Not just another iSCSI article: Microsoft Windows dynamic storage capability + iSCSI wall change direct attached storage forever.

By now, you've undoubtedly heard of the Internet Small Computer System Interface (iSCSI), the emerging protocol standard that allows SCSI devices to interconnect over TCP/IP And when you think of iSCSI, your first reaction is an evolutionary technology that provides Ethernet ubiquity to Storage Area Networks (SANs) mid block capability to Network Attached Storage (NAS), both of which are true. What's not much talked about, however, is the revolutionary impact that deploying iSCSI will have on Gigabit Ethernet and how this will forever change the Direct Attached Storage (DAS) landscape.

Dramatic improvement in storage technology, coupled with the explosive growth of the Internet, has fundamentally changed the way in which storage is used. Only a decade ago we grappled with files larger than 1.44 megabytes. Today we carry hundreds to thousands of megabytes on our key chains and in our portable music players. On the consumer side, graphics and audio files (.jpg, .mpg, .mp3) consume hundreds of gigabytes of our personal storage, while our office environments presentations and product documents (.ppt, .doc, .pdf) and the ever-dreaded e-mail files (.pst) do the same. Not only are these files bigger than ever before but also more daunting is the expectation that the data will always and forever be available at a moment's notice.

The best evidence of a shift in storage usage is the transition of storage's role, from tactical to strategic, in the minds of CIOs and IT managers everywhere. This is not to say that storage capacity wasn't an issue a decade ago, but compared to the complexity of storage management today, adding another disk to a server seems almost elementary. Quantitatively, since the peak of the Internet boom in January 2000, the storage networking market has grown from a $1.3 billion dollar industry to an estimated $4.7 billion dollar industry by January 2003. From 2000 through 2003, the total storage market is estimated at over $25 billion dollars.

Direct Attached Storage: What is it Good For?.

As the name implies, DAS is storage that is directly connected to a computer. Typically a hard disk or JBOD (just a bunch of disks) connected via a SCSI or IDE bus that moves data as blocks via low-level commands. In terms of quantity, DAS is still the most widely deployed means of adding storage to a network; however, fundamental limitations of the technology preclude its ability to address the ever-expanding user demands on storage.

* Physical limitations

* Lack of scalability

* Cost/complexity of management

On a practical level the major drawback of DAS is that, unlike NAS, DAS must be accessed through the server on which it is attached, not only using valuable resources of this "host" machine but also impeding direct assess to the storage, as you would get in a SAN environment.

Needless to say, in the minds of IT professionals, DAS is the ugly stepchild to NAS and SAN. As good fortune would have it, like Cinderella, DAS has a fairy godmother. The combination of iSCSI, transport offload, and Dynamic Storage (found in Microsoft Windows XP and 2000) will transform DAS into a robust storage networking solution.

Ethernet: Hold on, Not so Fast ...

In theory, Gigabit Ethernet provides throughput equivalent to that of a 32-bit 66MHz bus. Unfortunately, due to an unexpected computational limitation these levels are rarely reached and. if achieved at all, come at great cost to overall system performance. Ethernet relies on the Transmission Control Protocol and the Internet Protocol (TCP/IP) to ensure data integrity and to provide its robustness. TCP/IP's success in scaling with the ever-expanding Internet is a testament to its fortitude. Unfortunately, this capability comes at a cost: computational load. Industry experts agree that it requires roughly IMHz of modern processor operation for every 1 megabit of TCP/IP data that is being moved. Extrapolated for high-speed networking, this means that in order to achieve full-duplex Gigabit Ethernet a 2GHz processor would be fully consumed processing the TCP/IP protocol stack, leaving no headroom for applications or other functions (see Figure 1).

[FIGURE 1 OMITTED]

To address this limitation, an industry-wide initiative developed the transport offload computing model. It differs from legacy computing architectures in that it shifts responsibility for protocol processing from the host to the network interface. The network interface then becomes responsible for performing some or all of the functions corresponding to layers one through five of the OSI reference model (see Figure 2).

[FIGURE 2 OMITTED]

To support this new computing model, various vendors are producing transport offload engines (TOEs), which take the form of dedicated semiconductors that perform high-speed TCP/IP computation. Additionally, since some of these solutions are rendered in dedicated logic they provide the additional benefit of providing low latency throughput, effectively transforming Ethernet from a networking connection technology into a long haul high-speed interconnect bus.

iSCSI

iSCSI is a soon-to-be-ratified IETF (Internet Engineering Task Force) standard for encapsulating SCSI commands over TCP/IP. This allows for I/O block data transport over Ethernet networks in the same way that Fibre Channel provides the functionality for SANs. The beauty of iSCSI is that since it works on top of pervasive Ethernet networks, it can easily be added to existing network installations without requiring any new infrastructure.

For example, let's say that you wanted to add disk storage to your exchange server. In the old days (pre-iSCSI) the two most common choices were to take the machine down or have a SAN service your exchange needs. In the latter case, if you are fortunate enough to already have a SAN, great! If not, this is an expensive solution for a fairly contained problem. With iSCSI, you can assign storage anywhere from your network and make it appear to an Exchange server that it is a local physical drive, mail stays up, storage is ultimately scalable, and you don't have to touch your existing network infrastructure.

One Adapter Please

Unlike Fibre Channel, iSCSI also provides system-level elegance since it does not require a separate network adapter. Specifically, because the protocol is built on top of Ethernet, the same adapter can service both networking and storage traffic simultaneously. In addition, since the iSCSI standard calls for Ipsec-based security, the card will likely bring that capability to your system as well. Credit Suisse First Boston predicts that total Ethernet deployment will ramp from 2.6 million units in 2002 to 20.2 million units by 2006 with a CAGR of 67.2% estimated from 2000 through 2006, as 1T managers begin migrating their 10/100 Network Interface Cards (NICs) to 10/100/1000 NICs. By choosing an iSCSI-capable Gigabit Ethernet adapter, they will gain the benefits of much higher throughput along with the added benefit of iSCSI storage networking all in a single adapter.

There has been much debate on the need for gigabit throughput (vs. 100 megabit), especially as it relates down to the desktop. With the ever-expanding needs of user storage unlikely to be abated, iSCSI may very well be the killer application for Gigabit Storage.

Dynamic Storage Virtnalizes DAS

iSCSI and TOE technologically allow DAS to become immediately networkable, but that alone is insufficient to make it truly usable and, more importantly, widely deployable. As mentioned before, the use of storage has changed. A decade ago, drive partitions made sense; today, a user's Inbox and his MyDocuments folder are the most common means of segregating data. In the latest versions of its operations systems, Microsoft has introduced a new capability called Dynamic Storage. Found in Windows XP Professional and Windows 2000, Dynamic Storage capability allows a disk to be initialized as a dynamic disk, enabling it to contain dynamic, simple, spanned, striped, mirrored, and RAID-5 volumes.

From a user's perspective, the beauty of this solution is that Dynamic Storage allows one's storage to be infinitely scalable, whether fur home or work clearly matching the use pattern of today's computer user. From an 1T manager's perspective, the elegance is that by configuring all DAS, including desktop storage, as Dynamic Storage it all becomes immediately manageable and scalable.

In the same way that Windows 3.1 drove the mass proliferation of hardware acceleration in graphics cards, Dynamic Storage will drive the mass proliferation of iSCSI and TOE in Gigabit Ethernet adapters.

Mature Markets, Enabling Technology, Mass Proliferation

Previously, I bad the good fortune of working for the late Dr. Mark Weiser, who is regarded as the father of ubiquitous computing. Mark pioneered the idea that, for a technology to become truly pervasive, it had to be invisible to the user, and for it to be invisible, three vectors must intersect: (1) the use must be completely and intuitively understood, (2) enabling technology must be designed elegantly such that it performs the function without introducing new impediments, and finally, (3) collective solutions must be widely available such that access is not an impediment to its proliferation.

Users' ever-expanding needs fur scalable storage solutions are clear, well understood, and urgent; iSCSI and Dynamic storage exist and are architected in such a way that they are easily deployable; Microsoft Windows is the most widely used operating system in the world and iSCSI and TOE capability is being built into industry standard Gigabit Ethernet cards.

Looks like the DAS industry is about to change in a big way.

www.iready.com

Ryo Koyama is founder and CEO of iReady Corporation (Santa Clara, CA)