Welcome to the new episode of tech talks with outstanding Microsoft community members from all over the world. Most interviews are with Microsoft Most Valuable Professionals (MVPs), and if you are not familiar with that program yet, I recommend you reading my recent introductory interview. In this episode I talk with Didier van Hoye, a Belgium based MVP, Virtual Machine. I interviewed him during E2EVC in Hamburg and we were talking about his self designed disk2disk backup solution and how he see hardware vendors as a long time customer. Enjoy reading!
This post has no relation to my job or my employer. Everything I post is my personal opinion and I write complete independent.
Editorial processing done by Rafael Knuth
Flo: Didier can you please introduce yourself and your company to the community?
Didier: Thank you Florian, well my name is Didier Van Hoye and I am working for Agiv which is the Flemish government agency that provides geographical information services to both private and public sector. I am quite active in the Belgium Microsoft Community and that’s mainly due to my job, because that’s the technology I work with.
Flo: As an MVP what’s your professional focus area?
Didier: My MVP expertise might give that away, I am on Virtual Machine which actually means Hyper-V. So my focus at work is mostly on storage, networking, virtualization and anything that runs on top of that. We own actually the entire technology stack which makes it very easy and very fast for us to provide services to our business.
Flo: You recently built an awesome disk-to-disk backup solution. Can you tell us a bit about that?
Didier: Sure, the reason why we build it is that we have large amount of data to backup, and we need a backup solution that’s detached from our SAN environment because we build for failure, and we assume that we might run into an issue with our SAN, so if the backups are soley based and dependent upon the SAN we have a problem. Issues we ran into were that we needed a tremendous amount of capacity and also lot of throughput to maintain our backup window. While we were looking at the market, we found that most appliances that were being offered were too expensive, and that was mainly due to licensing issues. So we decided to see if we could build something that would provide us with the necessary capacity, the necessary performance and save us some money … and actually we achieved that using PowerVault disk bays to build a disk-to-disk based backup solution.
Flo: How easy is it to scale up this solution? You have 1.1 Petabyte at the moment, is that correct?
Didier: That’s source data with redundancy of data and protection. It’s replicated … so it’s not all unique data. But the storage capacity is that big, that’s correct. So we needed at least 300TB of backup capacity. But we can scale out easily by adding nodes.
Flo: Do you just need to plug in more nodes?
Didier: Basically now we have two backup media servers attached to the PowerVault storage, and if we need more bandwidth, if we need to do more terabytes per hour, we can add more nodes each with its own storage. In that way we can scale it out. The two nodes that we are running right now are at full capacity, so we can’t scale them up. But even if you scale up with a third node, and you only use half of its disk capacity from the start you could scale that node up later before you need to scale out to a fourth node. So it is pretty much “unlimited”. You can scale out to the numver of nodes you need. The only thing that you need to take care of is what kind of backup software you are using and how is their licensing model. For us it works just fine.
Flo: Are you running into any issues with your backup solution?
Didier: CommVault is a very good backup solution but it does have some drawbacks. One of the drawbacks is licensing and the cost of it. Let’s face it, we are all asked to do more with less and sometimes that’s an issue. Another issue with CommVault … it’s a very large product, its rather complex and in our opinion and experience it’s a bit too much consultancy driven. I think they would do themselves and customers a favor if they would be more open and if they had better guidance about their product, the usage in different scenarios … that sort of stuff. We cannot complain about the performance and the capabilities of the product, the issues I mentioned here are the most important ones to us.
Flo: How do you see Dell’s role in the IT industry change in the future? What are in your opinion the key opportunities as well as the challenges for us?
Didier: Well, there are a couple of things that are very good about Dell, and one of them is the company’s responsiveness and logistic capabilities. Dell is very easy to work with. As a customer I prefer to go to a shop where I am treated as a valued customer, where people help me to get the stuff that I need, and I walk out satisfied. Also, the price is very competitive … I can’t complain. The hardware is good, the pricing is good and the service is good.
I think that Dell has to take those strong points and leverage them for the future over the next five years or so. We see a tremendous push to have a more capable software. If you look at what has happened to Exchange for high availability, the entire application and shared nothing high availability and recovery model. If you see the same parts with SQL server 2012 for data base of availability groups. If you look at what’s in Hyper-V with Windows Server 2012 and the inbox replication. If you look beyond the hypervisor to the operating system … what you can do with storage spaces, what they are trying to do with SMB 3.0 as a file share for storage … for virtualization … for SQL servers. These are all attempts to create highly performance storage on commodity hardware with inbox. Those not will replace all the storage solutions from hardware vendors we have today. They are however a new player in the market and … it’s not just about making things as cheap as possible, it’s also about creating new possibilities because some things you want to do, you can’t do because they are economically unfeasible, their costs are prohibitive. But those new capabilities in Windows might might open up new possibilities in the market. People might start building solutions that previously they couldn’t because they were never able to afford to do so. It could even drive new business models, and what I expect from Dell is to take these opportunities and start putting offerings in place for customers to leverage storage spaces, to leverage SMB 3.0, to build Cluster in a Box building units that you can use … preconfigured for people who don’t want to build them themselves.
I have noticed a lot of hardware vendors that are not always up to speed about what’s possible now. This might be due to fear because they think: “This is going to eat our profit margin!” … which I disagree with. I think if you take the new possibilities and offer solutions you will not lose those customers. Otherwise somebody else would start offering them.
My personal advice to any Hardware Vendor is: Get into the game! Start offering solutions, and give your customers more opportunities based on new Microsoft products. … and you will keep your customers. They will be more diversified but you might open up new markets that you might otherwise lose. So don’t be afraid of the change. Embrace it and … leverage it.
Flo: One last question Didier. Can you explain what Cluster in a Box is?
Didier: To explain Cluster in a Box you need to know Windows Server 2012 clustering and possibly SMB 3.0. Let’s say it’s a couple of servers that are attached to shared storage. What that is depends on the vendor … but most probably it will be shared SAS as it make the most sense here and works well with storage spaces. It also has the networking in there, so basically you have all the material in a unit to build a cluster. That cluster will be used as a building block. If your company is smaller and you are satisfied with one Cluster in a Box … that’s your entire cluster environment. If you’re bigger you could start combining them, and then you get into technologies like SMB 3.0, RDMA (Remote direct memory access) being leveraged for cluster shared volumes; the sort of technology which customers can’t look into due to lack of time.
But it’s a form factor … that’s the way you have to think of it. It’s a form factor you can use to scale out but also you can use it to build a solution for a single small environment, which is a nice idea because it reduces a lot of complexity. Customers can order it, deploy it in their datacenter, use a small configuration wizard … and you’re up and running. You don’t have to deal with buying the storage yourself, configuring it, installing the services … everything is done for you in that box. As you can buy it as a form factor, as a building block you’re not limited or stuck. You can grow with those building blocks, so you have modularity in the datacenter on a lot of levels … in storage, in the clustering, in the network. Otherwise to get anything like this you have to go to that shipping container form factor. Let’s face it … for most of us its way too large.
Flo: Thank you very much for the interview.
Didier: Thank you, Flo!
Contaktinformation Didier van Hoye:
Didier’s Blog: http://workinghardinit.wordpress.com/
LinkedIn Profil: http://www.linkedin.com/in/didiervanhoye