One Guy’s Opinion: Managing the Real Risks of the Cloud   

One Guy’s Opinion: Managing the Real Risks of the Cloud   

July 31

On “One Guy’s Opinion,” Guy Baroan, founder and president of Baroan Technologies, discusses the technology world through the lens of a successful IT managed service provider.

This time, we asked Guy what his customers lean towards when it comes to cloud solutions, what can save them money, and what their cloud anxieties are. Guy also shares the story of a cloud vendor who left a customer high and dry after a hardware failure caused an accidental deletion of a lot of data. He’ll share how you can take matters into your own hands so that neither you nor your clients have to lose any data due to a cloud vendor’s mistakes.

StorageCraft: Do most of your clients use the cloud in one form or another or are some not in the cloud at all?

Guy: Most of the clients we have today have been transitioned to the cloud in one form or another, especially backups. We have a lot of clients who have offsite backups that we send to a data center that we have equipment at. We started with backup but now we’ve got all kinds of other applications like hosted Exchange and some people have CRM in the cloud.

StorageCraft: For those that don’t or for those that have fewer things in the cloud, is there resistance to moving to the cloud?

Guy: There is some. Some clients don’t have the same trust level with the cloud compared to something they own and that’s physically there. They often have to come to terms with the reliability. They wonder about things like uptime or say, “What if I lose my Internet connection, where is my data and who owns that data? What happens to my stuff if that company goes out of business?”

That was something we saw early on, but in the last three to four years with all the effort businesses are putting into the cloud, things are changing. There are big names like Amazon, Rack Space, and Microsoft who are all in the game now. They have the infrastructure and people know that as a company they’ll probably be around, so that’s not as much of an issue anymore.

Now the big question is how do people feel about having data offsite? I think more and more are OK with it because they like convenience and not having to do backups locally and not having to replace tapes and move drives offsite—they don’t have to worry about ongoing maintenance and that’s pretty appealing to them. Even the reluctant ones are becoming more comfortable with it.

StorageCraft: It seems like two things drive people to the cloud. One is convenience, which we touched on a bit, and the other is cost savings. From your perspective or from your clients’ perspectives, what are some of the most cost-saving uses for the cloud?

Guy: It depends on the applications. For instance, if you’re looking for cost savings, cloud really saves you in upfront costs. With a cloud solution, you don’t have to buy hardware or software licenses, and you don’t have to worry about implementation. Either way, cloud or local, you have costs going forward, but cloud eliminates a lot of those upfront costs. Another thing to think about with local equipment is the hardware refresh. Every three to five years, you’ve got to swap out your hardware and get new stuff. When you look at a comparison, especially Exchange, it’s often cheaper to be in the cloud.

With Exchange, let’s say a client moves to Office 365 with email only. On average, I’ve found it takes about eight to nine years for them to catch up to what it would cost them to purchase an Exchange server and the licenses to host email internally. When you factor in the hardware refresh we mentioned, you might never catch up—it might always cost you more because you’ve always got to replace that equipment.

With things like email, it’s a really a “duh” moment where you hit yourself on the side of the head. From a cost perspective it’s a no-brainer to move to the cloud, either through Office 365 or even with the Google Apps for Business. There are so many options it rarely makes sense to have an Exchange server in house.

But if you look at CRM, it might not be the same, and it really depends on who you work with. If you’re using Salesforce, they average about $35-40 a user but there are those that are $60-80 a month per user. If you have 10-20 people, then you look at these online CRM applications compared to buying upfront, they usually have closer to a 36 month catchup rate. You may be better off to keeping it internal if you have IT staff or an MSP to handle it. It just depends on the cost of the applications a client wants to use and what’s important to the company.

StorageCraft: What advantages are there in keeping things on premise?

Guy: The biggest disadvantage of the cloud is that the Internet connectivity people have is not the same as they get internally. Basically, it’s slower to use cloud-based applications than it is to use things that are locally networked. The biggest hurdle people have to overcome is Internet connectivity. If they have a T1 line and that’s all they have access to, they might not be able to do much in the cloud, especially if they use large files like those used in CAD or graphic design and video, it’s just not a good fit.

Once the industry gets over the connectivity hurdle and Internet speeds catch up to the internal network speeds, would it really matter if they’re in house when you can have them hosted in a data center with redundancy and full-time maintenance staff? If it’s all the same speed, does it matter? There are ISPs that offer 500 MB download and 300 upload, which is huge, and they really don’t cost that much for a business. If that’s something that was available for everyone, then you could just about move everything to the cloud.

But really, the connectivity issue is the driver that affects adoption because it prevents businesses from moving certain systems over. If they do move to cloud-based they might need to find simpler applications that don’t take up as much bandwidth.

StorageCraft: So aside from connectivity, there can also be security concerns. I recently read about a business called Code Spaces that was forced to shut down for good after a hacker gained full access to their network, which was hosted on Amazon Web Services (AWS), and demanded money to return control to the company. They ultimately didn’t pay up and the hacker destroyed their business from the inside by deleting tons of important data. Should instances like these deter people from using the cloud completely or should they be lessons about being cautious?

Guy: Well, I’ve got a story for you that sort of speaks to that. We have a client that’s a financial firm and they’re in the Amazon cloud 100 percent. They worried about what would happen to their stuff if Amazon had a problem, but they also needed to refresh their hardware. Really, it was a tradeoff. We went in and said, “Ok, your servers are five years old, these will need to be replaced.” Their cost was going to be $80,000-$100,000 upfront. The other option was Amazon Web Services. We told them we could put everything in AWS, but to make sure they had no issues with disaster recovery or backups, we would replicate their data from Amazon back to their office. Sort of a cloud first, local second approach.

One of the problems people often have with Amazon is that it’s really difficult to move off them once you’re in. So we’ve created a StorageCraft server on the Amazon cloud and we have that replicating back to their local equipment. We take snapshots of their whole infrastructure—which we can do with StorageCraft because it’s server level and not hypervisor level—and replicate those nightly from the Amazon cloud to their office. If something happens in the cloud, they have a complete copy of everything in the Amazon cloud on their local systems. From there, we use HeadStart Restore to pre-stage recovery of all of these servers. It would only be 15-30 minutes to spin up those machines locally if something went wrong with AWS. Clients can basically flip a switch and they’re ready to go.

So it really depends on the client. When clients move more and more things to the cloud, we tell them they need some kind of backup on their own, whether it’s back in their office or somewhere else so that they’re protected. You can’t trust that Amazon will have the solution in place for you—it’s the same situation for anybody, Citrix, Azure, whatever. You need to be able to recover if something goes wrong.

These things are certainly concerns, but you can plan for it when you’re moving to the cloud.

StorageCraft: That’s interesting, and in the story I mentioned, the company had no backups, just what was in AWS at the time. We were wondering if there was a way to replicate from the cloud and back to the office, something that’s actually backwards from our philosophy of local first, cloud second. Really, it sounds like your philosophy is that it doesn’t matter where your first backup is as long as you have one local, one cloud.

Guy: And that’s exactly what we talked about with the client.

Interestingly, I mentioned to a partner that we were working with Amazon and he said, “Oh my God, you should not work with them.” And he forwarded me an email that somebody he knew received from Amazon. Basically, this guy had no backups in place and trusted that Amazon would be able to keep track of his data and have everything up and running.

One day he received this email that said Amazon had had a hardware failure and this guy’s server was on the failed equipment. They said they’d had minimal issues like these in the past but in this instance they lost all of his data and that there was no backup. They had no guarantee that data would remain intact, and there was nothing they could do to retrieve it. The server was gone and there was nothing, not even an apology from Amazon.

This just confirmed to me that you can’t go to the cloud and think they’ve got everything configured for you. You’ve got to think about problems like these. You can do a replication from one region to another, or whatever, but you can’t assume they have it all taken care of. Some services have backups, but some don’t. You’ve got to take care of yourself.

StorageCraft: I’m guessing that’s buried in their service agreement. Somewhere it probably says, well, we’re not responsible if this type of thing happens, which means you there’s nothing you can do. Unless you find a cloud provider that can guarantee that there’s redundancy, but it seems like a lot of them don’t want to do that. It takes more resources for storage if everything is backed up.

Guy: Plus they just don’t want to be held responsible for that type of thing. Look at what everyone says, they have five or seven nines. The fact that they can’t say we have 100 percent uptime pretty much guarantees there will be some type of failure. Unless it says 100 percent—there will be a failure. It doesn’t matter if it’s eleven nines, they can’t guarantee it’s going to be up all the time.

And actually, we worked with a data center that at one time guaranteed 100 percent uptime. That was their selling point, they supposedly had redundant everything. Well they had a power outage during some construction that was being done nearby and it knocked them out for about five hours and bam, just like that, no more guarantee. They were the only one I’ve ever seen try to guarantee 100 percent uptime, but they couldn’t really do it.

StorageCraft: And it adds up fast. A few minutes can move you from five nines to four to three—just a little bit of downtime makes a difference and when big companies like Netflix host in Amazon and Amazon has trouble, you’re talking about thousands of paying customers without service. Sure the cloud is awesome, but I think they’re still figuring things out.

Guy: Yes. And when you’re thinking about the cloud, you’ve got a few different models. You have the applications side, like Exchange service– software as a service– or you’ve got the whole infrastructure, like what Amazon does and you’ve got the data center side, where you can use colocation and you’re maintaining things, but you’re still in the cloud. It’s tough to know what the right solution is, but the good news is that it’s evolving. The more people are coming into the industry, the more people are there legitimizing it. Plus there’s more competition and people are driving each other to be better and less expensive. This way everybody is a winner, businesses and consumers. Everything from bandwidth to services to reliability is being constantly upgraded. Who knows what will available in five years—it’s incredible the speed technology is moving at.

Photo credit: Horio Varlan via Flickr