I'm on a cloud kick this month. Not long ago, I wrote about how my company had gone "all-in" with the cloud. To be specific, we mostly went SaaS. This was my answer to the "cobbler's dilemma" (the cobbler's kids have no shoes); while we ran around bringing other people to the cloud or big data, we had neglected to take our own advice. This resulted in an emergency cloudification.
When you think about it, most of us are at the point now where are there a bunch of things we should just cloud thoughtlessly before something happens. Right now, before we regret not doing it:
1. Backups: Unless you have two data centers or DR centers that are far enough apart that nothing could possibly happen to both simultaneously, just backing up to "over there" in the traditional way might be insufficient. What happens if you have one of those 100 year storms take out the northeast or a strong east coast earthquake? Are your backups replicated across geographic areas and safe from the Sharknado or the upcoming Sharkuricaine? Cloud backups are.
2. Faxes: It is still necessary to communicate with the dummian world. Dummians continue to think the fax machine is not only relevant technology, but that it is somehow safe and secure or that businesses that don't have them are garage companies. Yet there's no reason to stoop so low as to have an actual fax machine -- nor the inevitable "Office Space" moment with a fax machine. There were cloudy fax services before cloud was cool or even a marketing term. You get a fax number, and when someone faxes you something, it ends up in your email. I've used e-fax for years.
3. Training videos: Every other Friday my company does a "training Friday" where everyone in the Durham office squeezes into a room or two, we telecast to our Chicago location, and we cover a technical topic. These are great epics because food is provided (usually Thai food because it is the only thing we can agree on besides "not pizza again"). We record these for future victims ... er, I mean hires. I'd been sharing them on DropBox as big fat files. Apparently this was seen as not very efficient. They're all on YouTube in our own private channel. You can't see them, so there.
4. CRM: If you have more than one salesperson, you're probably too big to self-host your CRM. Salesforce has been around for more than a decade, and there are less expensive alternatives like SugarCRM. The SugarCRM advantage is that if you start self-hosting with the free version, you can go cloud later with no-to-low data migration issues. On the other hand, Salesforce bought Pardot and is buying up the landscape around SugarCRM, which is a famously bad mumbo jumbo of PHP. I'm not complaining about PHP, I'm complaining about the kind of PHP that makes PHP developers look down on their fellow PHPers in the way that Lisp developers look down on everyone else.
5. Filesharing: We moved to Google Drive almost entirely. There is still some Dropbox around for some reason that eludes me. This stuff is cheap and has forms that work as a Windows share (aka CIFS/SMB/Samba) or mirroring, etc. You can even share things at http links. Seriously, if you haven't at least tried Dropbox or a similar service, I think you probably do live under a rock ... just saying. Constantly available (minus those 30 seconds a few weeks ago), constantly scaled, and reliable.
6. Revision control: If you need private or public repositories, GitHub and Bitbucket are your friends. These services have been tested and proven reliable by scores of developers. Your self-hosted repository has been sucking valuable development hours away from billable work for far too long.
7. Issue tracking/project management: There are plenty of tools for these activities -- and there really isn't a compelling reason to host your own JIRA (for instance).
8. HR management: As enjoyable as email and calendars are for vacation requests and other HR juiciness, you eventually need HR management software. This stuff can cloud pretty easily unless you're a gigantic organization with some massive customized Peoplesoft thing (in which case you have fun with that, I'll be enjoying my nice Web interface).
Stuff to think about, then cloud
Not everything will be effortlessly cloudable. Some things require more thought and planning. Consider the following:
9. Identity: We went Google. If it's good enough for the NSA, it's good enough for us! This may actually be a crucial choice before going all-in cloud, since the cloud offerings you select need to work with your identity provider.
10. Database: Scale in concurrency, scale in size, reliability, recoverability, and all of that juicy stuff. Especially if you are considering newer databases like MongoDB or HBase. These are available as utility offerings from your PaaS or IaaS provider.
11. Application servers: Do you enjoy installing and maintaining infrastructure around Tomcat or PHP or Node, etc.? Do you love waiting for WebSphere to start up? Maybe you won't be sending existing, complicated legacy apps to the cloud soon, but you can at least observe the first law of holes and stop digging.
12. Telecommunications: Why have a PBX onsite? Why have video or conferencing stuff? All of this can be maintained for you. If you have extensive legacy infrastructure, this is a little harder to do overnight.
13. Office suites: Seriously, you too can learn Google Docs or something that allows you to do attachmentless sharing and collaborative editing. Let's not understate the effort -- graphics and other things don't translate well, and there is still a feature gap. Granted, most people don't use those features, and many can be "done without" in exchange for better collaboration features (not having conflicting edits) but it took a relatively young and small company over a year to effect this change. Rome wasn't burned in a day.
14. Load testing: You're going to be hearing about this a lot in the next few months. Frankly, the traditional load test tool, Mercury, sucks majorly. Sure, it's very powerful, customizable, etc., but quite often you need some geometric multiplier of your server farm's horsepower in order to test with it. Moreover, it requires a lot of care and feeding. Also, people with sufficient expertise in it are hard to come by. Watch for some interesting new players in this field.
15. Continuous integration: Jenkins is the powerhouse here, and if your PaaS provider hasn't integrated it like Cloudbees has, I expect it will soon (or in the case of Google App Engine, through a partnership with Cloudbees). I'd throw this in the "thoughtless and now" category except for the number of dependencies. You could cloud your internal CI, but certainly it is easier once your source control and application server are clouded.
16. Wiki: Maintaining your own wiki is a headache that you don't need. Watch out, though: Many cloud-based Wikis are of the "really crappy imitation WYSIWIG editor" variety instead of the wiki-power-user friendly variety. Whatever you do, don't let anyone talk you into dumping your wiki into Google Docs, whose weakness is (surprise!) search.
17. CMS/website: I was once on a project where they insisted on spending millions of dollars on a custom corporate presence site. Granted, your highly interactive stuff won't be on Drupal or a flat CMS solution, but the stuff your marketing people edit should be. Don't go nuts with Teamsite or something like that as users won't understand it anyhow and it will be a cavernous thing to care and feed and debug. Go as close to click and edit (Adobe CQ style) as you can.>
These are the things that I've been watching people do with relatively ease or think that are on the immediate horizon. What else have you clouded pretty effortlessly or are looking to do?
This article, "17 things you should go ahead and cloud," was originally published at InfoWorld.com. Keep up on the latest developments in application development, and read more of Andrew Oliver's Strategic Developer blog at InfoWorld.com. For the latest business technology news, follow InfoWorld.com on Twitter.
This story, "17 things you should go ahead and cloud" was originally published by InfoWorld.