The days of legacy UNIX are all almost over as companies move towards the cloud. Enterprise IT infrastructure built on the need for interoperability resulted in vendor lock-in; cloud technology promises a solution. Freedom of choice, even on a daily basis, should be the cornerstone of any planned migration towards a cloud environment. In addition to choice comes the requirement for standards, both in terms of technology (APIs, communication etc) and in process – the ways systems are built, deployed and managed.
Linux and open source software has changed the way we think about IT infrastructure completely and it has been very successful in changing the way operational IT is run. The move to open source technologies, and the process that built them, provides a model for cloud adoption. Back in 1996, before there was an enterprise- ready distro of Linux, IT managers turned to Linux on SunOS and Solaris, as it was easier to manage and deploy. However, adopting Linux in the early days was met with some resistance, particularly from the risk averse who believed that the adage ‘no one ever got fired for buying IBM’ extended to other premium-priced IT products.
Persuading certain people that Linux and open source software was best for web services took some time. The premium IT solutions brought only marginal gain while the people who kept the kit running were in high demand. The 1990s were a boom time for IT workers and Solaris admins were expensive; Linux experts less so. Market forces in addition to packaged scalable solutions helped push open source into the mainstream.
Many early adopters of Linux in the enterprise are the same people who moved to open source and commodity hardware data centres between 2002 and 2007. Those businesses have reaped significant savings and performance benefits. These organisations could be seen as enterprise open source pioneers that designed and migrated to a new environment, and understood how to take what had been primarily community driven till then into the very commerical world.
They paid a higher price for adoption, both in terms of acquiring skills and leading the way on process and policies, as well as the raw technology. But the benefits made it worthwhile.
One of the more implicit benefits of open source beyond more affordable hardware, and lower licence/subscription costs, was standardisation. As estates of enterprise Linux got bigger, the need for effective means of management became clear. Open source systems management tools like Red Hat Network, Puppet and Cobbler now predominate in the enterprise, and a Linux environment provides significant management advantages over legacy UNIX or Windows environments. This forms what we now know as a Standard Operating Environment (SOE) and is a fundamental part of any current data-centre strategy. An SOE helps in managing and containing the complexities that inevitably arise when configuration variance has spread across the IT and one-off configurations and silo systems have multiplied. As numbers of platforms begin to grow, it is worth the time and effort to review existing systems and plan a migration to an SOE. Indeed, without the standardisation of platforms and process, it’s impossible to add further functions that make up a fully operational cloud deployment.
With a multifaceted migration, involving OS, virtual environment and applications, organisations need to retain authority of their new SOE and they can ensure this by designing it before moving existing infrastructure across. As a concept, an SOE is a collection of loosely coupled but functional tools and processes that manages an IT infrastructure. With the appropriate SOE in place, a business can be cloud-ready securely and in the knowledge that decisions made now will not prevent future technology adoption.
Our view is that there is simply no point investing time and energy in transferring imperfect solutions and processes into the cloud. Red Hat Consulting designs and develops SOEs, making deployment, updates and configuration management as efficient as possible. Red Hat measures that maturity with 10-15 enterprise customers per month, partly to assess their readiness for cloud computing.
Linux is rapidly becoming part of the establishment; even traditionally conservative organisations in verticals such as consumer banking, insurance or retail are adopting it. Some exorbitant renewals for older hardware and software solutions are pushing organisations away from UNIX and so the software suppliers selling the latter are seeking to maintain their profitability by raising the price for the firms still locked in. The return-on-investment to migrate to Linux makes a lot of sense. Migration provides organisations with the opportunity to update their system management, and to consider that in terms of a future highly agile data centre.
Adapting and adopting new technologies ahead of rivals can provide a fantastic competitive advantage. However, it is not without risk and those firms migrating to emerging technologies later often benefit from a number of advantages. There are tried and tested best practice routes for adoption. Furthermore, a greater number of technically skilled open source businesses and people exist. All of this helps improve service levels and push down prices.
In many respects, the new technology has become commoditized and the standard offerings for open hybrid cloud computing are now here; OpenStack and OpenShift are just two technologies that ensure the process of adopting standards, rather than more expensive custom- built solutions. A cloud solution off-the-shelf and running is going to be a lot more difficult than open source component-based projects like OpenStack: data gets locked into proprietary databases and as a result, firms get locked in to a single vendor. The vendor owns the roadmap, not the business, and the vendor controls the fees.
The parallels between enterprise Linux adoption of 15 years ago and virtualisation and cloud computing today are clear. Businesses should adopt what will become the standard and what is heading towards commodity. Choice of supplier (if you want one), standards adoption and community-led strategy are all key indicators. By working with a trusted vendor that can support and provide consulting on OpenStack, the required open source components can be brought together.
A final consideration should also be made regarding the importance of retaining architectural ownership of the project; piloting your own ship, even if the crew is comprised of hired hands, makes a lot of sense. Surrendering ownership can sometimes seem like the easier option, but it will take you in a direction where one small decision now may preclude a successful cloud strategy in the future. One thing should now be clear to most large enterprises: legacy UNIX has no future and it should have no place in any vendor’s cloud strategy.