Solaris

The battle cry of IT brinksmanship, “If it ain’t broke, don’t fix it,” means that risk has triumphed over cost. This fate is commonplace for legacy applications which can be found sitting on an outdated/unsupported box, running on ancient OS sporting a “Do Not Touch” sign.

Not important enough to fix; too important to fail. These applications are at risk of failure.  And everyone knows it. 

AppZero offers an alternative that changes the risk/cost math by eliminating the risk at a slashed cost/effort – with no re-engineering or coding required.

Over the last few years, we have helped a number of significant IT operations use our application virtualization solution to migrate their legacy Solaris applications onto newer systems that are reliable and powerful systems. Prior to learning about AppZero, these organizations lived with risk hunting spare parts for their hardware systems from Ebay and Craigslist – sites that, like the old buffalo grounds, are now hunted-out. At this point, risk becomes probability.

Most people reading this article understand that the evolution of Solaris brought big changes between version 7 and 8. So why are so many applications still running on 2.6? 

Because 2.6 was a very popular Solaris version, many customers invested in applications – both mission-critical/enterprise and line of business/departmental. As the Solaris operating environment grew, most mission-critical applications were moved to the more current versions of Solaris through application vendor upgrades, patches, etc. However, there were many organizations that had built custom applications, or had applications for which the vendor was no longer in business, and they had no easy way to migrate to the newer Solaris versions. Or the applications were not deemed ‘mission critical’. Enter the “wait and see” or “if it ain’t broken” approaches.

Fast forward to today. We at AppZero know firsthand that there remain a sizeable number of legacy Solaris applications which continue to be supported by some of the largest and most recognizable companies. In our interactions with these organizations, we have observed an odd mix of magical thinking and fatalism.  It has always worked; we will deal with it when it happens.

What is the truth? Sun/Oracle made significant efforts and investments to provide backward compatibility for their customers. They are one of the few vendors who have made significant investments in order to try and make it easier for their legacy clients to easily migrate to more current Solaris environments. The challenge for organizations running Solaris environments of the 2.6 era is that their applications usually don’t meet the criteria to be covered within the Oracle Binary Compatibility Guarantee. 

They had legacy clients in mind when they originally developed Zones within Solaris 10. For any organization running Solaris 8 or 9 applications, it’s a simple matter to put them into branded zones on Solaris 10 to enjoy the benefits of the newer OS and hardware systems. What many people fail to understand is that the branded zone approach is not an option for the older Solaris 2.6/7 applications.

Here’s where AppZero comes in. Our software encapsulates your old 2.6/7 application, packaging enough of its OS eccentricities that it will ‘feel’ completely at home when it is picked up and placed on the bright and shiny Solaris 10 OS and box. Instantly running.  Risk and cost dramatically cut.

If it ain’t broke …. check out AppZero’s solution and ask about our “no app left behind” service.

I am always looking for a way to communicate better and cut to the heart of the discussion. So, if you have thoughts on this subject, drop me a line at rwhitcroft@appzero.com, or tweet us at @appzero_inc. 

“So, VMware … separate hardware from OS = virtual server, right? Well, we’re like that but completely different – we separate application from OS = virtualized application.” Because we’re in the vast virtualization market space with a technology that is unique (patented), I am often asked to differentiate AppZero’s server-side application virtualization from the rest of the pack. Especially in the Oracle/Sun/Solaris/SPARC context.

And especially since we’ve been promoting a Webinar series drawing a straight line from ancient Solaris 2.6 and 7 applications to execution on OS and hardware from this millennium. (We broke the marketing bank naming the webinars: “Virtualize-2-Modernize; Solaris 2.6 + 7 applications run unchanged on 9 + 10” and the follow-on “How to run Solaris 2.6 and 2.7 applications unchanged on Solaris 9 and 10”)

Good news/frustrating news

So I’ve recently been immersed in SPARC-dom through the “Virtualize-2-Modernize” series, in tandem with doing exactly that type of Solaris application modernization work for some household name clients. The good news is that our customers clearly see AppZero’s value proposition and ROI as very high. Understanding and adoption go hand-in-hand – as do confusion and delay. What is so frustrating to me is that this clear value proposition is too often occluded by confusion surrounding the various flavors of virtualization in this market.  Specifically:

• Oracle VM Server for SPARC, the technology formerly known as Logical Domains (LDOM) – This server hardware virtualization and partitioning technology was released in 2007 by Sun and rebranded since Oracle’s acquisition of the company in 2010.
• Solaris Containers/Zones (global, non-global, branded) -- First made available in 2005 as part of Solaris 10, this operating system-level virtualization accommodates applications running on isolated virtual servers within a single OS instance.
• AppZero Solaris application virtualization – First shipped in 2006, AppZero separates an application from the OS, capturing that application and its dependencies in isolated capsules that can execute on a single Kernel OS instance.  For example, AppZero encapsulates Solaris 2.6 and 7 applications with enough OS distinctives that they can execute on the newer OS versions. (Copied and run, the applications are not installed in the traditional sense although they execute as if they were.)

It’s easy to delineate LDOM/hypervisors and application virtualization; it’s trickier to come up with a one-liner for containers/zones. Part of the difficulty stems from the fact that this Solaris technology and AppZero’s Solaris application virtualization solution both use analogous approaches to create isolated application executions on a single OS kernel instance. For those of you interested in differentiated elephants and nuances, I’ve included a comparative matrix at the end of this blog.

Differentiation by use case

In the meantime, AppZero use cases are clear. Tuned for rapid provisioning (what we call “ZeroInstall”), AppZero enables moving/copying a pre-installed, pre-configured application to any machine to be up and running in seconds. This resultant extreme agility has direct value for:

• ISVs who want rapid provisioning of demos and proof of concept (POC) as well as fast distribution of software.
• Provisioning of applications for scaling and de-scaling in a cloud.
• Moving applications from data center to the cloud … to another cloud … and back … etc.
• Reducing OS and image sprawl by dynamically applying “gold OS” and “gold application” images.

And in a category almost by itself ….

• Modernizing old Solaris 2.6 and 7 applications – separating them from the kernel – so they can move to Solaris 10, enjoying all the benefits of hardware and Solaris services from this millennium.

Differentiation by attributes

Inspired by a recent Peter Baer Galvin blog, “Pete’s all things Sun: comparing Solaris to Red Hat Enterprise and AIX” I offer this matrix comparing the technologies. I welcome your comments, refinements, and (should the unthinkable happen) corrections.

*Including apps that are deemed not compatible with Binary Compatibility Guarantee and are statically linked

Note: AppVirt: AppZero for Windows platform has a different interception level that Linux and Solaris 

I am always looking for a way to communicate better and cut to the heart of the discussion. If you have thoughts on this subject drop me a line at GregO {@} Appzero {dot} com or tweet me at http://twitter.com/gregoryjoconnor.

Imagine you are an IT professional or executive, and your teams are running a data center with 1,000s of machines and applications.  You know that there is a set of 10-15 year old applications running business critical functions on SUN hardware and the Solaris operating systems.

These applications, their OS and the underlying infrastructure are old.  In fact, measured in “IT-years,” where the lifespan of an infrastructure is 3-5 years, they are ancient.  By my calculations, an IT year is equivalent to 20 people-years, making these systems 200-300 years old …and ….

… stranded on an island.

It is not a question of if one of these servers will break or die; it is a take-it-to-the-bank-matter that these mission critical application servers will die.  The only question is when.

There is no replacement hardware available.  You can’t even find these servers on eBay anymore.  The poor old servers can not access new SAN based storage.  The cost of re-engineering the applications so that they can play in a supportable IT world is huge.

I have talked with many people in this situation and simply put, they do not sleep well at night.

Here’s their situation in a nutshell:

• Fact: Extremely valuable business application runs on infrastructure that is soon to fail.
• Fact: Huge pain and lost dollars to the business will result when these servers fail so there is a very tangible and measurable need for them to be fixed ASAP.
• Fact: I have talked with some IT folks who say that if their custom Solaris applications ever went down, it would take them weeks, months, or even years to recover.
• Fact:  These applications are integrated with other systems and do very specific tasks tuned for the business, making them poor candidates for prebuilt alternatives.

At this point four choices are available for these companies:

1.) Rewrite/port the applications

2.) Complain loudly so that when it hits you can say “I told you so”

3.) Run … find a job at a cloud company where systems are only a couple of years old

4) Take AppZero’s “virtual Viagra” to scoop and move your apps from the old OS to a new OS with zero re-write

The AppZero capsule, taken once, will last more than 4 hours without requiring any re-engineering or any trip to the doctor.

AppZero = engineering-free Solaris application modernization - 

How do we move old Solaris applications to new environments without re-engineering?  We separate the application from the OS and put it in a very portable capsule.

Just to be clear, separating an application from the OS is not something you can easily explain to your mother.  I know.  I’ve tried.  Why these old applications are stuck on old hardware and how AppZero can separate and move them to new hardware in a matter of hours excites me, but Mom?  When I hear, “That’s nice, Greg,” I know she has no idea what I am talking about. So instead we talk about how Jameson is doing in football this year.  (btw, he’s doing real well, thanks).

AppZero builds software that enables IT and software developers to create, control, and maintain virtual application appliances (VAA).  The VAA decouples an application from the operating system (OS).  It encapsulates that application, along with all of its required underlying infrastructure, so that it can then easily be transferred and run on a different system.

I sometimes compare the VAA capsule to the gelatin that surrounds everyday medicine you’d find in the average medicine cabinet.  The gelatin isn’t the medication, it surrounds it.  In a similar manner, the VAA is not the application, but contains or encapsulates it.  This capsule can then be picked up and moved to another system with very little effort or interruption.

VAA vs the cost to rewrite or port an application

An AppZero customer who modernized a few dozen applications with our technology recently spent a good 45 minutes telling me about the bidding process with IBM, Accenture and Wipro to “refactor” the systems that he later moved via our application virtualization.  The baseline number was $7M and 4+ years.

Our “capsule” solution was less then 4% of that number and implemented in 9 weeks.  Maybe I need to talk to my new board members about raising the price of this tool we have for Oracle and its newest crowd of faithful users?

Many of our prospects have been talking to us about how they can remain faithful to Solaris while still getting the thrill of running their applications on a “newer model”.  They are confident that Mark will inject some new life into the Solaris platform. 

In the meantime, long-time, faithful Solaris users can take advantage of “capsules” that allow them to share in this exciting new life today without having to divorce themselves from the applications that they have come to love and rely upon.

To quote Viagra’s tagline, “This is the age of taking action.  And getting the answers you want.  So you can get on your way.”  AppZero couldn’t agree more.