Thursday, March 25, 2010

Gartner on FCoE: The rebuttal

March 25, 2010 – Earlier this week I blogged about a Gartner press release and report that purported to point out the flaws behind the pro-FCoE arguments. Essentially, Gartner analyst Joe Skorupa argues that Fibre Channel over Ethernet may not, in fact, reduce the number of switches and ports, nor the power and cooling requirements, in a converged network. He also argues that a converged network could be more difficult to manage than two separate networks (Ethernet and Fibre Channel).

For the full press release and a link to the ($195) report, see my previous post, “Gartner exposes FCoE ‘myths.’”

Predictably, those views set off some controversial fireworks, in part because (for better or for worse) Gartner holds a lot of sway in the IT world, which is where the FCoE battle will eventually be fought. For the best rebuttal to Gartner’s arguments that I’ve seen so far, see J Michel Metz’s “Gartner on FCoE: Whoa There, Sparky.”

I’m just glad to see the FCoE discussions moving away from arguments about which vendor is best-positioned to take advantage of FCoE toward more discussions about the technological merits and hurdles involved with the technology.

Wednesday, March 24, 2010

Check out our financials section

March 24, 2010 – As anyone who’s gone through a CMS conversion knows, things can slip through the cracks. When we migrated to a new CMS platform last year, we inadvertently omitted tracking codes for one of the components on our site – the Financials section. Since I wasn’t aware of that, I assumed that this section was the least viewed element on infostor.com because it never showed up in our site analytics.

However, we recently realized the mistake, added the tracking code and, much to my surprise, the Financials section is one of the most popular sections on infostor.com. In fact, it sometimes gets more page views than our News Analysis, Featured Articles or blog components.

Its popularity may be due to the fact that in addition to all the information you’d expect to find on a financials page (stock quotes, history, etc.), the section also includes a wealth of content related to specific storage vendors, including press releases, news stories, blogs, podcasts and conference call alerts and archives.

You can also track the storage market via the InfoStor Market Index.

If you haven’t been using the Financials section of our site (which is located in the bottom right corner of the home page) click here and enter the symbol of a storage vendor. Try EMC (and check out the news stories on Joe Tucci’s salary) or NTAP (and note that NetApp’s stock is trading at close to its 52-week high).

Friday, March 19, 2010

Gartner exposes FCoE "myths"

March 19, 2010 – Much of the talk about the demise of Fibre Channel relates to its prospects as a disk drive interface, where it is rapidly losing ground against SAS and SATA drives. But Fibre Channel as a SAN interconnect is still going strong, despite rapid growth in the NAS and iSCSI markets.

In fact, the Fibre Channel SAN market racked up record sequential revenue growth in the fourth quarter of 2009 – more than 15% quarter-over-quarter, according to the Dell’Oro Group market research firm. And the leading players in the Fibre Channel switch and HBA markets – Brocade, Cisco, Emulex and QLogic – turned in strong sequential performance numbers. Q4 2009 results were strong enough to put Fibre Channel revenues close to par with the record results recorded in late 2008.

Dell’Oro vice president Seamus Crehan attributes the strong performance to “improving economic conditions, overzealous budget cuts during the first half of 2009, government stimulus, alleviated supply constraints . . . and the server upgrade cycle that started in the second quarter of 2009.”

However, the future success of Fibre Channel as a SAN interconnect rests largely on the success of the Fibre Channel over Ethernet (FCoE) protocol.

InfoStor readers are by now very familiar with the potential benefits of FCoE and converged (LAN + SAN) networks (see Related Articles links at the end of this post). But a recent report from Gartner suggests that the concept has been over-hyped and that, in fact, many of the key arguments in favor of converged networks are flawed.

Here’s the Gartner press release, which includes a link to the full report (10 pages, $195):

Gartner Says Don't Assume That a Single Converged Data Center Network Is More Efficient Than Two Well-Designed Separate Networks

Converge All Data Center Traffic to a Single Technology, Not a Single Network

STAMFORD, Conn., March 16, 2010 — The notion that a single converged data center network makes for fewer switches and ports, resulting in a simpler network consuming less power and cooling, is flawed, according to Gartner, Inc. Gartner research shows that a converged data center network requires more switches and ports, is more complex to manage and consumes more power and cooling than two well-designed separate networks.

"The industry is abuzz with the promise of a single converged network infrastructure, this time in the data center core," said Joe Skorupa, research vice president at Gartner. "Alternatively described as Fibre Channel over Ethernet (FCoE), Data Center Ethernet (DCE), or more precisely, Data Center Bridging (DCB), this latest set of developments hopes to succeed where InfiniBand failed in its bid to unify computing, networking and storage networks."

"The promise that a single converged data center network would require fewer switches and ports doesn't stand up to scrutiny," Mr. Skorupa said. "This is because as networks grow beyond the capacity of a single switch, ports must be dedicated to interconnecting switches. In large mesh networks, entire switches do nothing but connect switches to one another. As a result, a single converged network actually uses more ports than a separate local area network (LAN) and storage area network (SAN). Additionally, since more equipment is required, maintenance and support costs are unlikely to be reduced."

In addition to the financial barriers to the success of a single converged data center network, Gartner also believes that there are significant design and management issues to be addressed. When two networks are overlaid on a single infrastructure, complexity increases significantly. As traffic shares ports, line cards and inter-switch links, avoiding congestion (hot spots) becomes extremely difficult. Mr. Skorupa said that over time, emerging standards, such as Transparent Interconnection of Lots of Links (TRILL) may make it easier to avoid these hot spots, but mature, standards-compliant implementations are at least two to three years away.

Debugging problems in the converged network are also more difficult since interactions between the LAN and SAN traffic can make root cause analysis more difficult. Since many problems are transient in nature, events must be correlated across the two virtual networks, increasing complexity. Should an outage be required for solving a problem or simply for performing maintenance, a downtime window that is acceptable for both environments may be required. This increases complexity and may increase cost, as well.

"It's clear that the barriers to a single network range from a dearth of available products and the price premium charged for those products to the requirement to "forklift upgrade" your entire network to long-standing organizational conflicts," said Mr. Skorupa. "However, while the promise that a unified fabric will require fewer switches and ports, resulting in a simpler network that consumes less power and cooling, may go unfulfilled, that doesn't mean that enterprises should forgo the benefits of a unified network technology."

Mr. Skorupa said that there is clear benefit in standardizing on a single technology for all data center networking if that technology adequately supports the needs of applications. This will simplify acquisition, training and sparing. However, settling on a single technology does not require that the networks be combined. Design, operations and troubleshooting is much easier with two separate networks, and it may also cost less to build two separate networks.

Additional information is available in the report "Myth: A Single FCoE Data Center Network = Fewer Ports, Less Complexity and Lower Costs."

Related articles from InfoStor:

Guidelines for FCoE deployment

The shift to FCoE is underway

Is FCoE the future of data center networks?

For a cogent rebuttal to Gartner's arguments, see "Gartner on FCoE: Whoa There, Sparky"

Tuesday, March 16, 2010

Startup’s SSDs top 1,000,000 IOPS

March 17, 2010 – According to independent test results, Pliant Technology’s Lightning Enterprise Flash Drives (EFDs), which are based on solid-state disk (SSD) technology, topped one million IOPS in a configuration with 16 drives.

Many storage vendors have issued press releases claiming performance of more than, or at least very close to, one million IOPS. Examples include Broadcom (press release), Emulex (press release), and Microsoft and Intel (blog post and video).

The reaction from most people (including me) would be “So what?” and/or “Who the heck needs one million I/Os per second?” It’s a legitimate response, because almost nobody needs that level of horsepower, even in the context of rapidly growing virtual server environments, multi-core processors, etc.

But that was also the response when 10GbE, 8Gbps Fibre Channel, 6Gbps SAS, etc. came onto the scene. And we know how those technologies are doing. The fact is, end users’ need for speed is insatiable, and if you don’t need it today – you will tomorrow.

The testing of Pliant’s SSDs was done by OakGate Technology, which supplies testing tools for IT technology providers.

As with any benchmark tests, it’s important to (a) have a grain of salt on hand and (b) look at the test configuration. The 1.1 million IOPS results were achieved with 16 drives and an 80/20 read/write ratio with 4KB block sizes.

The testing also revealed that that configuration can produce 2.3 million IOPS with a 100% read stream and 512-byte blocks.

Two ZT Systems Linux platforms with dual quad-core Xeon processors were used in the Lightning EFD tests, each running the OakGate FireOak test suite, which was controlled by a Java-based GUI running on Windows XP. Each system included four Adaptec 1045 SAS HBAs, with each HBA connected to two 3.5-inch, 300GB Lightning LS300 EFDs from Pliant.

For complete test details, see OakGate CEO and founder Bob Weisickle’s videos on YouTube:

· Benchmark 1: Real-world Workload
· Benchmark 2: Maximum I/O Rate
· Benchmark 3: Maximum Bandwidth
· Benchmarks 1-3 (complete version)

Friday, March 12, 2010

And the top 5 storage software vendors are . . .

March 12, 2010 – In yet another sign that a rebound is underway in the storage market (albeit a painfully slow one), the storage software market posted a slight increase in revenues in the fourth quarter of last year compared to Q4 2008 -- $3.094 billion vs. $3.079 billion, or a 0.5% increase. Q4 revenues were up 6.3% over Q3 (although Q4 numbers are almost always better than Q3).

That’s according to IDC’s recently released Worldwide Quarterly Storage Software Tracker report.

In the fourth quarter of 2009, EMC retained its top dog status with $734 million in storage software revenues and a commanding 23.7% market share, followed by Symantec (17.5%), IBM (13.2%), NetApp (7.9%), and HP (3.9%) and CA (3.8%) in a virtual dead heat for the #5 spot.

The only thing I found surprising here was NetApp’s strong showing. In fact, NetApp posted the greatest revenue gain from Q4 2008 to Q4 2009 – 7.9%. That increased the gap between NetApp and HP, which experienced a 12.1% decline in revenues from Q408 to Q409.

Looking at the full year, the picture doesn’t change much, although HP drops out of the top 5 list. On 2009 revenues of almost $2.7 billion, EMC was #1 with a market share of 22.7% (down from 24.4% in 2008), followed by Symantec (17.9%), IBM (13.5%), NetApp (8.0%) and CA (4.0%).

Under the storage software umbrella, IDC includes the following product categories: data protection and recovery, archiving, replication, storage management, device management, storage infrastructure, file systems, and “other.”

Related blog post:
Who are the top 5 array vendors?

IDC press releases:
“Storage Software Market Has Typical Fourth Quarter Jump, as Well as a Slight Increase from Last Year”
“Total Disk Storage Systems Turn a Corner, Posting First Year-Over-Year Gain in More Than Four Quarters”

Thursday, March 11, 2010

Who are the top 5 array vendors?

March 11, 2010 – Market researcher IDC recently released its Worldwide Quarterly Disk Storage Systems Tracker report, which provides revenue statistics for the external and internal disk array markets, as well as vendors’ market shares.

Although it’s no surprise that 2009 wasn’t the best year for disk array vendors, the market did end the year on a high note. Fourth quarter revenues for the total (external and internal disk systems) market accounted for almost 30% of the full year’s revenues, and represented the first year-over-year growth since the third quarter of 2008, according to Liz Conner, IDC’s senior research analyst, storage systems.

Q4 revenues were $7.273 billion, up slightly from $7.26 billion in Q4 2008. Total revenues for 2009 were almost $24.5 billion, compared to $27.8 billion in 2008.

There were no surprises on the leader board. Racking up 2009 revenues of $4.1 billion in the external disk systems market alone, EMC earned its 800-pound-gorilla epithet by being #1 by a long shot. EMC grabbed a 22.7% market share, trailed by IBM (14.2%), HP (11.7%), Dell (9.1%) and NetApp (8.6%).

Looking at the total (external plus internal) disk systems market, the numbers slide toward server vendors HP and IBM. With $4.5 billion in revenues, HP earned the #1 spot with an 18.4% market share, followed by IBM (16.9%), EMC (16.8%), Dell (11.4%) and NetApp (6.4%).

Those rankings were pretty much expected, but what struck me as weird was that there was virtually no change in those vendor’s market shares between 2008 and 2009. In each case, there was less than a 1% difference, despite all the high-stakes battling in this high-margin market.

When you slice the disk systems market into segments, there were a few mild surprises.

For example, EMC led the NAS market with a 50.5% share, followed by NetApp with a 20.2% share. I had no idea the gap between the two was that large, but NetApp must be making up for it in other areas (see “NetApp hit$ a home run” ). The NAS market grew a respectable 12.6% year-over-year, and now accounts for about 20% of the entire external disk storage systems market.

But if EMC and NetApp are hogging more than 70% of the NAS market, that doesn’t leave many scraps for the hordes of other NAS vendors.

The iSCSI SAN array market posted a very impressive 30% revenue growth in Q4 vs. the same period a year ago. Dell led the iSCSI market with a 31.5% slice of the revenues (maybe EqualLogic was worth $1.4 billion), followed by EMC with a 15.7% cut. Again, a little surprising because I would have guessed NetApp was in the #2 position.

You may have noticed that Hitachi Data Systems does not show up in the top five anything in the IDC report. That’s because IDC does not include OEM sales in its tracking. As such, HDS sales revenues do not reflect their OEM sales to HP, nor to Sun -- a partnership that’s set to expire on March 31.
See Kevin Komiega’s new story, Oracle-Sun ends HDS partnership
and my blog post, Who wins/loses in the Oracle-HDS breakup?

And you can read the full IDC press release here.

Monday, March 8, 2010

Who wins/loses in the Oracle-HDS breakup?

March 8, 2010 – In the fallout from the Oracle/Sun-HDS split, I imagine that the angriest bunch will be Sun shops that have been happy (yes, some Sun shops are happy) buying their high-end HDS storage arrays from Sun in bundled server/software/storage deals that translate into reasonable prices. Then again, there are plenty of other ways to buy HDS arrays: direct from Hitachi, from HP, and through HDS’ many channel partners.

If you missed the news last week, see senior editor Kevin Komiega’s story: “Oracle-Sun ends HDS partnership.”

Some observers suggested that the breakup was a deep blow to HDS, but I think Hitachi will get its share of the business with or without Oracle. (And bear in mind that the Sun-HDS reseller deal only involved HDS’ high-end USP arrays.)

And if not, the breakup could be good news for other high-end storage vendors, most likely EMC. In fact, EMC and Oracle are best buddies and maybe EMC had something to do with the Oracle/Sun-HDS breakup. (Then again, since when does Larry Ellison make decisions based on outside pressure?) It’s possible that Oracle will fill in the new gap at the high end of its storage lineup by inking an HDS-like reseller deal with EMC, where Oracle could leverage a much broader partnership than it had with HDS.

However, that goes against the prevailing opinion that the real driver behind the Oracle/Sun-HDS split was Ellison’s goal of becoming a soup-to-nuts IT hardware/software behemoth with total control over its stacks – and that means no hardware reseller deals.

If so, it’s not good news for LSI, which supplies Oracle/Sun with mid-range storage arrays. The survivability of that relationship has been under question since the Oracle-Sun deal was first announced.

And it will be interesting to see if this whole thing has any implications for Pillar Data (which is independent of Oracle but funded by Ellison).

The real winner in the fallout between HDS and Oracle, in my opinion, is the channel – and that includes both HDS resellers/integrators as well as VARs that carry competitive disk arrays and software. If HDS doesn’t grab the bulk of the business in Sun’s installed based via direct sales, that will be a plowable greenfield opportunity for the channel.

Some observers were surprised by the Oracle/Sun-HDS dissolution, and also noted that it was “sudden.” Far from it: This had been brewing for a long time. (In fact, I sort of predicted it back in April 2009; see “Oracle + Sun: Whither – or wither – the storage?” )

As for reseller opinion on the breakup, here are some comments from Greg Knierieman, vice president of Chi Corp., a Cleveland-based solutions provider that resells HDS (and other vendors’) hardware and software:

“First, I think it’s interesting that Oracle is not out in front of this with an announcement of their own in an effort to position themselves and to give guidance to their customers. Most of the enterprise Sun accounts are already working with HDS directly or with an HDS channel partner. In this respect, I think HDS and their channel partners are well positioned to maintain and grow with their Sun/Oracle customers.

“The other side of this is the Sun/Oracle relationship with LSI, which currently provides mid-range storage for them. Short of a specific announcement on LSI’s future from Sun/Oracle, there is now room for those very same HDS partners, like Chi Corp, to provide Sun customers with guidance and options for their mid-range storage.

“On who wins/loses: Oracle’s lack of a response has probably raised the concern of some enterprise customers. I doubt they are panicking, but they certainly have questions. While Sun’s business has been deteriorating for some time, there is still a solid base of dedicated Sun customers. Oracle also has a notorious reputation of charging an enormous amount for service, support and maintenance. Some of our customers who use Oracle software claim that their highest maintenance fees for any vendor is Oracle, so there is already skepticism about standardizing on Oracle hardware and what the long-term financial consequences will be.”

Tuesday, March 2, 2010

FCoE CNAs: HP/IBM tap Emulex, Cisco taps QLogic

March 2, 2010 – For quite some time, archrival QLogic has had bragging rights in the market for converged network adapters (CNAs) that support FCoE, iSCSI and 10GbE, but Emulex is now able to ante up with recently announced OEM design wins from HP and IBM. But QLogic still has the edge in the number of design wins.

On Friday, Emulex announced that its OneConnect Universal Converged Network Adapters (UCNAs) with support for 10GbE and FCoE have been certified by HP for use in select ProLiant rack, tower and blade servers. Support for iSCSI is expected in the second half of the year.

It’s important to note that HP will brand the Emulex adapters. In typical HP fashion, the three adapters will be succinctly dubbed the HP NC550SFP Dual Port 10Gb/s Server Adapter; the HP StorageWorks CN1000E Dual Port CNA (for ProLiant rack and tower servers); and the HP NC550 Dual Port Flex-10 10Gb/s Ethernet Adapter (for ProLiant blade servers).

Emulex’s OneConnect CNAs have full hardware offload for TCP/IP, FCoE and iSCSI, and users can start with a 10GbE NIC and add support for FCoE and/or iSCSI when/if they need it.

Today, Emulex announced a significant expansion of its OEM agreement with IBM. Big Blue will offer Emulex’s 10GbE NICs (upgradable to support FCoE and iSCSI) across eight IBM server lines, including the recently announced eX5 line and three BladeCenter families. Emulex already had an OEM design win with IBM for its Virtual Fabric Adapters. (QLogic's 8100 series CNAs are also supported on IBM's System x and BladeCenter platforms, for both 10GbE and FCoE connectivity.)

Also today, however, QLogic announced that Cisco has certified its 8100 series FCoE CNAs with the Unified Computing System (UCS) platform. Cisco has also certified QLogic’s CNAs for use with the Nexus 5000 series of switches.

If I were QLogic, from a marketing perspective I would continue to crow about my lead in OEM design wins. (In addition to Cisco, QLogic has FCoE CNA design wins with EMC, IBM and NetApp, as well as a number of second tier storage vendors.) But from a competitive perspective, it really doesn’t matter at this stage of the game.

It’s safe to assume that virtually all of the leading vendors will eventually qualify CNAs from both Emulex and QLogic. It’s also likely that OEMs will qualify CNAs from other vendors, such as Broadcom and Intel.

The good thing here is that users will have a wide variety of CNAs to choose from – unlike in the Fibre Channel HBA market, where it boils down to the Emulex-QLogic duopoly followed in a distant third place by Brocade.

Users will be able to choose based on price, performance, and support/performance for whatever protocols they’re using – 10GbE, FCoE and/or iSCSI.

Meanwhile, Emulex and QLogic may be headed for a court date. See "QLogic sues Emulex, but not over technology."