Simon Bisson 

Information overload

Simon Bisson explains how we can deal with the billions of gigabytes of data created around the world
  
  


It took 300,000 years for the human race to accumulate 12 exabytes of information. That's 12 billion gigabytes. Now the University of California in Berkeley estimates that we're creating 250MB of data each year for each man, woman and child on the planet - another exabyte or so - every year. It's no wonder businesses are looking for technologies to help them manage and store the information flowing in and out of their computer systems.

Storage systems have always been a critical part of any business IT infrastructure, whether it is a server's own direct access storage hard disks, or shared network storage.

Network storage allows organisations to consolidate data, simplifying backups and giving the businesses the ability to develop common data models as part of the drive towards more flexible business processes. It's also a solution that can scale easily, so you can add more storage hardware when it is needed. Vendors such as EMC offer upgrade paths that take a storage network from a workgroup system to a service that can handle storage for a whole enterprise as you grow.

Two technologies are at the heart of network storage strategies: storage area networks (SANs) and network attached storage (NAS). Often seen as very different solutions, targeting very different markets, they are both part of a trend that is leading organisations to centralised storage systems, accessed by a mix of high-speed local networks and slower wide-area networks.

Traditionally, SANs are good at handling data a block at a time rather than a file at a time, so they're ideal for databases. High-speed fibre channel connections give servers quick access to large amounts of data, and specific storage resources can be allocated to specific applications. With fibre-channel switches and management tools, SANs become part of flexible arrays of storage hardware known as storage fabrics. A storage fabric can manage storage for several different applications, as well as linking backup systems to the storage network. SANs are usually an expensive solution, but lower-cost SAN implementations like Apple's Xserve RAID bring the price low enough for smaller businesses.

NAS solutions let you access files over Ethernet connections: think of them as dedicated, pre-installed file servers. They're quick and simple to set up: you can have a large NAS system like a Network Appliance Filer up and running within 30 minutes of installation. Smaller NAS devices like the Linux-based GigaDrive from SME-friendly Linksys, with its web-based user interface, can be ready for use as soon as you switch it on and name it. The access method usually depends on the operating system the appliance is intended for: NAS file systems for Unix use NFS, and CIFS for Windows uses CIFS (Common Internet File System). Some systems offer file access over HTTP.

As network storage technologies mature, and management tools become more powerful, businesses can begin to take advantage of storage virtualisation. Like virtual servers, storage virtualisation lets you treat storage as a resource pool you can allocate on demand. In the past, SAN architectures have had to be based around single-vendor solutions. That could change with the storage management interface specification that the storage network industry association is pushing as a basis for interoperability. This uses existing management standards like CIM (the common information model) and WBEM (web-based enterprise management), along with web services.

It is important to check whether your applications have been certified for use with NAS or SAN systems before investing in a solution. Some applications, like Microsoft's Exchange 2000 messaging tool, are only certified for use with block-level interfaces. While that would normally rule out NAS solutions, newer NAS technologies do offer block-level access using technologies such as iSCSI.

Network storage means more than just magnetic disks. Backup technologies and servers are an important part of any storage infrastructure. One option is to use a virtual tape appliance like Fujitsu Siemen's CentricStor. Virtual tape appliances use disk caches to manage access to tape storage, ensuring that data is backed up on demand, even when tape drives are in use. Similar solutions can make network storage part of a hierarchical storage system, holding indexes and caching information that is normally kept in long-term storage, on tape drives or in optical storage.

Network storage has an important role to play in business continuity. Consolidated storage is easier to replicate to backup sites than a myriad hard drives on small servers and PCs. Technologies like iSCSI and FCIP can help connect data centres, and allow remote offices to access the same data sources as the rest of the business. The resulting distributed infrastructure should make it easier to unlock the information inside an organisation.

Implementing network storage infrastructures isn't something to do overnight. The infrastructure requirements for a SAN fabric can be expensive, and need a particular set of administration and management skills. If you don't have those skills in house, you can look at a service like that offered by Cap Gemini, Ernst and Young and EMC, which gives you a managed storage network infrastructure for a fixed per-gigabyte cost.

So what does the future offer? One obvious trend is the merging of SAN and NAS technologies. Already some vendors are producing SAN equipment that also offers NAS interfaces, while others are selling hardware that acts as a gateway between existing SAN systems and the wider Internet Protocol world.

Open, interoperable network storage network can be a key tool to help build a resilient, agile IT infrastructure for businesses of all sizes, from the smallest SME to the largest Fortune 100 enterprises.

 

Leave a Comment

Required fields are marked *

*

*