天极Yesky
  • 笔记本电脑
    笔记本
  • 台式电脑
    台式机
  • 手机
    手机
  • 电脑硬件DIY
    DIY硬件
  • CPU
    主板
    音箱
  • 硬盘
    显卡
    键鼠
  • 内存光驱
    显示器
    机箱电源

  • 数码相机DC
    数码相机
  • MP3播放器
    MP3/MP4
  • 数码摄像机DV
    摄像机
  • 电脑外设
    外设
  • 网络
    网络
  • 服务器
    服务器
  • 数字家庭
    数字家庭
  • 群乐
    群乐
  • 产品报价 行情 商家 新闻 评测 | 软件 设计 网页 开发 安全 论坛 E时代 游戏 图片 壁纸 下载 网摘 博客 索尼专区 Vista 科技奥运
    天极网
    Bell Canada answers desperate storage call
    作者: Mark Lewis
    出处:
    责任编辑:
    [ 2004-06-17 18:23 ]


    Editor's note: This is the last in a series of three Storage Innovator articles on companies honored recently by SearchStorage.com. On Sept. 17 at the Storage Decisions conference in Chicago, Bell Canada was given the Fall 2002 Storage Innovator award in the category of "Return on investment." Judging for the ROI category focused on how a particular storage operation most dramatically affected a company's bottom line.

    Comedian Jerry Seinfeld always wondered why humans continue to swim and surf in the ocean when the waves just keep spitting them back out. If you think that's hard, just imagine trying to push the ocean back with a spoon. This is an apt description for how Canada's leading telecommunications company, Bell Canada, felt two years ago when it tried tackling the challenges brought by its own storage data growth.

    With 38,000 employees, Bell Canada needed to account for a huge amount of personal data stored by workers on the company's servers. It was a significant network management challenge.

    According to Chris White, manager of corporate security at the Montreal-based company, data growth was so intense in the company's Windows NT-based network that conducting full backups each week, even during weekends, was almost impossible. By the middle of 2001, Bell's network consisted of 11 single-node file servers with 1 terabyte of space and 32 cluster servers with between 400G bytes and 800G bytes per cluster.

    "We were running projects manually, load balancing and data migration," White said. "We could get the servers down to 80 percent full, but after a week or a month later they were 99 percent full again."

    With the flood of data and no clear relief in sight, management stepped in and told White's team to "fix the problem."

    As a trial, the storage team looked to its London operation as a test bed for a series of new quotas it was going to impose on employees. The company sent out a series of e-mails telling employees and later reminding them to clean out nonessential data from the company servers. But apparently not everyone complied with the quotas they were given, and when the storage team moved to reduce the storage allotted to the office, the system crashed. The test failed, but it provided an excellent learning experience.

    When they decided to roll out more strict storage quotas throughout other Bell Canada operations, the team, under the guidance of the company's chief information officer, had a better feel for what the company could handle.

    The imposed quotas limited the amount of file storage space each of the employees was allowed on the company's servers. Employees were also asked to remove files that were not business-critical, such as duplicate media files, MP3s and personal files.

    "This forced a lot of people to clean," White said. "There was a lot of outdated archival data that &users didn't even know what it was. It was just as easy for us to move it to tape backup or CD and get it off the servers. That created free space."

    White also knew where the heavy users were located, and he closely monitored their capacity usage. If a user went over quota for a business reason, White was able to allocate needed space quickly. But it wasn't just handed out. Users had to state an immediate and real business objective in order to get more space.

    Aside from setting quotas, Bell Canada was able to track its storage efficiently with a storage resource management (SRM) product. Using Precise's (formally Wquinn's) StorageCentral SRM solution, the company was able to push down quotas and get an ongoing look at its storage levels.

    As a result, 3T Bytes of space were immediately recaptured. Having the ability to see the whole topology of the network also allowed the team to reallocate space accordingly. This saved the company from having to go out and spend even more on hardware, as well as turn off legacy servers with the space it found. Through these initiatives, Bell Canada achieved savings of about U.S. $500,000 during the first year it used the tool.

    The problem of server availability was also cured. At times, data logjams would have application servers running at only 85 percent, but with StorageCentral and quota enforcement, server availability was boosted to 99 percent. Judging by Gartner Inc.'s estimate that one hour of downtime costs $18,000, the company saved an additional $22 million dollars in the first year by boosting availability.

    White's team had "fixed the problem" and saved the company a bundle of money in the process.

    "We didn't want to affect business, so we had to be flexible and available to re-adjust quotas right away," White said. "You can't be too hard-nosed about it. Honestly, I was surprised we didn't get more flack."

    Click for more information on Bell Canada.

    Go to Precise's Web site for more information on StorageCentral SRM.

    More on this topic:

    笔名:
    请您注意:

     遵守国家有关法律、法规,尊重网上道德,承担一切因您的行为而直接或间接引起的法律责任。

     天极网拥有管理笔名和留言的一切权利。
    相关内容