THE CASE FOR A MULTI-VENDOR STORAGE STRATEGY
By Pete McCallum
There was a time, not so long ago, that a company could afford to be an “<Insert Vendor Name Here> company.” There was a go-to supplier, go-to VAR, and a go-to product line. The go-to company provided training and lunches, and that built up loyalty. In all of that loyalty, the hope is that the go-to company innovates regularly enough to continue providing value, and maintaining goodwill even after the sale. However, what I believe the consumer space is discovering is that trusted vendors didn’t live up to their end of the bargain. Perhaps it’s not the vendors themselves , but industry practices that need to change?
When I was running datacenters, and being entertained by vendors who wanted to sell me things, the big differentiator was always the extra-mile folks. Usually, technical resources who improved my craft a bit – and didn’t just rest with teaching me about their product. I worked for an aerospace contractor who had to follow government rules about three-bids, single-source, and diverse vendor strategies – mostly for security reasons. I was able to buy technology based on my business needs and trade technology coaching for purchasing tips with my preferred vendors –This practice helped me to vet out vendors very quickly into “providers” and “partners.”
Now that I’m out of that role, I see things a little differently. I know now, why long-term vendor affinity can crush innovation. Without explicitly naming names, take one of the best-known technology providers of the past decade. Most of their revenue has traditionally been tied to years 4-5(+) of maintenance on a CAPEX basis. Typically, your initial investment is discounted, upgrades and expansions are at a premium (to make up for the discount on initial purchase), and maintenance runs at least 20% of initial purchase price annually for as long as you want support. If you do the math, you actually re-purchase your product every 5 years. This should not be an epiphany to anyone as it has been a standard practice in the technology sector for decades.
Financials aside, this kind of practice causes some trickle-down effects on product innovation. If I stand to make 60-70% of my revenue in years 4+, why would I destabilize that base by asking a customer to upgrade or switch product lines sooner rather than later? My goal is to keep my customer happy on their current portfolio as long as possible. So how do I do this if I have to keep my hardware somewhat “flat” in capabilities? Software, of course. I can innovate all day long in my software as long as I only allow my software to speak with my hardware. Software also requires training and has both soft and hard loyalty due to skill, process, and product affinity. The effect of this practice is that you keep my hardware for years past its prime and lock deeper into my revenue stream. So now, instead of using technology to support my business, I change my business to support my technology.
I believe that today, this paradigm is being rejected more and more, causing an interesting reaction in the client and vendor landscape. The CIO cries out from the wilds and shouts: “I have too many silos, too many processes, too many people, too many moving parts!! I need simplicity!!” And the industry reacts with glee: Let’s give them convergence. So now the big companies band together to hold onto market share and bundle up end-to-end solutions with “cross-platform” orchestration. Datacenter in a box, we were promised. That’s all fine and dandy as long as you like and only use their boxes. Now we are even MORE tied into the revenue stream – and that one throat to choke requirement ends up being yours, Mr. CEO as you allow your business to slowly get muscled by your technology providers.
We followed this trend with Hyperconvergence which took out the multiple vendors, and built a consolidated, optimized reference architecture of best-in-class capabilities and built little tiny boxes of lock-in. As long as we like the hypervisor, and don’t mind the restrictions of scale, and the lack of mobility, and have the ability to toss our whole old datacenter (admittedly not a bad choice for many…), then Hyperconvergence is a great concept. But that’s a lot of very important “as long as you...” caveats.
Now we see the growth of open-source in the crazy juggernaut of OpenStack. Let’s get all the big players in the industry to contribute their best-in-breed feature differentiators to a “free” platform that gives everyone of any size an amazing platform of features and capabilities that will completely remove the need for your commercial software, and will tend to commoditize the value proposition of your hardware further into the ground. “[throat clears] sure. I’ll give you my best! [rolls eyes].” Technology socialism, for sure, but it does have lots of promise.
While I’m partially bashing on everything else, I have to make sure I include the cloud in my rant, right? Let’s get this definition crystal clear: the cloud is somewhere else, with some other technology, who I pay to maintain my data for the lowest possible cost and lots of contractual assurances. Heh. Yes, some clouds are puffier than others, but at some point, financials and competition will require cloud providers to push prices to the floor – and force lower standards. Am I saying you tend to get what you pay for? Yes. Yes, I am. The great thing, though, about the cloud providers is that they can be a greenfield technology transition without that investment on the part of the customer and it is about as simple as it gets to consume.
So the echo in the wind back to the CIO is this: “Now that you’ve asked, we’ve provided very little of value other than to offer you a place to offload the responsibility to someone else!!” And now the cloud provider has to deal with the vendor-lock-in and silos and cost fighting, and the CIO fidgets at her desk with almost zero control over what happens to her business’s critical data.
What is it that we need, then? There is something of value going on here, even if you can’t see it under my somewhat cynical overview: there is something happening . Movement is taking place. New technologies are creeping in that change the paradigm of consumption and bring cloud-like models into on-premise IT. What convergence taught us is that orchestration across storage, networking, and compute are critical for IT operations. What hyper-convergence is teaching us is that focusing down into workload-optimized topologies and optimizing operations through analytics-infused automation is critical to IT operations. OpenStack is answering the call for hardware-agnostic, agile platforms that leverage software to provide the right personality of services for a given workload at a given stage of a workload’s requirements. The cloud is teaching us that perceived simplicity and optimizing efficiency for cost is possible and viable. And we’ve proven once and for all that the old way of vendor “techstortion” (technical extortion, anyone?) is finally not welcome in business anymore.
What is the direction things are moving, then? Unless your business is simply hobbled by indecision and freezing all technical assets until something makes sense again we feel ya! You aren’t alone by any stretch; you have been looking at everything I have mentioned in this rant (er… blog!). You have small SAN, mid-SAN, server-san, all-flash array, good-old’-boy vendors, NKOTBs (New Kids on the Block), hypervisor-based, containerized, SDS, SDx, traditional, modern, memory-resident, SaaS, Cloud-enabled, gateway-ed, referenced, marketed, funded, and struggling vendors knocking on your door all day long asking you to listen. And you are listening, and making your choices. You are buying something, even if it’s only time. In the end, you will dabble. IT always has. We take bits of all the cookies on the plate and spit out the bad tasting tech. We keep a handful of old standards and some of the weirder cookies we didn’t think we’d ever like.
What I’m saying is that you will end up buying or trying some of almost everything I have mentioned. Your gut is telling you that something needs to change – and I think I’ve shown you what is broken. The industry will react to what you keep and what you spit out. What you will need is a way to protect your data assets while you test drive and kick tires. You need subscription-based purchase plans that do not obligate you to a return-on-investment time. If a technology does not provide value, you will simply stop using it and your money will go elsewhere. In order to facilitate this, you will need a platform that allows you to bridge your infrastructure across locations, form-factors, vendors, technologies, providers, company names, and other traditional barriers – all while maintaining your processes, skills, and guarantees of service levels to your business. Yep, you need a chaos platform – a middleware between the uncomfortable known and the exciting unknown.
Some of you may be forced into a multi-vendor strategy for compliance or procurement practices, but all of us will be goaded into it simply because the old model is broken and there is simply too much to choose from these days that tangibly helps your business.
I was this customer. I had all these challenges, I saw these changes coming. I embraced a variable technology architecture with multiple locations, and chose a platform to bring it all together and solve the challenges of shrinking budgets, compliance issues, reductions in force, changes in IT leadership, mergers and acquisitions, migrations of technology, consolidations of infrastructure, monolithic vendor platform overturn, and other challenges.
With all the chaos you will be finding in your technology options today, I would highly recommend kicking the tires a bit on a product that will help you to make better sense of what you have, and enable you to safely constrain the chaos of what is to come.