Help get my head in the cloud
October 8, 2013 8:08 AM   Subscribe

My company has about 40 employees and five Windows based in-house servers. About half our staff work remotely (using a terminal server) and we have redundant Internet connections. Over the years we have built up a decent data center involving battery backup and tape archival. As I look to the 2014 budget, which would involve upgrading two servers, I have been asked if we should go to the cloud. I'm honestly finding it hard to evaluate this and hope some others may be able to help me improve my pro/con listing. Extra credit for suggestions to evaluate the costs.

We run a small business server (2008R1) with Exchange 2007. This also acts as our file server that shares our network drive. The lion's share of data that we backup and utilize day to day lives on that one server. Our other four servers are for terminal services and a handful of specific applications.

The impetus to look into the cloud was while researching some on site disaster recovery appliances (Axcient and DattoBackup) and the question was posed if we would have less need for a fail over server if we were cloud based. To my mind this would only be true if the hosting provider, as a matter of contract, had spare hardware and was prepared to build up a new server in the event of a disaster. Moving the hardware and/or data onto servers in a co-location facility only eliminates risks that threaten the hardware in our building. A legion of other risks that threaten the data still exist if it is hosted.

The Pro side I see to getting out of the small business running a server closet scenario are:

1) We can dial down our Internet connectivity requirements. We pay good money for Metro Ethernet because downtime affects 20+ people.
2) A data center would likely have spare hardware that I don't have in the event of a disaster
3) Their facility is likely more secure
4) Presumably some offerings involve the ability to scale up your cpu and disk space needs as you grow, versus upgrading servers wholesale.

The Con side seems:

1) Everybody works remote. Presently half of the staff enjoy LAN speed for data access. While terminal services is empowering, it just isn't as fast as using a workstation on a LAN.
2) I have a good working relationship with an IT support company local to me. We have worked together for seven years and I have some doubts if I'll get as good of support and service from a hosted facility as I get with them coming to me on site for routine and urgent matters.
3) I still think we would need a disaster recovery system. Data and systems fail in data centers too.

Am I missing something? From what I can tell moving the servers off site doesn't help very much. I sense the problem may be that I'm thinking primarily of this in terms of where the server resides while cloud computing proper is about provisioning of resources. The problem is that everything I read about it seems to over promise and is light on the details of just how your applications run in the environment. This uncertainty makes me inclined to the devil I know, but I'm really open to points of view on this.
posted by dgran to Computers & Internet (5 answers total) 1 user marked this as a favorite
 
You've got your pros and cons about the actual environment correct, what you need to do is think like a finance guy.

If you own equipment it's an asset, and one that you can take depreciation on. If you in essense LEASE the equipment, that's an expense.

You really do need to run the numbers on this. Is there a cost savings? Personally, I doubt you'll be able to downsize your metro-E. It only comes in 10 and 100 M speeds, so even if half the folks VPNing in are using the bandwidth, you probably need the other half for the folks in the office.

People who don't understand networking think the cloud is magical. It's not. It's a server-farm some place else.

Now for some, that's great! They don't have the wherwithal to run their own servers, they don't have on-site IT or they don't want to deal with it, for whatever reasons.

If you see that there's a lot of growth in the future, involving a massive investment of resources (cash) and the business would rather not have that many depreciating assets on the books, well, time to start looking at off site hosting.

I'd start with discussing the financial aspects with your accounting folks, and then with the CFO to understand what's really driving this discussion.

It's always about the money.
posted by Ruthless Bunny at 8:37 AM on October 8, 2013


The one place you could win with cloud is that you'd likely virtualize those Windows servers onto somebody else's cloud resource pool, not just colocate your hardware in somebody else's facility. Failover then becomes a function of their infrastructure -- for a VMware-based cloud, for example, if your physical server goes down, the High Availability function kicks in and the workloads on that server are restarted on another physical server. The other place that cloud wins is, as you say, if you need rapid growth.

You would get a similar benefit by virtualizing your local servers in-house. Virtualizing with a hardware refresh usually allows server consolidation -- you might be able to replace those 5 servers with 2 or 3 and get HA functionality besides. Some things, like backup and maintenance windows, also become easier with virtualization. If that interests you, check out the Spiceworks Community, which has a fair amount of wisdom for small business IT, if you aren't familiar with it already.
posted by troyer at 10:16 AM on October 8, 2013 [3 favorites]


Duane -

I just sent you some MeFi Mail on the subject.
posted by tgrundke at 10:46 AM on October 8, 2013


It isn't clear to me what your role is in all of this, but in general, the best way to ensure your job in IT is to work to put yourself out of a job in IT.

You seem to be looking at this from the point of view of what are the pros/cons of doing things the way you do them now, only with your servers in co-lo rather than on-site. Instead, I'd suggest looking at it as how can we use "the cloud" to better solve our business problems.

Start by looking at the functions provided by your current servers.

Take your use of Exchange, it helps with communication and scheduling. Those problems can also be addressed by moving your Exchange server to co-location, but with downsides as you note, or you could solve them via a hosted Exchange service. This will have its own advantages and disadvantages, but the advantages will include the fact that someone else worries about not just the hardware and the operating system but also the network, backup, software updates, and, even end-user support. It also simplifies cost and capacity planning, because costs are steady and predictable. No need to anticipate the cost of equipment upgrades and repairs, or the point at which you outgrow your old hardware. Going further, this function can be filled by a non-Exchange service, like Google Apps.
posted by Good Brain at 4:30 PM on October 8, 2013 [1 favorite]


Response by poster: Thanks everyone for chiming in here. The suggestion about the spiceworks community lead me to one of the more clear-minded explanations I've read:

http://community.spiceworks.com/cloud/servers/reviews?guide=1#

In particular the "making your case" section about half way down gives a set of compelling use cases for when cloud computing makes the most sense. For my part, none of these are compelling and ultimately some privacy issues surrounding the sensitivity of the data we manage compels me want to keep the data in-house. This addresses the "where it lives" portion of the question.

I had a nice talk with tgrundke about disaster recovery and he really knows his stuff. One approach I'm researching is the ability to get all our servers running Hyper-V (as virtual servers) replicated to an off site server. There are a myriad of approaches to doing this and I'm just starting to get my head around it. The ideal solution will involve something that makes it easy to synchronize data off site and only begin paying for cpu+ram resources in the event of a disaster.

As it happens, the decision isn't financially motivated other than general reticence to spend money.
posted by dgran at 12:49 PM on October 9, 2013


« Older Legal CBD without THC? Really?   |   How Do I Take in Better Stuff? Newer »
This thread is closed to new comments.