IDEAS Insights

Fujitsu Offers a Solid Storage Virtualization Solution

When we read about storage virtualization products today, we tend to read more about solutions from companies like IBM, EMC, NetApp, and HDS.  However, there are other companies that also provide solid block-level storage virtualization products. Some are emerging companies, such as Incipient, and others are more established like Fujitsu.  Fujitsu is a very well known company, but as expected it tends to get less press in data storage market than the big guys.  But that doesn’t mean that the company doesn’t have a solid set of data storage products, including a sound storage virtualization solution.  In fact, Fujitsu currently offers a very comprehensive set of virtualization solutions that cover server, storage and network I/O.

A complete virtualization infrastructure requires a multi-system solution that involves the virtualization capability combined with appropriate management software.  Fujitsu employs its “TRIOLE” approach, which includes virtualization, automation, and integration.  At the heart of the storage virtualization segment is the ETERNUS VS900 CPU.  This switch technology is provided in conjunction with autonomic software that helps to maximize the benefits and cost savings of a virtualized environment.  Although the virtualization software resides on the VS900 CPU, it is managed by software on the management server.  The virtualization and management software communicate and share information, such as configuration, copy, and migration data.  This two-prong approach of dividing virtualization functionality and configuration management over separate servers enables streamlined management, higher performance, and simplified troubleshooting.

Another key component to Fujitsu’s virtualization solution is its management software.. This software is designed to integrate the total system, i.e. servers, storage, and networks.  Managing resources in a stovepipe environment, even when virtualized, does not provide storage managers and systems administrators with that totally encompassing high level view that is so critical with today’s complex infrastructures.  Rather than relegating storage managers to manage only storage, sever managers to manage only servers, and network managers to manage only networks, Fujitsu provides total autonomous management of the entire infrastructure.   As business systems become more intricate, this level of integration is critical.  Moving forward, there will be a definite pattern of server, network, and storage management integration. Management integration is key to providing a seamless, virtualized environment

On the downside, Fujitsu’s ETERNUS VS900 supports only Fujitsu storage – it is not a multivendor storage solution at this time.  However, if a customer is using or evaluating Fujitsu’s ETERNUS storage, the VS900 is a viable storage level virtualization alternative. 

Liquid Computing Brings Telecom Technology to Traditional Computing

Every start-up company claims it has unique technology that will change the marketplace. But most of these companies quietly fade off into oblivion, never to achieve their goals. We at IDEAS have seen dozens of these startups come and go over the years. However, Liquid Computing’s LiquidIQ fabric computing server caught our attention as one idea that may actually end up changing the market.

Liquid Computing was started in 2003 by two industry veterans: CEO Brian Hurley, who spent over 20 years at telecom giant Nortel, and CTO Mike Kemp, who amassed over 30 years of experience building high-end servers for organizations such as Nortel and the U.S. Defense Advanced Research Projects Agency (DARPA). Together, these two with their team of engineers have built an entirely virtualized server with a high-bandwidth, low-latency backplane to connect the processors, memory, and I/O subsystems. Best of all, they can offer this product for about the same cost as traditional high-end servers. Throw in autonomic system management that is straight out of the telecom market and you have a very interesting and desirable server. We recently had the chance to spend an hour with Brian discussing his current product and a looking at what Liquid Computing may have in store for the future.

Today’s LiquidIQ server supports twenty processor blade modules and five I/O blades in a chassis, delivering aggregate I/O bandwidth of 200 Gbps per chassis (supporting Gigabit Ethernet, 10 Gigabit Ethernet, and Fibre Channel). Each processor blade supports four dual-core "Rev F" Opteron 800-series processors and up to 16 GB of memory per socket. Two of these chassis, with a combined total of 320 Opteron cores, can fit into a standard rack, and in the current release a dozen chassis can be lashed together to form a 960-socket computing behemoth if desired. There are a couple secrets that contribute to the performance and scalability of this server. One is the proprietary IQInterconnect midplane that features a low-latency, 100 GB/sec bandwidth between processor blades; the other is the use of AMD’s Direct Connect Architecture and Hyper-Transport technology. The server can either be deployed as an 8-core coherent memory SMP server (16 cores when the new Opteron Barcelona blades are available), or it can remain non-coherent for traditional message passing, similar to that used in Linux clustering. Because it is 100% virtualized, configuration changes can be made on the fly and resources can be dynamically allocated where needed.

As expected, the LiquidIQ server is garnering quite a bit of attention from the high performance computing market. Anytime you mix fast processors, a high bandwidth low-latency interconnect, and a good price, the HPC market immediately takes notice. Add to that a world record STREAM benchmark result (demonstrating high memory bandwidth) for four-socket servers and a victory over the Cray XT3 in the HPC Challenge (HPCC) benchmark set, and you have a very nice HPC server. But what about the commercial computing market? That’s where the real money is to be made. This server already contains many of the features commercial customers are seeking, such as virtualization of all resources, integrated high availability capabilities, automatic detection and recovery of failed components, and capacity upgrades that can be performed online. To be successful there, Liquid Computing will need to achieve scalability beyond the 16-core level or end up competing against commodity server vendors such as Dell. It will also need to ramp up its support infrastructure. Commercial customers need a higher level of support than HPC customers. No matter which direction Brian chooses to take the company, IDEAS feels this is one start-up that is worth watching closely.

Niagara 2 Chips Announced, but No Servers Yet

Sun has formally announced the multi-threaded UltraSPARC T2 processor chip, previously known as Niagara 2. Like the predecessor T1 chip (Niagara), the T2 chip has eight cores per chip and is aimed at workloads with natural parallelism, as opposed to serving job streams requiring fastest single thread performance. T2 enhancements include, among other features, eight threads per core compared to the T1’s four per core, a floating point unit per core compared to one per chip on the T1, plus the addition of built-in dual port 10 Gigabit Ethernet and a cryptographic unit per core. As announced, T2 clocks at the same 1.4 GHz as T1, but has 4 MB of L2 cache compared to 3 MB of its predecessor. Performance on a per thread basis will be enhanced by the presence of the floating point units as well as the Ethernet and Crypto offload engines. Nonetheless, a major difference is that T2 supports 64 threads compared to the T1’s 32 threads. Highly parallel workloads, such as web-facing or Java servers can likely take advantage of the increased number of threads.

Although systems using T2 were not officially announced along with the chip, their unveiling is anticipated soon. IDEAS expects 1U and 2U servers, similar to today’s T1-based T1000 and T2000 servers. Note that T2 does not support SMP (neither did T1), so rumored servers with two (or more?) sockets would operate as independent operating environments that share a rack-mount enclosure. Next year’s Victoria Falls (aka Niagara 2+) will support two chip SMPs. Thus, introduction of multi-socket servers for T2 could provide not just rack density but might serve as a presumably-upgradeable place-holder for Victoria Falls.

Sun indicates sales of T1-based systems are up 225% year-over-year, with $200 million T1 systems sales booked in just the last quarter. The T1 chip has addressed a customer need for low power consumption, high density computing. Many, if not most, T1 systems appeared to have been deployed in the front-end web-facing tier. With the T2 refinements, Sun will likely target more general purpose computing, looking to extend into mid-tier midrange application servers as well as continuing to satisfy the highly parallel front-end workloads.

Ubuntu and Gentoo Linux distributions are available for today’s T1000/T2000 and are expected to be available for T2-powered servers. Note though, that Linux support is “from the community;” Sun only formally supports Solaris 10. Solaris 10 Containers and LDOMs are available on T2. That is, the 64 threads of T2 can support 64 independent LDOMs, each with its own software stack.

Bottom line: T2 exists and servers using this chip are anticipated to be announced soon. For environments seeking low power consumption, high compute density solutions, the eight-cores and 64-threads of T2 should be a serious competitor against Xeon and Opteron servers (that are only now becoming available with four cores per processor module). There are many applications that can take advantage of the high multi-threading of T2, but sufficiently parallelized software may limit Sun’s push to extend beyond the front-end computing tier.

Tape Is Alive and Well -- As It Should Be

When it comes to backup and archive media, all the discussion of late seems to be around disk based back up.  We read about VTLs and disk libraries, as well all the technology that goes along with them, such as data de-duplication, compression, and encryption.  Disk has the reputation of being faster and more reliable than tape – and much better suited to restoring data.  Faster and more reliable is still being debated, although I do agree that restores from disk are much faster and convenient than from tape. All the attention to disk would lead us to believe that tape obsolete medium and is on its way to being replaced completely by disk based technologies.  However, this is certainly not the case – tape continues to represent an essential component of the data center infrastructure.  Ongoing advancements to performance, capacity, and price continue to ensure tape’s role as a critical factor in business continuity management.
 
Although disk based backup does get most of the attention these days, there have been some notable milestones in the tape space.  Probably most noteworthy is the arrival of LTO-4.  As the name indicates, LTO-4 is the 4th generation of the Linear-Tape-Open Ultrium technology that was developed jointly by three of the major tape vendors:  HP, IBM, and Quantum.   A major impetus to the development of LTO Ultrium was to provide an open format technology that would allow users to obtain these tape products from a variety of sources, as opposed to the vendor restrictions presented by proprietary offerings.  By definition, open source will facilitate the compatibility of products and media across numerous vendors.  In addition to IBM, HP, and Quantum, LTO can be licensed from vendors such as ADIC (now Quantum), Exabyte, Fujitsu, Overland, and Tanberg.  See the LTO web page for a complete list.

There are multiple benefits to users for moving from of LTO-3 to LTO-4.  First is capacity.   Assuming a 2:1 compression, LTO-4 doubles the capacity over LTO-3 from 800GB to 1.6TB.  Secondly, throughput is increased from 160MB/sec to 240MB/sec.  But perhaps the most significant feature of LTO-4 is the implementation of the Advanced Encryption Standard (AES).  Already incorporating Write-Once-Read-Many (WORM) functionality, data written to LTO-4 can also be encrypted for added security and compliance requirements.  Encrypted tapes that are lost or stolen cannot be read if they fall into the wrong hands; therefore, lost or stolen tapes (that are encrypted) may not necessarily turn into headline news. This added functionality does come at a price however. LTO-4 is currently over 50% more expensive than LTO-3, which could cause many users to examine the need for these added benefits.
 
The arrival of LTO-4 is an indication that tape technology continues to progress and provides a viable role in the data center.  Disk based backup is a sound strategy for a first tier backup (e.g. data 30 days old or less), however, tape remains the most cost effective and most widely implemented method of storing archived data.  Many data center managers are not asking themselves whether or not they should move from tape to disk, but rather how a combination of disk and tape backup media can meet their needs for business continuity and regulatory compliance. With the recent emphasis “green” initiatives, storing data long term on spinning disks does not come close to matching the cost efficiency of storage data long term on off-line tape.