Or, at least, your Violin arrays.
A few weeks ago Violin Memory introduced their Symphony management app. They asked me to put together a brief intro video.
As work progressed a larger question presented itselt: how can enterprise IT compete with low-cost competitors like Amazon Web Services? That will be the topic of the next StorageMojo post.
Symphony points the way to a feasible answer. Enterprise IT can – and must – provide services that cloud-based providers can’t. When Amazon burps, who do you call?
The StorageMojo take
I’ll be writing more about how IT can compete with Amazon. But let’s face it: IT organizations simply aren’t used to competition that is faster, cheaper and easier than they are.
It will be a tough change for IT pros, but a necessary one.
Courteous comments welcome, of course. Stay tuned for more on competing with the cloud.
faster, cheaper, easier? I assume you have used amazon right? because I would attribute none of those words to their service.
It is more expensive (by far) – it takes infrastructure and sets it back 15 years and slaps a IT-unfriendly API on front(in a nut shell)
It is generally much slower (esp for storage, since you write a storage oriented blog)
it is SIGNIFICANTLY more complicated, the whole “built to fail” model, the lack of capabilities and features etc. People think you don’t need as much experience running in amazon because you don’t have to worry about hardware. In fact the opposite is true, you need MORE experience to deal with the poor infrastructure that they have. Another fact is most organizations do not built their application stacks to the built to fail model. Really only large hyperscale companies do this (it makes sense at their scale). The time & money is not worth it at smaller scale. Look no further than the large swaths of web properties that go offline whenever amazon has a hiccup.
I could go on for hours, literally.
Amazon has really one use case:
1) highly variable workloads (constantly spinning up and down stuff) along with an application that can operate within the “built to fail” model. Anything else is a flawed usage of their system.
Doesn’t mean it won’t stop people from trying. I can run a Hadoop cluster off a bunch of 3PAR V800 storage arrays. Not what they were designed for(and would be very $$), but they’ll do a damn better job at Hadoop than Amazon does at anything else.
Yes you did hit a nerve 🙂 I do have Amazon PTSD after having used it for a few years and experienced the pain first hand. It was the single worst experience of my professional career. I’ve spoken with many companies who have similar experiences to mine, many have moved out, others continue to leave. People don’t realize how terrible it is until they get in. Most management types haven’t a damn clue what is going on, and that makes the job for folks like me all the more difficult.
I’m more than happy to get into more technical details at some point if you wish, not something I want to cover again at 10pm. I’ve yet to see a cloud:in house ROI model that had a pay off of more than 12 months (hosting in house vs in Amazon or any other cloud for that matter).
Most of the folks that use amazon and like it for the most part haven’t a clue the sorts of innovations we’ve had in the IT space over the past decade. Other’s are adamant that they don’t want to deal with any hardware(no matter the cost – one of my friends is at a big company where they spend high six figures/month on cloud (not amazon) because the boss doesn’t want to have to deal with hardware. ROI for doing it in house in that situation was 4 months by his calculations).
Others still are pressured into using “the” cloud because they don’t know any better and it’s all the rage right now.
So please, whatever you post – I hope it’s accurate.
in the mean time if you’re bored you can check out a recent post of mine where, in a nutshell I talk about how Software Defined Networking is a crock of shit (and I have tons of ammo to back that up in the post).
After literally years of being confused as to what exactly SDN encompassed I had the opportunity to ask the inventor of the thing himself a few weeks ago following a keynote speech from him, and the result is now that I’ve confirmed SDN is what I thought it was (no more doubt) I can rip into it.