|DS6000 - The specs|
They might not be called sharks, but IBM Corp.'s two new arrays will make the storage waters a bit more perilous for its competitors. Big Blue has eschewed the shark moniker for its new TotalStorage DS6000 and DS8000 arrays; but make no mistake, their underlying architecture and scalability are definitely predatory.
The two new systems, while different in many respects, have enough software in common to provide a powerful one-two tiered storage punch. In developing successors to its ESS storage line, IBM borrowed heavily from its experience building high-performance servers. Both new storage boxes are powered by proven server processors--the PowerPC 750GXfor the midrange TotalStorage DS6000, and the Power5 processor in the enterprise-class TotalStorage DS8000. Key benefits include:
- Shared software and compatibility with current Sharks to facilitate data movement among tiers
- Amodular approach to scalability
- The DS6000's small footprint
- Processor-based partitioning in the DS8000
Acursory look at the spec sheets for these systems won't blow you away. Storage capacities, connectivity and other tale-of-the-tape numbers are healthy, but can't compare to the stratospheric stats of Hitachi Data Systems' (HDS) new TagmaStore boxes. But IBM's new systems are the first leg of a product roadmap that leverages new chips, cache algorithms and advanced server partitioning to deliver enhanced scalability for capacity, performance and manageability. "The new box is going to take them forward because they're using the new technology chips," says Joe Furmanski, technical project director at the University of Pittsburgh Medical Center (UPMC), "and that's going to make an order of magnitude difference."
IBM touts the DS6000 as offering enterprise-level performance at a midrange price. The modular design allows the array to scale up to 224 disks and a maximum capacity of 67.2TB. It has eight host and eight storage ports, and can attach to open systems and mainframe hosts. The ability to connect to mainframes is unusual for a system positioned as a midrange product. The new enterprise entry, the DS8000, can house up to 192TB of storage. Initially, two models will be offered: the DS8100 featuring a dual two-way processor configuration and a maximum capacity of 115TB, and the DS8300 with a dual fourway configuration and support for up to 192TB of storage.
Leveraging server technology
The DS6000 and DS8000 are both built around IBM server technology, and while the processors and controllers are different, both systems are highly modular and are easy to scale upward. They will be available in December, with full production in the first quarter of 2005.
IBM also implemented its LightPath Diagnostics system on both arrays. LightPath appeared several years ago on IBM's xSeries servers; it's an onboard diagnostic system with a series of LED indicators that are used to monitor performance and predict component failures. This marks the first time IBM has used LightPath in a storage system. The DS8000 takes diagnostics a step further by exploiting the Power5 chip's self-healing capabilities that can make adjustments to avoid operational failures.
Management of the two arrays is consistent because of the high degree of software commonality. The code base that runs the systems is nearly identical, with the DS6000 using about 97% of the code that powers the DS8000, providing easy interoperability between the two new models. Interoperability extends backwards, to older ESS arrays that new DS units will coexist with and ultimately replace. "About 75% of the code that runs on the DS8000 is the same as that on the ESS800 today," says Mike Hartung, an IBM fellow and an architect of the DS systems.
By maintaining considerable code compatibility with each other and ESS boxes, DS models will use IBM datamoving tools, such as XRC, FlashCopy, Global Copy and Metro/Global Mirror. This allows for the copying of data among any mix of DS6000s, DS8000s and ESS models.
Nancy Hurley, senior analyst at the Enterprise Strategy Group (ESG), says this "very cohesive software family allows you to smoothly integrate between the two, and to use them in a tiered storage environment."
A new caching algorithm developed at IBM's Almaden Research Center is used in both boxes. Adaptive Replacement Cache (ARC) builds on the Least Recently Used (LRU) cache algorithm prevalent in most storage systems and used for years in server architectures. ARC effectively melds LRU with another caching technique, Least Frequently Used (LFU), to dynamically balance the "recently" and "frequently" criteria to improve cache hit ratios. Hartung says the ARC algorithm is so effective that under some test workloads it made the cache appear to be twice its actual size. ARC also does a better job of cache management, especially in segregating random and sequential reads.
"The ARC cache management notices a sequential process and keeps it from flooding all the memory," says Bob Venable, manager of enterprise systems at BlueCross BlueShield (BCBS) of Tennessee in Chattanooga.
Another shared trait is host support. Both systems can attach to mainframes or open-systems servers. For the high-end DS8000, mainframe support is to be expected, but for the DS6000, mainframe support in an array positioned as a midrange box is a bit unusual, although EMC's DMX800 offers this support, too.
|DS6000 and DS8000 - Pricing|
The Power of the DS8000
IBM hasn't just swiped some CPUs and a few spare server parts to tweak the DS boxes. The DS8000 incorporates two full pSeries P5 servers, featuring dual Power5 processors that offer high-end performance, larger caches, improved bandwidth and advanced features such as partitioning. The DS8000 scales linearly with the addition of processor pairs. Rich Lechner, IBM's VPof storage, calls this type of scalability a "new infinity" that will eventually allow the DS8000 togrow to more than 1PB of storage.
The partitioning capabilities of the Power5 processor are a key part of the DS8000 picture. Logical partitioning (LPAR) is built into the hardware and firmware, and can effectively control and allocate all system resources. Initially, two partitions are possible on the four-way model. Partitioning isn't just a matter of allocating storage and cache; each partition gets its own dedicated processors, cache, storage and adapter slots. With HDS' new TagmaStore, virtual machines can be similarly created, but with TagmaStore, "you can't do anything as far as allocating the processing power," says Hurley. Comparing IBM's and HDS' implementations, Mike Kahn, managing director of The Clipper Group, says, "The customer is presented with two different approaches, with different scaling and economics."
In the initial configuration that allows two partitions, there's a 50-50 split of processing power where each partition also has a dedicated bus, the RIO2--or Remote Input/Output 2--yet another element in the Power5 server architecture. Soon, IBM will offer the capability to partition on a 75-25 split, in which each side gets a dedicated RIO2 bus. "At this stage of the technology, we can't share the RIO interconnect between LPARs," notes Hartung. "In a future release, we will." He indicates that in future revisions, partitioning can be a dynamic process in which policies control resource allocation.
In IBM's server world, partitioning can be done on an even more granular level, with Power5 configurations divvied up into chunks as small as 10%. While redundancy and high-availability demands make it unlikely DS8000s could be sliced into such small segments, IBM has indicated that additional levels of partitioning are in the offing. When that happens, "you're going to be able to start tiering performance within the box," says ESG's Hurley.
For example, a single DS8000 can be partitioned to effectively operate as two separate arrays--to isolate a production environment from development storage, for example. Or one partition could be dedicated to open systems, while the other handles mainframe storage.
UPMC's Furmanski likes that IBM is applying some of its newest server technology to the DS8000 and says, "we've been waiting for it for quite some time." He sees possibilities for the DS8000's partitioning capability in his shop, which has about 140TB of storage on a mix of Hewlett-Packard (HP) Co. boxes and a few IBM Shark F20s. "We have to control and segregate the performance we get for our production environments vs. our development environments," says Furmanski, who envisions using LPARs to ensure the availability of his firm's electronic health-record system.
But slicing an array in half isn't the only application for LPARs. LPARs are independent but can interoperate, so a partition could run a storage management program or some other application. The benefit would be that the two partitions--the application and the storage itself--are closely coupled and would communicate at memory speed rather than across the network. "We're looking for applications that are going to benefit from co-location with the data," says BCBS' Venable, "and TSM and SVC are two primary workloads that we think will be good."
Furmanski is also intrigued by the prospect of running an application in a partition, notably SVC. "When you link it up with their storage area network (SAN) virtualization software and host virtualization software," he says, "they have an orchestrated environment of virtualization."
Another application that would take advantage of the DS8000's partitioning would involve running a NAS head in one partition, making it possible to have file and block systems in a single array, a capability NetApp has offered for some time. "The model is that of storage-based middleware," says IBM's Hartung, meaning the NAS file system runs within the array, rather than on an appliance.
Partitioning and the ability to allocate essentially all system resources puts the DS8000 in a unique position. The closest in concept is HP's StorageWorks Grid design in which "cells" containing processors, cache and storage have their own "personalities" (such as a NAS file or block file system), and can be added to the grid incrementally. "HP and IBM are the only ones talking about not just partitioning off capacity, but being able to partition off the computing power," notes Hurley.
|DS8000 - The specs|
DS6000 up close
"Midrange" is probably a misnomer as applied to the DS6000; it's difficult to hang that label on a storage system that could do duty in small- or medium-sized businesses as well as enterprise departments, remote sites or even data centers. "It has a lot of potential, both as a mainframe connect or an open-systems product--or both if that's what a modest-sized enterprise would need," says Kahn. The Clipper Group, he adds, has christened the DS6000 as an "upper-class storage product" because it represents a second tier within the enterprise class.
Size is everything with the DS6000, and scalability is the name of the game. In its most modest configurations, it has a footprint that occupies a 3U slot in a standard 19-inch rack, and starts out with about half a terabyte of storage for $97,000. Drawers--3U enclosures that house 16 disks--can be added to the rack to boost capacity to 67.2TB. "The DS6000 is an enterprise-class storage device that's one-tenth the weight of an equivalent offering from EMC, the DMX800," says IBM's Lechner.
The modular design should simplify upgrades, and take some of the complexity out of operation and maintenance. The system is "designed to be installed in less than an hour by a non-technical professional," adds Lechner. Maintenance is largely in the hands of users, relying on PathLight diagnostics and self-healing features. DS Storage Manager is another key operation and maintenance component. The software ships with the DS6000 and DS8000, and has been upgraded with a new GUI and a series of wizards. In addition to monitoring and management functions, the program features a configuration tool that lets administrators simulate configurations, save them and apply them to the arrays later.
The DS6000 doesn't support partitioning, but IBM says SATA drive support will be added in the future, so the DS6000 can house a mix of SATA and SCSI disks. Given its scalability and enterprise-class performance, the DS6000 will undoubtedly find its way into data centers for critical or somewhat-less-than-critical applications, and as part of a tiered system. With the addition of low-cost disks, tiering alternatives will be even more diverse. "It becomes another tier to which it might be hard to compare existing products immediately," says Kahn.
Contrasted to HDS' high-end offerings, "IBM is aggressively going downstream," adds Kahn. He suggests HDS will do some repositioning to gain traction.
"To me, the DS6000 seems like more of a category killer than the 8000 series," says Steven Berg, VP and senior analyst at Punk, Ziegel & Company, a New York City investment banking firm. He says he was surprised the DS6000 doesn't support iSCSI,which he considers a promising technology for companies with decentralized operations. Berg notes that EMC's DMX line supports iSCSI and "was really surprised that IBM and Hitachi didn't follow."
DS and Sharks
The two DS offerings are so broad in scope and scalability that they overlap in some respects, notably storage capacity. And the fact that they share nearly the same code base while maintaining compatibility with existing Sharks, should not only ease their management, but encourage their coexistence as well.
A basic scenario might feature DS8000s in the corporate data center, with DS6000s installed at remote locations. With the same copy services running on both DS boxes and the legacy Sharks, it should be relatively easy to move data among all the machines to create a tiered, data management system. "When we get into disk mirroring across locations," notes BCBS' Venable, "we'll look at the 6000."
Common management and software across the lines will encourage the installation of both DS systems in the same environment. "It's a very cohesive software family that allows you to smoothly integrate between the two and use them in a tiered storage environment," says ESG's Hurley, adding that there wouldn't be a "need to have a number of disparate software systems."
With HDS and IBM making significant storage announcements, one naturally listens for EMC's shoe to be the third to drop. While EMC hasn't hinted at a direct response to these announcements, its upcoming storage router product could compete on some levels--but likely without the breadth of either IBM's or HDS' new offerings. Hurley isn't confident EMC can respond quid pro quo: "I don't believe EMC has anything that would enable them to deliver something similar to what IBM announced," she says.
Hitachi may feel some of the wind leaving its sails, too, with IBM's rollout coming a mere month after its own. While Hitachi's TagmaStore focuses more on virtualization, capacity and performance, IBM seems to be taking a different tack with the DS architectures. While scalability appears to be IBM's main thrust, Hurley says "they definitely have the speeds and feeds," but what IBM has accomplished architecturally is more important.
At The Clipper Group, Kahn's assessment is similar, although he points out that the DS8000 "doesn't solve all the virtualization problems the Hitachi will by putting all the other arrays behind it." But he adds that IBM can counter Hitachi's virtualization with its SVC.
While the merits of each company's approach can certainly be debated, the good news for storage managers is that the new offerings should provide clear-cut choices that address specific requirements.