Cloud computing: Anyone successfully using it to replace the local file server?

These days, people continue to repeat we don’t need all this hardware anymore. Just laptops for the employees, and everything else should be in the cloud. What about the file server that’s used for employees in the office to edit word processing files? Wouldn’t there be a lag time for opening and saving documents across the internet connected to an AWS data center? I guess it would depend on how near you are to the region where it is stored.

Anyone successfully using their office file server in the cloud with it being hundreds of miles away?

Objects tend to be far better than files, but yes, I move clients to the cloud every week although I typically work with much larger needs that this.

Use object stores whenever possible, file-systems do not gain most of the benefits of horizontal scalability and reliability that the cloud provides as they intrinsically are incompatible with a distributed model without jumping through hoops.

This will be due to features which you probably don’t need, and probably use a file server due to historical familiarity.

How is using an object store going to address the speed of the connection between the traditional office filled with employees and their actual data they would use for word processing, excel spreadsheets, power point, etc. if it is located in an AWS data center hundreds of miles away? Wouldn’t there be a performance lag to deal with, where the employee does saves to a document? Or are you still having the employee edit it locally and when they are finished with the document upload it to the cloud?

I gather you aren’t using something like AWS EC2 with Samba (SMB) to store their file server in the cloud?

In the accounting firm I work for, we store a lot of documents in the cloud so they can be accessed from any computer that has Citrix. However, I don’t like editing them while they’re in the cloud for a reason that you noted: input lag. When the process that opens the file runs locally, then there’s no input lag. So I can open a Virtual Office Excel window and have it load an Excel file from the cloud drive, but after doing so, the document is being edited on my machine until I’m done with it, and no one else can look at it until then. It’s perfectly fine to work with in that case, because the only input lag is on opening the file. If I look at an Excel document that’s stored in File Cabinet and want to see more than the preview sheet, I have to open it from File Cabinet in a special window that runs Excel inside of File Cabinet such that the file doesn’t lock, but it’s hell to work with. If I have to work with such a file for more than a couple things, I will save it locally, edit it, then upload it to File Cabinet again overwriting the old one.

While there was a time that people stored most everything they worked with in Virtual Office, I personally have never liked doing it because there’s an extra step in the process. If I want to share a document I’ve worked on, I can upload it to File Cabinet or save it on the virtual cloud drive, but a great deal of the time I have no desire to share it and it wouldn’t be presentable to share because I often don’t waste time labeling things. Scrap Excel workbooks don’t need to be saved to the cloud. And I still need a machine that can run Excel. I still need a full featured computer even if all my software is stored elsewhere, because unless it’s actually running on my computer, it’s hell to actually use.

Cloud computing sounds good until a workman with a digger cuts the data cable. Or your cloud provider gets hacked. Or your cloud provider goes bust. Or… I’m sure this is rat avatar’s bread and butter and he can give chapter and verse on mitigation strategies.

We’ve migrated from a big piece of onsite tin for file storage, out to Microsoft Azure on a DFS server - there’s still a small server at each site that does some clever caching and replication magic so that frequently used files get better performance - but over a 100mb fibre line, there latency of opening a file isn’t too bad, even if it isn’t in the local cache.

As for security, your own company’s data center is very likely less secure than a cloud provider like AWS or Google Cloud.

A workman can cut the data cable to your company’s data center also. The smart thing to do, is to have redundancy, and with more than one cloud provider if needed.

Not familiar with Azure. Are they offering a product that does the caching for this? Or what is being used to accomplish this?

It is turning out to be a great solution for school districts like ours that can’t afford to pay an experienced network manager to maintain a relatively small but yet vital network.

Most of our stuff is cloud based now. We have Google services for email, all of our apps, and for all of our accounts, and our Power School is net based also. We have a domain controller, special ed server, FOG server, and an HVAC server locally housed. Eventually, we want to virtualize everything.

My company uses hosted SharePoint as a file server and yes, there are small delays in opening and saving but it has negligible impact. If you a highly file-intensive business and staff are constantly opening and saving things then it might have more of an impact.

Citrix is more than just a file server solution, it’s a virtual desktop. And it can be slow. I use Citrix to access a federal system and it has gotten better over the years but it is slow compared to a physical desktop.

Less likely to cut a cable from the server room down the hall.

I do this stuff for a living and Cloud storage is a godsend. I work for a megacorp with redundant data lines and dedicated staff maintaining the Cloud servers hundreds to thousands of miles away. It works just great and is much, much better than local servers. Performance is just fine but we have the best infrastructure available.

What I can tell you is that Cloud storage is the future and it is better to get started on the migration sooner rather than later. The pluses greatly outweigh the minuses. The only minus I can think of is loss of local control of the server itself but even that can be a benefit if another responsible party is maintaining it for you.

Its called Dropbox and its been around for over a decade now.

I can’t think of anything more slow than Dropbox to use, but thanks for the laugh.:smiley:

I know there are people pushing back from using the Cloud, because they justify their existence in an organization by having the large budget for a huge computer room and staff to handle the hardware. Also the concept of it being located elsewhere makes them uncomfortable. And there are some people who simply don’t want to change and learn new things.

Other than the concerns of local lag for users accessing and editing files, I don’t see any reason to keep a local data center any longer. I loved walking into them and being surrounded by the hardware, but that’s just my association with the experience, and that experience isn’t worth the overhead any longer. CIOs and CFOs see the cost savings, and that as usual is going to drive every action.

Years ago, a friend of mine was telling me I should have a server running at home to host my personal websites instead of paying for a web hosting service. For fun, I figured it out and to keep a single server running 24/7 cost twice as much in electricity than what those huge web hosting services were charging to host a bunch of my domains. Also, my ISP’s terms of service didn’t allow me to run a public server anyway.

You are right, the Cloud is the future.

I’m sure a risk assessment expert could tell us, but a worker in the building could also cut a cable too. In any computing system, redundancy is the key.

If you saw the guys managing my server you would think it’s a plus. :wink:

Dropbox is a replication scheme, not exactly the same thing as a file server, although it does maintain a replica in the cloud. I use it for my personal files, and I think they have an enterprise solution; not sure how good it is.

One gotcah I can think about would be some kind of network latency or throughput concerns, but I’d also bet that the vast majority aren’t data-center side issues, but customer-side issues.

The only other one might be security; my last employer was a large healthcare company and we kept all our data in-house due to security concerns. Whether or not that’s ridiculously paranoid, I don’t know, but I do know a lot of their other security decisions fell squarely into the ridiculously paranoid and controlling arena. (they basically shut down ANY method of getting data out of your PCs save emailing or using the company’s network, and even the emails had some kind of big brother monitoring software that got bitchy if you tried sending anything compressed or encrypted).

A lot of office work is data entry, analysis, reporting. And a lot of that can be done quickly and easily on low powered devices being fed both the data and the application from the cloud. If you are experiencing “lag” on a cloud based application, I would submit that this application is not properly architected.

The experience should feel responsive, native, and functional, even when there are issues with internet connectivity.

Can you give us the name of the application you are referring to and devices you have working in place to accomplish this?

The biggest issue for the sectors in which I was involved was security. Obviously if you’re doing secret squirrel stuff the data has to be on site. Less obviously, if you are competing internationally, your data could be vulnerable if it is outsourced, particularly if you are competing with Chinese companies. But consider this: if you were competing with American companies, should you really trust your data to Amazon, an American company?