Daily Archives: May 19, 2014

10 Mistakes You Make When You Try To Build a Private Cloud

VentureBeat recently ran a great article on 5 Mistakes You’re Making When You Try to Build a Private Cloud that did a great job of covering 5 mistakes you make, but why stop there? SI can easily come up with 10 mistakes, more if it gives the issue a second thought. So, since some of you still don’t believe that The Cloud is Filled with Hail, let’s review the VentureBeat 5 and throw 5 more into the mix to see if that’s enough to convince you that The Cloud is Not a Magic Mirror — especially when you take a do-it-yourself approach!

VentureBeat’s 5 Mistakes of a Private Cloud are:

1. You believe the cloud will solve all your problems.

With so many vendors touting it, you believe that a cloud must be the answer, so why not control your own? There are a host of reasons, including those that will be discussed in response to the other wrong assumptions, but the most important thing to remember is that not all applications are good candidates for the cloud. Applications that are intermittent, that run full tilt, or that spike unexpectedly are not always good cloud candidates — public or private.

2. You think everyone will automatically love the idea.

You keep hearing that clouds bring agility, adaptability, and actionable data — so you think that you can convince everyone else to fall in love with the cloud too because you believe that these are reasons to fall in love with the cloud. A cloud is as adaptable as the software that drives it, as actionable as the data you can get into it, and as agile as your organization — if it takes 3 months to get a product to market using the best processes you can come up with, it takes 3 months to get that product to market — cloud or no cloud.

3. You think it’s cool.

Clouds aren’t cool (although the rain they bring may cool you off). And unless you are in the business of selling “cool” technology (i.e. private clouds to suckers who buy private clouds), the last thing you should be basing a business decision on is the “cool” factor. You buy technology to solve your problems, not because it’s cool.

4. You think you will succeed in boiling the ocean.

A private cloud is a huge IT project similar to trying to replace 3 ERPs across your global organization that have been entrenched for 10 years across 3 continents in one fell swoop while trying to add 4 modules you never had before. It’s like trying to boil the ocean with a single giant magnifying glass — brave, maybe even visionary, but ultimately stupid.

5. You think your plan will fit the organization.

The typical private cloud relies on converged infrastructure (CI) stacks which break down the typical organization walls of application teams, server teams, network teams, and storage teams. How many Global 3000 organizations have one single version of the truth across the enterprise? Maybe the few dozen organizations that successfully achieved enterprise wide deployments of SAP and Oracle?

That’s just the beginning. Here are 5 more mistakes courtesy of SI:

6. You think a private cloud will be cheaper than a public cloud.

You might think that a cloud is a fluffy magic box that can be obtained with a handful of magic beans that you can get by trading a simple cow, but that’s about as far from reality as you can get. Clouds require hardware, software, dedicated network connectivity, and power. Lots of power. You will need backup generators in addition to a wall of UPS units (to keep the machines humming until the generators kick in), multiple fibre connections, racks of machines and storage area networks, and a lot of specialized software. And, instead of sharing the cost, you get to pay for it all — as well as the staff to build it and maintain it 100% — 24/7/365.

7. You think all modern technology was built for the cloud.

A lot of software is, but not all — and chances are that a lot of the software you are using, even if still under maintenance, was not built for the cloud. So, you’ll have to update your current software and migrate your current data stores while you are at it.

8. You think it is the best way to interact with your trading partners and the private clouds you wrongly assume they have.

The cloud is connective, but only if it is shared. Otherwise, it’s just one massive local area network that needs to talk with other massive local area networks used by your trading partners. Clouds don’t create connectivity – data interchange standards do, and you don’t need clouds for that!

9. You think you can secure it better than the experts.

Hi Ho, Hi Ho.
It’s off to work we go!
We block the ports and tune the firewall
In our ‘Net the whole day through
We block the ports and tune the firewall
It’s what we like to do
It ain’t no trick
To lock down quick
If ya block the port
With a sniffer on a ‘NIC
In the ‘Net …

And you can block every port, patch every firewall, and sniff every ‘NIC, but the reality is, your network is only as secure as the weakest link — which is probably the software you’re using and the ports you need to have open. Which you don’t know how to protect because your IT staff is struggling to patch your firewall, scan the ports, and upgrade SSL before the heartbleed bug bleeds you dry of your corporate secrets. When it comes to security, you need true security experts — and you’re not going to have them in house.

10. You think the cloud can actually be secured.

The only way to truly secure a network is to unplug it. So if you think you have a hope in Hades of securing your private cloud …