22nd May 2012

Blessay: Cloud Convergence, Skills and Your Career

With the rise of Cloudy Computing the boundaries between networking, servers and storage are changing. As the marketplace chooses new technologies, products and vendors our professional careers are going to change and evolve at the same time. In this article, I wanted to consider those changes and perhaps come to terms with the new environments from a career and personal perspective. In one way, I’m planning my career here by thinking out loud.

There is no constant but change

Perhaps the first thing to understand is that IT has always been changing, with several seismic shifts that changed the industry. The last thirty years have shifts from mainframe to minicomputers to distributed desktops to server centric computing. Perhaps, the last such change was in the late 1990ís when server centric designs (started by Novell and later Microsoft) became the fashion, and hasnít much changed in that time. Perhaps because the server-centric era has lasted for so long, we have forgotten that all things are temporary in technology.

The Fear of Change

To my mind, one of defining factors of the push into cloud computing is the fear of change. That is, cloud computing appears to offer significant savings but IT departments are wary taking risks to self-integrate this technology. Today, IT Departments select best of breed technologies and then self-integrate them into the existing infrastructure. Thus, customers might select one vendor for servers, one vendor for networking and another storage, this selection process means that products matched skills and requirements for the needs of that specific business.

But historically, IT departments has a two vendor policy for all equipment to reduce the risk of failure. It was common for companies to use two types of firewalls, two server vendors, two networking suppliers. The impact of a failure by a single vendor to provide a quality product or suitable maintenance meant that an alternative was readily available. Additionally, competitive pricing of two vendors meant that vendor pricing could be kept under control. By inviting pricing from competitors the risk of technology companies for overcharge is mitigated.

On the other hand, a dual vendor policy means that two skill sets for operation and administrations must be maintained. I would maintain that having experience of two vendors is a far more rational decision since you will be aware of alternate methods and approaches and more flexible in your planning. Really, it’s not that hard, it just seems like it.

The Fear of Risk

Managers (probably fresh from their MBA courses) have decided that a single vendor policy reduces training costs (but conveniently forgetting that this increases purchase costs and risk by using a single vendor) and by reducing the technical breadth, operational costs are also reduced.

This has serious impacts on IT teams today, as many people believe that their careers are linked to single vendor. But lets look closely at what those skills mean.

The Skills Overlap

Lets consider the engineering skillsets that we have had for the last ten years, in this handy chart:
cloud-skills-1.jpg

I think this graphic summarises the separation between the Infrastructure teams in the current market.

The Impact of Virtualisation

The arrival of VMware and their much improved systems management tools ( i.e vCenter ) rapidly displaced the existing virtualisation products and has gone on to create some disruption. I see this disruption represented in the following diagram:

cloud-skills-2.jpg

The net effect of Server Virtualisation has been to break the direct contact between the Server OS and the external infrastructure. In many ways, this has been reasonably straightforward since nothing much changed. Networking still presented ports and VLANs to servers, and Storage continued to present LUNs and Megabytes. Different perhaps, but not really a fundamental change.

Enter Convergence

Since convergence of Storage and Networking has started, we are now seeing the skills graph look more like this:
cloud-skills-4.jpg

Storage and Networking will integrate much more deeply, and VMware is moving to provide management tools that configure the Network and Storage layers. In my view, VMware is more about management tools and visibility than a virtualisation platform. These tools mean that costs can removed (by abolishing menial tasks) and visibility improved (thus reducing risk).

New Skills Maybe, but Just Skills

So lets break this down to skills. At least, lets pick the main skills and put that onto the graph. Here is a rough and ready reckoning:

cloud-skills-3.jpg

So what ? The thing to notice is that the skills didn’t go away. All the skills that we use today remain vital and important. And the skills of tomorrow are just new skills. They could be storage networking, or the new VMware tools for managing storage configuration and network configuration. The traditional skills around Server OS, Internet Firewalls, and Drive Arrays don’t change and remain relevant in the future.

A Whole New Skillset

The new skill is what I call Converged Engineering. The skills that blends networking, storage and virtualisation experience and knowledge into a single point. Because the infrastructures all overlap, most organisations are going to need at least one person who can cover all the technologies. I suspect that this is new “hot job” for the next ten years.

The Stack / vBlock

I think that this also demonstrates why the the vendors are offering a solution stack. They need to make, †train or create these Converged Engineers who can design and lead the early adopters into successful deployments. In a couple of years, there will be more people who have the new skills, and most likely the stacks will collapse into best of breed solutions that are integrated by the customer. In the short terms, the skills shortage means that vendors will need to cover the gaps with engineers, probably speciallly picked and heavily trained, who can cover the spectrum of skills in networking, storage and virtualisation.

Also, as the products develop and technologies progress the knowledge will change rapidly. For example, DCB is not yet finished and may yet change, and the final practice is yet to be seen.

The EtherealMind View

First, Networking isn’t going away. The skills you have today, will still be needed tomorrow. You ARE going to have to learn new skills, and change your working relationships. The integration with storage and virtualisation is going to intense and challenging. Second, most of the menial work will shift into automated tools on the virtual platforms. For example, VLAN creation, port allocation and IP addressing will be done by automated software in VMware.

Just because the last ten years hasn’t seen much change or innovation doesn’t mean that innovation has stopped. It’s time to knuckle up and reach for new knowledge and skills, I need to be thinking about where my career fits in the new model.

And that’s the same as always because change is the only constant.

This post is copyright of Thropos Ltd ©2008-2011 at Etherealmind.com - contact | email: greg.ferro@packetpushers.net - twitter: @etherealmind | All rights reserved
About Greg Ferro

Greg Ferro is a Network Engineer/Architect, mostly focussed on Data Centre, Security Infrastructure, and recently Virtualization. He has over 20 years in IT, in wide range of employers working as a freelance consultant including Finance, Service Providers and Online Companies. He is CCIE#6920 and has a few ideas about the world, but not enough to really count.

He is a host on the Packet Pushers Podcast, blogger at EtherealMind.com and on Twitter @etherealmind and Google Plus

  • Fernando

    This is a great post to get the discussion going, thanks. Mid and senior level technical professionals ignore this debate at their peril.

    A few rambling thoughts evoked by your article:
    - To me, one of the biggest challenges for the current architect/engineer will be the high frictional cost of change. If one adopts a vendor stack, especially in a “non-cloud” arrangement (i.e. deploying HW within an enterprise), the impact of moving from vendor A to vendor B is now exponentially greater, particularly if one designed things with “vendor enhancements” in mind that cannot be replaced by an open standard. We’ll be creating significant challenges for our organizations if we don’t help them understand that the decision to align with a stack must be made at the highest levels of IT and business management.
    - For the individual, this broadening of skillsets will mean a further dilution of attention between even more technologies. We’ve seen it before in other areas of IT, this is just a continuation of the process of becoming a mature infrastructure industry. I imagine we’re just coalescing around a three-tier professional environment, with different requirements/rewards:
    1. Architecture professionals: broad understanding but not necessarily deep expertise.
    2. Engineering professionals: deeper expertise in particular areas/problems.
    3. Technicians: implementation, task-focused resources.
    - From this we derive that medium or large organizations will not only require all three types of professionals – either in permanent or per-project basis – but that the ability to communicate effectively within a team will be even more valued.

    Perhaps the IT training outfits should start including copies of “Who moved my cheese?” with the course materials…

    Cheers,
    Fernando

    • http://etherealmind.com Greg Ferro

      I take the view that the “stacks” are temporary as the new technology arrives and achieves early stage adoption. My current speculation is that the management software that enables the cloud needs to start with a limited set of hardware to be successful. Then when stable it would expand to include other vendors in attempt to grow market share.

      Therefore, the current stack is most likely temporary until customers and market developments call for multi-vendor solutions.

      • Fernando

        I truly hope it is this way, but I’m not holding my breath. Let me count the ways…:
        - I view the “stack” as a lock-in play for the major vendors, not as a technology ramp-up environment. Cisco, HP, IBM, Oracle, … all want lower cost of sales and repeat customers. For an enterprise customer, I think the decision on which stack to go to will be the mother of all political battles – servers vs applications vs networking – played at the highest levels of IT/business management.
        - Once a stack is chosen, the sheer size of most customers will likely guarantee inertia. Think replacing Cisco with HP/Juniper/… on the networking layer is cumbersome? What about ripping out everything, including servers and storage?
        - Even if the management/orchestration component is able to expand, there will be a lot of internal pressure to build a better mousetrap with the stack vendor’s own components than venture out and make it easy for a competitor product to integrate.
        - Two words came to mind when you mentioned “expand to include other vendors”: Cisco MARS. It was supposed to be exactly what you described. We know where *that* ended up…

        My hope for best of breed components is for them to have strategic alliances with all major vendors – everyone can run VMware, for example (OK, except Oracle than got virtualization with Sun), or F5 can play with HP/Dell/NetApp/…

        This discussion is focused on enterprise customers running their own “clouds” (I still don’t like the term…). I don’t have enough exposure to service providers’ internal networks to see how they’ll deliver these services, but I would expect some more heterogeneous deployments there (maybe each SP will have a copy of each major stack and build the orchestration glue between them as intellectual property).

        The pseudo-Chinese proverb “May you live in interesting times” certainly applies here…

        Cheers,
        Fernando

  • Jake

    Great article Greg. Although I’m only a college senior who is about to get thrown out into the “real world”, this pretty much sums up what I’ve been observing from the sidelines as far as the increasing prevalence of virtualization and converged data center networks is concerned. My degree program does a good job at getting us lots of exposure to both the networking and server administration areas of the IT industry, but my focus has mainly been on network engineering. The decision I’m faced with at this point is whether to dive as deep as possible into networking to obtain what Fernando might call engineer and architect level knowledge, or if I should sacrifice some of that super in-depth knowledge and skills to gain a broader virtualization skillset.

    My gut feeling is to study as much virtualization as time reasonably allows while still staying focused on networking and obtaining the engineer or architect knowledge on an as-needed basis. Do you have any thoughts or comments on this philosophy?

    Also, I just wanted to say that I like your site quite a lot, and that when you pronounce router as “rooter” on your podcast, it’s pretty funny to me for some reason :-D

    - Jake

    • http://etherealmind.com Greg Ferro

      It worth noting that WAN and Security networking are mostly unchanged (although virtualisation does change parts of that as well). But your skills are useful, just needing to be applied in a new context.

  • Felix

    Great Artical :-)

    Jobs will change fast. It also depence where you work at. I started of at a / Telco on ATM, guess what DSL was added. Oh the Management System is the same as the SDH one… I moved on because I liked that Internet. I then started with leased line customers… Well guess what… End to End Configuration to the pop and the routing….
    Then I changed jobs to a SMB enterprise, Doing Router and Switch Managemenent ….. Cause I had linux/unix skills add server administration. Because of tight budget, but well we need this service…. There is a software produkt which adds an hardware abstraction layer, so that you can add another operating system to that linux server… lets try that.. Guess what, this works we need more power and wouldn’t it be nice to get a little more uptime and assurance in.. if this server has a failure this and this server will also fail…
    oh we need a shared storage for that solution… well gues what…. there are company’s who provide that solution.

    Oh we need this server to not reach that server on the same machine… lets do some real interesting routing, switching and security..
    My current job title is called “IT System-Manager” mayby called TroubleShooting Expert, Network Man, (L)Unix Guy, Storage Man, Solution Architect, Security Man, Telefon Man or Cable Guy, ( not to mention all that specific ITIL Rolles)

    Yes and that funny thing is… it is stressfull, but also awfully interessting :-) Having that Skills and knowledge… It is still good to have certification???? I’m currently holding my CCNP, will try my CCIP in a few month, I want my VCP. Next Thing maybe Storage?

    IT is changing…good.. and we all are already a part of it.. Think of MPLS and VRF of Network Virtualisation. Firewalls are named context.
    Network Engineer have started to “virtualize” there Brains already :-) Converged System Engineering Sounds a little better

  • Pingback: New skills needed | Timothy Riendeau