Are You Ready for Software-Defined Everything?

The hottest buzzwords in IT technology today are software-defined networking (SDN), software-defined storage (SDS), and software-defined data center (SDDC). These are part of a broader trend that we might as well call software-defined everything. The movement towards a software-defined infrastructure is about decoupling the bare metal that executes the point data transactions from the software layer that orchestrates them. The hope is that by separating the smarts from the brawn, the underlying hardware can become cheaper and interchangeable (avoiding vendor lock-in) while the overarching software becomes more capable and faster-evolving.

The technological nuances of a trend this broad can be mind numbing; even those people in your organization most connected to the advances behind software-defined everything will tell you that there is too much going on to really track all of it. But despite all the moving parts – the veritable Alphabet Soup of acronyms – the biggest impediments to unlocking the promise of software-defined everything are not tied to engineering at all.

Ultimately, software-defined everything is about using software to bridge the technological and organizational gaps between discrete IT silos to create an underlying infrastructure that can be managed holistically as part of the business. Rather than individual elements (compute, storage, and networking), infrastructure will be treated as a set of resources required for specific workloads. In this world, the application, the end user, and hopefully the business are king.

But these changes are more than technological.

If teams that have been traditionally housed in silos must act more in concert, it means things like purchasing decisions, budgets, reporting structure, and even performance metrics must be re-thought. Organizations must be prepared – not just from an equipment perspective – to implement a software-defined infrastructure. With this in mind, here are a few things to think about:

Purchasing decisions. Who has a seat at the table when new equipment is purchased? Today, if your company is buying new storage, it is likely that the storage team makes the decision. Similarly, compute folks make the server decisions. But if these must work together, it is likely that people from different disciplines must get involved. The temptation here will be to include people in meeting invites to vendor presentations, but mere surface involvement will not deliver meaningfully better results. Teams must understand that their job is to evaluate deeply how purchases in areas adjacent to theirs might impact how they can work together to deliver optimal user experience.

Budgets. Once you get people involved in purchasing decisions, the question of budgets needs to be addressed. It creates organizational tension when teams who do not have budget control have an input into purchasing decisions. If the networking team owns the budget, what happens if input from the compute team suggests that the more expensive solution should be purchased?

Reporting structure. Similar to the budgeting impacts, what are the reporting implications of cross-team collaboration when making decisions about things like priorities? The term matrix is bandied about with reckless abandon, but are your teams truly operating as a matrixed organization? Are priorities universal? And communicated broadly?

Performance metrics. In a siloed world, performance metrics tend to be tied more to availability than end user experience. For example, networking teams might be measured on total downtime in a given period. However, downtime is a poor measure of productivity and experience. In a software-defined world, performance metrics should be reconsidered and applied broadly across teams.

The IT arena is certainly moving towards a software-defined future. But for any of us who have been involved or even watched from the sidelines as large transformation initiatives start and fail, we will know that the technological changes are perhaps the least interesting. It can take years to transform an organization – if you can do it at all. Even if the realities of software-defined everything are still years away, this means that organizations are likely already behind. Getting the people and processes in place to help drive this kind of change will be paramount for any company looking to take advantage of these next-generation technologies.

The clock is ticking. Forget the technology — is your organization ready?

Mat Mathews is co-founder and VP of product management at Plexxi.

Go Back to Top. Skip To: Start of Article.