The Storage Explosion Is Here

There's a dizzying selection of storage solutions out there. Which one is right for you? The answer will often be "more than one."

Suppliers of storage hardware and software are presenting what appears to be a huge list of options. Which ones are best isn't always clear. Furthermore, it's not clear if there's a single option that addresses an enterprise's needs. All of the suppliers, however, promise that their solution is the best, the most cost-effective and makes the best possible use of <insert the name of your favorite storage technology here>.

Marketecture Abounds
Like just about every other area of the IT market, suppliers of storage are always doing their level best to one-up their competitors, seeking ways to out-gun the others in the areas of storage performance, reliability, scalability and overall cost.

Although there's a great deal of noise in the market, a few things are clear. For one thing, each of the suppliers believes it and it alone is uniquely qualified to be the only source of storage technology. Also clear is the fact that there are a number of different types of technology from which enterprises may choose, and an overwhelming set of combinations for how this technology may be used together. Finding the right fit largely depends on what question the enterprise is asking.

Increasingly, storage services are being offered by suppliers of managed services, colocation or cloud services.

Why So Confusing?
There are a number of different approaches to storing applications and data, and each is useful in the right place and at the right time. Some approaches require the storage media be directly connected to clients and servers, while others attach storage devices to a storage appliance or a storage server.

If the storage device is directly connected to the client or server, there are many different storage interconnects from which to choose. Each offers a different mix of price and performance, and can control what media options are available.

As with direct-connect approaches, there are several different storage interconnects in use in today's datacenters when the storage devices are attached to a storage server or appliance. These servers may be connected to computing systems using a general-purpose LAN or special-­purpose SAN.

The industry is also seeing the increasing use of system memory being used as a special form of storage for compu­tationally intensive, extremely high-performance applications. Sometimes suppliers call this "distributed cache" or "in-memory database."

To add to the confusion, cloud services providers have begun to offer an array of new Storage-as-a-Service products. They're trying to convince enterprises that it's better, less complex and less costly to use those services rather than purchase, install and operate their own storage.

Different Technology for Different Needs
The industry has used various types of technology over the years, including:

  • Tape. Different suppliers have offered paper tape, cassette tape, and reel-to-reel tape products. Several suppliers have offered direct access tape devices that could replace rotating media for large-scale storage applications.
  • Rotating media. Different suppliers have offered rotating drums and a whole herd of different types of disk storage. While most of these were based on magnetic recording, some were based on optical recording technology.
  • Solid state. Although solid-state storage has been available for decades, and its access times and throughput made it extremely desirable, the cost was prohibitive for most applications. Recently, however, the introduction of new technologies has resulted in the rapid adoption of flash memory.

As suppliers seek ways to offer flexible and inexpensive storage options, the market is seeing the emergence of distributed cache solutions using the system memory of low-cost, industry-standard servers, blades or distributed NoSQL database solutions using server clusters.

You've Got Options
Like most areas of IT, there are many different types of storage technology, and each has the ability to serve a different set of needs. If the enterprise carefully reviews its application portfolio, it will soon become clear that each application has a different storage profile.

Some applications require the storage and retrieval of huge amounts of data, and longer access times are acceptable. Other applications access huge amounts of data, but the access time must be kept to a minimum. Still, others require immediate access to data and any delay is unacceptable. Finding the right solution necessitates understanding your environment's unique requirements. Most fall into one of these categories:

  • Long-term storage. The requirements for this type of storage usually include massive capacity and low cost per megabyte or gigabyte. Applications using this data typically are batch or analytical jobs.
  • Medium-term storage. The requirements here lean more toward finding a good balance between performance and cost. This often means storing applications and data for remote or VDI desktops, servers or even handheld applications. Enterprises are often willing to compromise on storage performance to reduce overall cost.
  • Short-term storage. The requirements for transactional or business intelligence applications often include the need for very short access time and medium levels of throughput. Enterprises are often willing to compromise on cost to obtain performance.
  • Storage for high performance or technical processing. The requirements for this type of workload often include extreme storage performance (seek performance and access time or latency), extreme needs for data throughput, and huge volumes of data. Shared cache, clustered NoSQL databases and in-memory databases are often used to address these requirements.
  • Flexible storage. When the enterprise faces ever-changing, dynamic requirements, it will often turn to some from of distributed or hybrid storage. A local cache made of high-performance storage is deployed to improve the performance of off-site or cloud storage. The off-site storage may be at another enterprise-owned site, at a site managed by a managed services supplier or in the datacenter of a cloud services provider.

How Are Suppliers Addressing These Requirements?
Although each supplier is addressing these storage requirements differently, there are some common threads:

  • Slow, but reliable storage for huge amounts of data. A form of tape or optical technology often satisfies these requirements. Some cloud services providers are suggesting their Storage-as-a-Service offerings might be a replacement for this type of storage. It's not clear what type of storage they're actually using to address this need. Low-cost, low-performance rotating disk storage is very likely part of the cloud services provider's offering.
  • Fairly fast and inexpensive disks. These can be used for client-side applications or server-side applications for small to midsize businesses.
  • High-speed, expensive disks. These target server-side applications that need both a larger amount of storage and high levels of performance.
  • Flash and other forms of solid-state storage. They're packaged as storage devices that address the needs of applications requiring very low access times or high levels of throughput. Typically these devices are much faster and more expensive than traditional disks, and offer less capacity.
  • Internal solid-state memory. It's packaged by storage virtualization technology so that it appears to be a storage device.

What's the Best for Me?
The enterprise must take the time to survey its portfolio of workloads to learn the answers to the following questions:

  • How much is the enterprise willing to pay for storage? High-performance storage typically is expensive.
  • Does the enterprise really need massive storage capacity? There are many ways to address this type of need. The best answer usually is a compromise between cost, performance and storage capacity.
  • Is the enterprise willing to use off-site storage? Enterprises in regulated environments may only be able to use on-site, locally controlled storage for regulated applications. Collaborative applications, e-mail and other non-regulated applications might be candidates for off-site cloud storage.

The Golden IT Rule
Most enterprises, by necessity, rely upon many different types of storage. This is partially due to the different needs of each application, and also due to the application's age. Older applications, for instance, are likely to be using older types of technology. Enterprises often follow the golden rule of IT, "If it's not broken, don't fix it," when dealing with these applications.

The more enterprise decision makers know about their applications and their requirements, the easier it is for them to select the right storage technology, storage location and determine if cloud storage is even a reasonable option.

(Editor's Note: This article is part of a special edition of Virtualization Review Magazine devoted to the changing datacenter. Get the entire issue here.)

About the Author

Daniel Kusnetzky, a reformed software engineer and product manager, founded Kusnetzky Group LLC in 2006. He's literally written the book on virtualization and often comments on cloud computing, mobility and systems software. He has been a business unit manager at a hardware company and head of corporate marketing and strategy at a software company.


Subscribe on YouTube