming back and we can all focus on Open Solaris distro which I strongly believe is a way moving forward.

ps. but why did they scrapped jumpstart for that over-complicated replacement?....

Peter Tribble said...

Robert: my primary criticism is the change in deployment technologies, principally the removal of jumpstart (and that's tied to packaging). So, if you think that the replacement for jumpstart doesn't stack up, then you're agreeing with me 100%. In areas other than installation and packaging, there's precious little difference, and all the other changes are good, even if they are overblown (seriously, I've used a shell with command line editing for over 20 years, it's nothing new).

Rand Huntzinger said...

I fall between your position and mikek's, or perhaps it might be better to say that I understand both positions. There is a real difference between a development and a production shop. Developers tend to love change and quickly embrace new technologies. Production shops tend to be focused on getting their work done, which isn't computing per se, but their company's business. To them change, and in particular a disruptive change like this one, is a costly distraction from their real work. If change is needed, it should be slow and incremental so that it can be integrated with minimal disruption.

I don't think the transition to the IPS packaging system is a bad thing in the long term. That said, the transition is going to cause costly disruptions which won't go down well with production clients. I personally think that more should be done to ease the transition for existing Solaris shops.

milek said...

Rand - and for such shops there are Solaris 10 zones on Open Solaris. So one will be able to deploy a Solaris 10 zone on-top of Open Solaris.

Then is a customer is creating packages Open Solaris will make his/her life easier as spec files for Open Solaris for IPS are basically RPM spec files - it means more knowledge and experience can be re-used while building custom packages from Linux world.

Peter - re jumpstart - I'm not a big fun of the replacement but then I haven't looked for quite a while and I was looking at the very beginning. It might be that things are getting better. At least I hope so... the IPS packaging in Open Solaris is much better that the old packaging in Solaris and in theory one could make the good old jumpstart to work with IPS packages - that would probably be the easiest part. AFAIK there is no final decision that jumpstart is gone however it is likely.

And no, I do not agree with you Peter that Open Solaris distro should be scraped - actually I'm a strong advocate for the opposite which is finally hapenning. Especially the new packaging system will not only bring OS on par with other Unix systems but already delivers much better experience in some cases even if it is not polished yet - but we will get there.

ps. you can still install/build your old SVR4 packages on Open Solaris even without using Solaris 10 zones if you want.

Binary Crusader said...

Sorry, but the old SVR4 package and patch system has no future and is untenable from a support and engineering standpoint.

Even with significant reworking and support of the old SVR4 codebase, the very processes involved for packaging and patching SVR4 were costly and error prone.

No one likes change, but the new installation and packaging systems were necessary for the revitalization of Solaris.

Mark Allen said...

If you think SVR4 packaging makes OpenSolaris more attractive to system administrators and/or architects, you're just flat out wrong.

Packaging is something that's been *broken* in Solaris for a long time and I am so glad that IPS is replacing SVR4 packaging.

(Most medium/large) enterprise CIOs don't care about packaging choices. They care about TCO and SVR4 in Solaris 10 makes TCO for that platform a steaming pile of fail compared with other server platforms.

benr. said...

My position is well know... I like IPS, AI and Caiman, but I simply don't think they are ready for prime-time. SX:CE should remain available untill such time that the new technologies are solid and production ready.

AI needs more stability. Caiman needs a text-based installer. IPS needs a defined method for system config. Sparse zones need to work. etc.

The new technologies just aren't ready yet. Plain and simple.

Garrett D'Amore said...

Binary Crusader:

Please please please stop the myth that the problem with the SVR4 stuff is related somehow to patching.

The way we used to do patches, with sparse packages was undoubtedly a mess. But if we just agreed not to deliver sparse package patches anymore, nearly all of the major problems with SVR4 packages simply fall by the wayside.

The notion that we had to throw out decades of experience and technology because our processes around patching were messed up is nothing more than a slanderous lie IMO, particularly given the fact that IPS makes no bones about the fact that it doesn't support anything even remotely like old style patching.

The transition hurdle for administrators and end-users who have paid millions of dollars in the past, going from Solaris to OpenSolaris is very high.

I admit OpenSolaris' LiveCD install is simpler for the situations that it can be used. But most of those situations are folks downloading it for personal use. While this is wonderful for increasing adoption from Linux, it doesn't address the main concern I have, which is making this stuff useful to people who will pay for it -- this is how I make my living after all.

For large scale deployment by enterprises, the current OpenSolaris install technologies fall far short of the mark.

The Live CD is an interesting toy to play with, but it is currently not Enterprise grade, I'm sorry to say.

Binary Crusader said...

SVR4 stuff is related to patching; it's an enabler of the dim-sum patching model that Solaris currently uses and must die.

SVR4 is inadequate when compared to almost any other modern installation or packaging system on such a level that I dare not call it more much more than a glorified tarball manager.

As for Enterprise users, Sun marketing and others have made it very clear that the current releases are not enterprise targeted so it should be no surprise that they are not enterprise ready.

Change has to come some time, and change is always painful. Deal with it.

milek said...

Ben - while I agree with your other comments I don't understand why "IPS needs a defined method for system config"? Old packaging system doesn't provide it and what IPS offers currently is already better. Granted when you really need to run a script during package install currently it is a little bit cumbersome but if you need to restart other SMF service or add binding to /etc/driver_aliases, etc. then IPS offers much better way of doing it than old packaging when you had to use scripts which is prone to errors.

Re other technologies - yeah, they need more development, true. But this is exactly why re-directing some resources spent on SXCE to Open Solaris is important, especially currently when resources are even more limited than ever...

milek said...

Garret - SVR4 is old and clunky, missing lots of features people are used in other OSes for years. Implementing these features and fixing many other issues wouldn't probably save any time it would perhaps even take more time than starting from scratch. And the end-result would be a frankenstein piece of software barely re-assembling original code. So what's the point?

When it comes to installation of Open Solaris in enterprise (or rather lack of it...) - well, we're definitely not there yet.

nadkarni said...

Over the past month a bunch of us have been visiting a large number of enterprise customers on two continents, providing the direction and pointedly asking if this is an issue. What surprised me most was the level of positive feedback. The #1 issue for these customers is patching. And having spent some amount in the old patch code, I find your response "sparse package patch" the cause very flippant. For one you miss a key issue, which is the screwed up dependency mechanism. But that's just one of the many issues.

Rand Huntzinger said...

I hope to get a look at Solaris 10 zones soon (it should be in the OpenSolaris distribution in a week or so) but I don't think this addresses all of the transitional issues a shop will face. Basically, it allows you to maintain older applications which don't easily transition into the new environment. Nevertheless, you still need to be able to support the new environment underneath the zone. I suspect that many shops have operational procedures for installation and maintenance procedures which depend upon the old environment which will have to be reworked before OpenSolaris can be used in production scale. This will be disruptive.

I'm sympathetic to the arguments against SVR4 packaging. I think IPS is better and agree that it is the future. My concerns are with the present. I've tried the automated installer (AI) and it works but there are still things I can't figure out how to do with it. I don't doubt that it will get there eventually but it isn't there now. I also still have some concerns of the stability of pkg. It has improved over time but I'm still not convinced it is where I'd like to see it for production work.

I understand the argument that stopping SXCE will allow work on OpenSolaris to proceed more rapidly due to limited manpower. I don't doubt that this is true. Your customers also have manpower issues which must be considered. A transition to a new system will require extra work on their part which many of them are likely to dread.

I don't know where the best balance point is among all of these concerns. Change is usually difficult, even when it is worth it.

milek said...

Rand - agree with you 100% Please keep in mind that we are not talking about Solaris 10 here - there never will be IPS, AI, etc. in Solaris 10. So enterprise customers who want to stick to what they know they can stick to Solaris 10. We are talking about Open Solaris distro which is more of a work in progress / development snapshot of what's coming. The more resources can be focues on it and we can finally drop putting too much resource into sustaining legacy technologies in the development branch the quicker we will get there...

That Sun managed to get some customers to run their production environments on SXCE and Open Solaris is the ultimate testament of the quality their deliver (yes, I've been one of such customers for many years now).