 Well, good morning. This is John McArthur. Let's get started This is November 13 2012. This is our Pure Insight research meeting and today's topic This discussion is enterprise class data protection on a small business budget Just a few reminders before we get started if you're not speaking Please mute your line with star six and you can unmute with star six Conference is being recorded and as a reminder the conference call is being recorded is being recorded If anyone is tweeting today wants to tweet out comments We'll try to monitor that if you have questions Dave Vellante who's here in the studio with me will watch the the tweets a hashtag wiki bond So wi ki bon So any questions or comments just please tweet those out We have with us today Paul Martin Paul is the IT manager at Poulin grain And I met Paul at a recent VMware technology user group Meeting here in Marlboro, Massachusetts just down the street from the wiki bond offices And so we were having an interesting discussion and I asked Paul if he would be willing to come on and discuss His environment his challenges and some of the things some of the technologies that he's looking at to try to improve his Data protection capabilities. So Paul. Thanks for joining us today All right, we also have with us David Fleuer on the line. I I think I I think I heard David on Nick Allen also a wiki bond analyst and here in the studio with me is Dave Vellante Dave and David our co-founders of wiki bond so So Paul, why don't you that the topic of today's discussion is enterprise class data protection on a small business budget And you really do have a Small business budget. Why don't you start by just telling us a little bit about Poulin grain and and then some of the Challenges that you're trying to trying to address All right, John. Thanks. Yeah, Poulin grain has been around since 1932 We are a fourth generation Vermont family owned company We've got three manufacturing facilities to in northern Vermont one of which I happen to be at Currently our corporate office up in up in Newport, Vermont In addition to the two Vermont locations. We also have a manufacturing facility up in upstate, New York Those three locations keep me quite busy on the only IT guy we've got two data centers one primary here at the corporate office and then one disaster recovery location and We service about a hundred and twenty independently owned distributorships all around the northeastern part of the United States we've We're a small company, but but what we do service a servant service a large population We I mean as as such we we find ourselves kind of in a unique position in that we are environment as We see very real problems. We see very real IT issues that are occurring today even in our limited environment because we are trying to do so much with so little That'd be in the case is kind of what has brought us here today. How many No, go ahead No, I was gonna say we we find ourselves in a position where our backup current backup infrastructure isn't meeting our needs So go ahead. I'm sorry. No, no, that's fine Tell us a little bit about your current backup situate environment technologies that you're using What's working what's not working and where the challenges are In in 2008, we virtualized our our entire infrastructure And it's worked well for us. We've adopted theme as a as a backup solution backup software provider and that That software is backing up to a primary backup repository made by Exegrid the Exegrid is being is is de-duping and and Replicating itself to a disaster recovery location, which is one of my remote offices And right now we find ourselves running out of capacity on those exit devices on both sides on both sides You're running out of capacity Yeah, initially a couple of months ago. I started really noticing that the the The back of capacity at my DR location. It's a smaller capacity exegrid Because we weren't Sending all of our virtual machines. We weren't sending all of them over to the DR site The thought was that we would only go ahead and send critical data over So initially what what I was seeing was that the the DR Locations Exegrid was was filling up At the same time too, I mean it goes without saying that the primary one was also Starting to fill up, but it wasn't as noticeable because it's a higher capacity box Now we're nearing budget time We're nearing the end of the year and I'm also nearing the end of capacity on that on that the our locations Exegrid appliance I Also finally noticed that the the primary locations Exegrid is Also in a position where we're it's starting to run out of capacity So now initially I thought that we were going to have to buy Capacity for the DR location, but as it really turns out The true fix is going to have to be to to buy additional capacity for both sides Or figure out some way to to give myself more storage capacity of both locations. This is always There's always kind of a standard standard question. What do you have a sense of what your data growth rate is? right now what we have well, we have found we bought the first Exegrid in 2008 we bought the second one in 2010 and Based on what I've seen on those two appliance that we're growing at just under about a terabyte a year our Backup data is growing for just under a terabyte a year. Okay, and and what are the now? What are the applications that are driving the growth? Primarily primarily the I have to say that the the biggest application driving the growth is our line of business software I Mean you can do whatever you want as far as we know limiting people's mailbox sizes and telling them to delete old files They may no longer need but you know unfortunately to run the business and then the largest offender For lack of a better word in the in the capacity Area right now is our is our ERP are our line of business software. Okay so what are the what are some of the options that you've considered in terms of Dealing with the the growth in data and dealing with the The need to back up and then replicate to the recovery site Well, I've visited this actually a number of times in the past Directly or indirectly depending on your point of view in that what I've done is instead of Replicating all of my servers the first thing first step that I did was I started separating out the critical service from the non critical ones and then to Take that stuff further. We have you know tried to prioritize and take ones out that maybe weren't quite as critical. We we've Added capacity we over time. We've we've we've trimmed files over time We've changed retention policies. I think I'm down to like an eight-day retention policy on my file server Ideally, I like to keep a couple of weeks I mean, it's not out of the question for somebody to come up to you and say they need a spreadsheet that they Overwrote from a couple of weeks ago. It happens to all of us But I've had to adjust my retention policies because because of the capacity constraints they have and now I'm looking at the hopefully the ultimate solution of Cloud backup And for for you, what's attractive about cloud backup? I think first and foremost the I don't want to say limitless capacity, but nearly limitless capacity I I don't like and I think it's an extremely inefficient waste of time to have to keep going back and revisiting The same problem over and over again I grant you if the problem was fixed right the first time but really who knows how much your data is going to grow And how much your business is going to grow in a perfect world both would would grow exponentially The business would grow and as a result of the business growing the data would as well I think that a Driver for me is gonna is gonna be cost It's gonna be a huge cost Are you I'm sorry a huge driver for me ease of use and and the fact that For me it needs to really implement with the technologies that I've already worked so hard to put into place And that seemed to be working so well for me so there needs to be so the beam so the beam I'm sorry the beam software is working well for you and The and the exit grid is working well for you. Is that what I'm hearing? So you want to incorporate they are both working well The problem is that the exit grid appliances as a costly solution it works extremely well But but it's cost-prohibited for me to continue I believe Using using exit grid as my as my backup storage solution. So Paul. This is Dave Vellante. I wonder if you could just address Some of the higher level trends here. We've seen for the last say five to seven years data deduplication really came in as a way to reduce cost and it it sounds like in your case it might have The solution might have addressed some kind of a near-term tactical problem, but it didn't solve your issue long-term which seems to be Elasticity you want the ability to just basically dial up capacity when you need it So the the data deduplication is a kind of a one-time hit that draws capacity down And then you're back to adding adding more capacity. Am I getting that right? And I obviously can't speak for everyone But in my case, I think the biggest advantage to deduplication was realized when I entry John entry and exit times are off Sorry, sorry Paul. I'm sorry Paul. We got we got cut out there for a second. Can you just repeat what you were saying about in your case? Yeah, absolutely deduplication I think the the the biggest advantage to deduplication is Is realized when you first put it into place Largest advantage the largest that you'll see is when you first implement deduplication after that if business is usual I mean backups as usual is The amount of data is the amount of data is the amount of data and over time That's just it's just going to continue to grow. It's just the nature of the beast You it's nice to say that yet deduplication is is is the answer to all of our problems, but After that initial implementation where you see all of the all of the duplicate blocks or all of the duplicate files all of a sudden, you know Fizzle down and and and and you know down down to their smallest form Once that's been done. I mean you're just you're backing up data after that Can you talk about more? Can you talk a little bit more about your your cloud backup solution and specifically what you did all right right now? If I get if I've got that right you're you're looking at but you haven't implemented cloud backup, correct? That's correct. In fact, I'm looking at a lot of different players I've talked to some of them that have been very very helpful. I'll talk to Interestingly enough a couple of them that haven't been very helpful I mean first and foremost I'm coming into this, you know really with with a learning curve ahead of me I don't understand the technology. I don't I shouldn't say that I don't I didn't Understand the technology and not that I'm any expert now, but I didn't understand the requirements of it I didn't understand the cost behind it I mean Like any other project in our world we there's a learning curve ahead of it and and you know There's there are a lot of a lot of things that you don't necessarily know up from going into it And thankfully I did run into a couple of people that that were very helpful in sharing the information that I needed Again to answer your question. I've looked at Gosh, I'm going to get probably 8 10 12 different cloud providers so far and I'm only a little bit closer than I was two weeks ago to finding a Solution so but basically the the the justification the business case is as we were talking about before that ability to get capacity on Demand use you call it nearly infinite So what have you learned in your research and and what are your concerns and you know? How are you evaluating the the risks in the rewards? You know which which technologies of Do you think are promising and which are not? Well, I mean given my my position My team manager of an SMB My budgets are very limited So cost is going to have to be a huge consideration for me of all of the solutions that I found I mean, I don't know if you want me to say specifically who I've found for providers This is a this is a pure insight community And so we want people to share both their pain and their and their progress so that other people can All right well The the the the one the one Provider that that seemed to really stick out or stand out for me Amazon has a has a project called called the Glacier project the the resounding theme regardless of the provider that you're going with Seems to be that they all charge on a per gigabyte data amount per month Translated that comes out in Amazon's case one penny per gig per month Right now I'm looking at about about probably three to five terabytes that I'm that I have for for data that includes That includes my retain data as well my retention policy So the so that's one penny per gigabyte per month Very easily fits into the budget that we've that we've allowed for for data growth for next year Does that also enable you to start does that also enable you to start protecting applications that you're not currently protecting? No, and and and that penny is only for the cloud service There's actually two pieces to that not only do you need to pay for for a cloud service service provider But you also need to pay for a cloud enabler you need to have somebody sitting in the middle that's going to translate Me into the cloud language and the cloud into my language Some broker that needs to sit in the middle to act as a as a caching appliance I mean they call it a caching appliance my personal opinion It sounds to me like like the primary storage because and because what they're doing in a lot of cases is They're using this caching appliance as a point to restore from I Don't know. I guess the definition becomes a little gray between local storage and caching appliance But it becomes a little more clear I suppose when you look at these caching appliances in the end from the aspect that some of them are physical appliances Okay, others are virtual machines an OBS file that I throw into my infrastructure Well now that tells me the caching appliance in that case becomes my own storage infrastructure So I guess really throw your own definition in there and and and is this thing a local storage or not But again, you need an enabling device In most cases and at least in the ones that I've seen they call us a hybrid method of cloud storage There's a couple of companies That were that were very with us a very very good a panzura And twin strata the folks that I spoke to at both of those companies Specifically a particular the panzura Person that I spoke with very very very friendly very knowledgeable in fact that when I was speaking with the rep Steve with the gentleman name the the owner of the company I think was the owner of the company or the president of the company was walking by his office at the same time stepped in the office when he heard Steve and I talking and and and and and proceeded to have a dialogue with me and Teach me About cloud and teach me about their solution even when it went so far as to tell me that even if they're not the solution for me Maybe this one is maybe this one isn't and actually gave me a bunch of literature Gave me a bunch of names of companies And Steve was even willing to give me his contact at Amazon Web Services to tell me a little more about their Glacier cloud project The cloud cloud service offering so I mean and that means Quite a lot because he's selling somebody else's product in that case So let's talk about this a little bit so Glacier for those of you don't know is the Amazon's new service It's a deep archive service essentially it's as Paul said It's a doll a penny a pair per gigabyte per month. So the the price is right But it's not designed for retrieval. It's really designed for you know, you put it there and you may never see it again But but I wonder And then of course you you're talking about the panzor and the twin strata, which would be a Gateway and a cloud on ramp essentially writing, you know Translating to Amazon's API's which is the other thing that you've got to do, you know, the restful API's so Paul I wonder if you could address that that Glacier notion of sort of deep archive Is this the type of data that you sort of hope to never have to get back again? Or is it Go ahead. That's a very good point I guess the first thing that I should clarify is that I'm trying to find a disaster recovery repository now Right in my case since I'm looking for a disaster recovery repository Let's say I'm looking at secondary backup storage the AWS the Glacier project actually fits very very nicely into that So to you to your to your question. No, I hope to never ever touch this again Because that would mean that the big meteor dropped down on on my office and probably took me out with it My hope is that that never happens But yeah, the Amazon Glacier is more of a long-term archiving repository as opposed to a live cloud storage Well alive exactly that I spoke alive alive cloud storage destination. Yeah, it's not an active archive It's a deep archive that you you're hoping you never have to get to and we're not talking about restoring application We're not we're talking about restoring applications. We're talking about restoring data recovering data Yeah, and Amazon will of course charge you Restoration costs depending upon how fast you want to get it back and how much data So you're I'm sure well aware of that. They're not the only one They're only one let me clarify that there there are additional charges when you When you subscribe to some sort of a cloud service provider in addition to the to the to the appliances and the stuff that you Need to buy you need to pay for the support you need to pay for the product of the appliance You need to pay for what they call put it you need to pay for what they call get and in addition You're going to pay for bandwidth charges Depending on how much data you're sending across the pipe They may they may they may charge you based on on the amount that they're sending Interesting lineup and this is kind of funny the the initial backup Of course is gonna is gonna take the longest of all of them some of the providers that have looked at actually will ship you an appliance for the initial streaming of data the initial backup and And and then you send it back to them and and and that's how they do their cloud service It's it's really been an adventure to learning All of the different techniques that these different yeah, you're right seeding seeding the cloud is Oftentimes one of the hardest things to do and it's a one-time Activity, but it's time-consuming and the different suppliers have different ways to throttle up or throttle down the pace at which you seed and But that's a big deal obviously getting getting your data into the cloud Yeah, and in your case we're talking about three terabytes. We're not talking about you know a mid-size or large company where you got tens or hundreds of Terabytes, so but even three terabytes That's what I'm saying. So that's what I'm saying at three terabytes. It's an issue What is it when you get to 30 to 100 or 200 so so that's interesting Yeah, I Wanted but I wanted to stop for a second because I want to make sure that People on the phone have an opportunity to ask some questions of you Paul or or anyone else that's Hi Paul, this is Scott Lowe. I do have a couple of questions for you You know, I was a CAO in a small environment for quite a long time Do you find that working in a small environment? It's just what's you a sort of a disadvantage when you're working with some of these companies like they're more used to working with with quote-unquote real enterprise companies versus the SMB or Has that not been the much of a factor? Scott, thank you very much for asking that question first of all, hi, thank you very much big fan and There's a company that I ran into that was recommended to me called the Nassuni and They they sell what sounds to be a typical product there. That's not very unique In any way, I don't think but what was really interesting to me was was the the person that I spoke to About more information the person that I that I called for more information Seemed to go out of her way to try to talk me out of their product when she found out that that we were an SMB She didn't do a whole lot to try to sell the product She did a lot more to try to talk me out of their solution Did she direct you to something else? Did she direct no, and then that was the interesting part She wanted to get me off the phone as quickly as possible I'm not quite sure what what sales technique that was but it didn't work You're not a prospect onto the next one technique Yeah And Scott thank you again for that question But but I think because we are a small business Maybe folks don't don't want to deal with us because they don't see us as being any kind of viable source of revenue for them Certainly not long-term revenue, but yeah Now when also when you know when you're looking at the cloud service providers you talked about a number of times There's a there's a charge for this as a charge for that there's a charge for you know Puts you get to the charge for being with do you feel like that you maybe jump it into this and be kind of like okay? When we got our first bill who knows what we're gonna see Because it's really kind of tough to predict. Yes, like the first iPhone bill I like I like the idea. I like the idea that panzora had I think I think it was they had they had a one flat fee 19 cents per gig per month That covered their support their product all your puts all your guests Unfortunately, that's that puts it out of my pricing range 19 cents. It's just a little bit too much. Yeah, that's that's a lot Yeah, and panzora is you know, they're they're considered an enterprise grade You know companies so you're gonna pay a little bit more for those guys for sure How about their nirvonics? Have you looked into the nirvonics club? No, I have not seen nirvonics I've got I've got five different cloud providers that that that I was that I was looking at and nirvonics is one of them Amazon is the other That that seem to be doing well Google cloud. I want to look at HP cloud and rack space as a cloud as well I Don't know cost wise where any of them stand But or if any of them have an enabling device that they sell with their service or not You're not looking at Oracle They're not gonna be a necessarily an SMB play, but they do have a virtual appliance that wasn't horrendously pricey Yes, yeah Tara was another one that actually seems to be They seem to want a lot of a word to send a lot of attention CT eRA Does anyone heard of that one? Can you say that what you said? Did you say centera or satara? The Tara CT eRA So that one seems to have won some awards and gotten some attention Door simple was mentioned a lot By folks that I spoke to Yeah, satara store simple satara cloud Again just a reminder to those Star six to mute your line We've got one loud line there star six to mute please For anyone else on the phone and one of the one of the goals here was to it was to try to Help identify any other players that Paul should look at and any other issues that he needs to consider Anyone want to contribute here No, I think I Whoever speaking who's speaking please this is Scott. Oh, hey Scott. Sorry. You were cutting out a little bit Okay. Yes. So Scott. Are you asking the RPO RTO question? I know there's a little bit discussed before but if you Paul's like basically said, you know, does your company have a formal? policy on Basically recovery time recovery point objective When there's a when there's some kind of an outage or something that takes place No, we don't have a an objective beyond it needs to be faster than it was the last time That was about 24 hours. I took to recover manual server rebuild. So The expectation fortunately for me that the bar has been set kind of low Anything that could do that's going to be better than that is it's going to be it's going to be a success I do understand and to speak to that the the cost piece of it does have a lot to do with the amount of Time to recovery the the Amazon cloud From what I've been able to see actually they offer the slowest service of all of them And that being being that the case maybe that's not the solution for for somebody that wants to use Wants to use it for primary backup storage But for me, I think it's going to work well because I'm not looking for that that kind of performance I'm looking for for the archiving piece of it and the lower cost solutions Some of some of the hosting providers and sort of regional resellers who are building out private hosting environments or managed services Environments are starting to offer services where they not only do