Latest Updates: virtualization RSS

  • Is the new DroboPro for you?

    admin 8:07 on Tuesday, 14. April 2009 | View Comments Permalink | Reply
    Tags: , appliance, argumentation, , compliance, data robotics, device, , drobo. drobopro, express, , firewire, harddisk, , iscsi, jbod, marketing, mount, , , pci, pcie, presentation, rack, raid, san, sata, , smb, , terabyte, , virtualization, webcast

    Last week Data Robotics, the company behind the Drobo personal data robot, announced their new product: The DroboPro

    Like the current Drobo it’s a shiny black box that will fit any kind of 3,5” SATA disk you fit into it and create a virtual storage volume using up as much space as possible while keeping all your data save from loss due to disk failure. The improvements are mainly found inside the box. The Pro fits 8 instead of only 4 SATA disks, reduced noise by utilizing two bigger fans on the back and sports a 1Gb/s Ethernet port that’s dedicated to using iSCSI.

    This means you can now hook the DroboPro to any Ethernet port you have and connect it to your your computer via iSCSI. The only requirement is that your computer has to have a second Ethernet port, that you can configure to only use the iSCSI protocol. This will give you read/write speeds up to 100 MB/s on the Drobo disks.

    This all sounds great, so where’s the catch?

    The DroboPro will start at $1299. If you a small business and want to include the DroboPro into your server rack, there’s a rack mounting kit available, for another $150. If you want to add 8 reliable 1TB hard drives, this will be another $800.

    $2200 isn’t that much for a SMB storage solution, but there are some things about the DroboPro that bother me more than the price.

    1. The device is marketed as an SMB solution with rack mounting capabilities, but it’s only 12” wide and 3U high, so there’s 7” of wasted space. Usually a 19” 3U NAS has space for 12 3,5” or 24 2,5” disks, the DroboPro only fits 8 3,5” SATA disks. The DroboPro is totally single device, so there’s no way of extending, combining multiple DroboPro to act as a single device. This is bad news for everyone that needs to scale beyond the current 16TB barrier or 8 disk limitations.

    2. The rack mounting kit adds another $150 to the DroboPro price tag while it’s included for free with most competitor’s devices. I’m not absolutely sure what exactly is worth $150 from the pictures,  but for a mere metal mounting plate with plastic clip-ons it’s way over top.

    3. The DroboPro only has a single power supply. I’m not sure about Data Robotics or you, but I like my power supplies redundant, just in case one of them fails. Imagine you company unable to work because the DroboPro’s power supply is broken and you have no ways to access your 16TB of files until the replacement DroboPro has arrived.

    4. Security: The device itself cannot be locked. The only theft prevention is a Kensington lock connector on the back and the disks can easily be placed into another DroboPro. So, if you want to prevent a massive data theft, you’ll have to lock the DroboPro itself into a rack or shelf.

    To answer a few of my questions I attended a webcast on Thursday in hope that some of my doubts might be addressed. Instead of the expected product presentation for SMB customers it turned out to be a “for dummies” presentation. The host took unnecessary lengths to explain the different RAID systems (limited to 1, 5 & 6) and why Drobo’s “Beyond RAID” was much superior to them.

    With the info I currently have, “Beyond RAID” is a simple JBOD with a virtual storage volume file on it. The controller takes care of the data redundancy on the physical disks, so a single or dual (on a DroboPro) disk failure won’t lead to any data loss. Although this is a kind of storage virtualization, it’s not nearly as sophisticated as a professional storage systems or Data Robotics marketing would like to make you believe.

    Just as I expected from the setup of the webcast, none of my questions where answered.

    But for whom is this DroboPro really made?

    Data Robotics says it’s for professionals and small businesses that need fast high capacity storage. what they didn’t mention or answer is what a small business is for them. I’ve been in the industry for a few years now, and one thing I know is that every product vendor has it’s own definition of the SMB market. The range of defined SMB company sizes varies from 1-100 to 100-10,000. For the DroboPro my guess would be more in the 1-100 employee range.

    These are companies that in most cases only have one or a part time administrator with limited knowledge. No offence, but I met a lot of them and you usually only get what you pay for.

    If you’re one of these businesses and don’t care if your file services go offline for days in case of failure, go an buy one. It’ll be much better than any NAS of fileserver you might currently have. Just make sure it’s not placed on the Administrator’s table, secured with a Kensington lock.

    Should your company approach IT services in a more professional way, the DroboPro might not be your first choice. The missing power redundancy, limited scalability will drive you into a dead lock sooner or later and lack of device security could be a serious compliance issue.

    If you’re a professional working with large quantities of data like music, photos or videos this could be a rather price worthy, flexible and fast solution for you.

    In case You’re only looking for a easy way to store data and need neither iSCSI nor more than 16TB of space get a “normal” Drobo. Even if you only have a PC without FireWire 800. there are plenty of FireWire Cards for your PCI Express slot, that give you faster access than USB 2.0 does.

     
  • Who needs a Desktop Computer anyway?

    Mirco 19:24 on Wednesday, 14. November 2007 | View Comments Permalink | Reply
    Tags: budget, computing, , desktop, infrastructure, optimization, , terminal, thinclient, virtualization, vmware

    Imagining most of the corporate environments I’ve seen… Desktop PCs on every desk, users mostly use Office software or some business applications, and security policies don’t allow saving files on the local disk.

    Asked how to reduce the fixed costs and consolidating servers, my first question usually is: “Does everyone really need his/her own desktop computer?”

    Why am I asking? Corporate Desktop PCs produce a lot of costs. You need to replace them every 3-4 years, you need to maintain the OS and the software installed on them, they consume a good amount of power to display spreadsheets and text documents, they need space, and produce noise as well as heat.

    Do not underestimate the heat! We noticed that after some of our customers switched their old CRTs with new shiny TFTs the room temperature went down, considerably. Ok, this slightly increased the heating bill in winter but the savings on cooling the office space during summer were higher.

    Replacing the Desktop PCs with Thin Clients will have roughly the same effect.

    “But wait. Replacing all of the computers with thin clients will not be accepted by the users!”

    Are you sure?

    Companies are not a democracy. If the users have no choice they will accept the new thin client strategy, if they want to keep their jobs.

    On the other hand companies try to squeeze out the last cent out of every investment before turning to new one.

    So, what will it cost?

    Now, that depends largely on the software you use, how much you spend on your average corporate desktop, and which kind of solution you’re targeting.

    Yes, there’s more than one way to thin down your computers.

    The classic approach would be a terminal server solution. All applications and the desktops run on a shared terminal server which can host roughly up to 80 users (depends on applications and server size).

    Or you could go with the virtualization hype and provide a virtual machine for every user which more or less behaves like the common desktop computer, only it’s running on a large server farm.

    Alternatively you also could use application streaming where preconfigured packages of application are synchronized with the application server but running locally. This could be an intermediate solution for existing desktops or used for traveling users on a notebook.

    Ultimately, if you need to provide high performance graphic applications like CAD or DTP tools you now can bind a whole blade server to a thin client. This is the most costly solution, but enables your users to work on his/her machine from any of these thin clients while the system itself is secured in the datacenter.

    This is the place where you start saving money. All of the solutions above focus on concentrating the user environment in a place where it can be managed most efficiently: Your datacenter.

    You won’t need highly trained IT personnel spread across the globe to maintain your desktop computers. I’m not saying you should lay them off, but they can remotely manage the centralized infrastructure. Think of it like a 24h server system administration plan with the people you already have. These people can keep your server running, so the users can work with the remote desktop environment.

    If a thin client breaks, someone just has to plug in a new one and all is set. If you have to shut down one server the sessions can be relocated to another, which reduces critical downtimes.

    Since all of the computing power is concentrated in a single place you won’t need local servers in every branch office, outlet, site, whateveryoucallit. This will save money for the hardware, system maintenance, and will simplify your backup strategy.

    And the best part: All data, documents, files, etc are stored in a (hopefully) safe place with periodic backups. Just imagine never having to search a broken harddrive for that crucially important Excel file, which should have been stored on the department file share in the first place, again.

     
c
compose new post
j
next post/next comment
k
previous post/previous comment
r
reply
e
edit
o
show/hide comments
t
go to top
esc
cancel