Taneja Group | Blog Feed

When the Light’s Red, Whose Throat Do I Choke?

2012-05-16T16:43:28+00:00

Application performance management (APM) solutions get at the system-wide nature of application performance issues. They follow important parts of an application transaction across IT domains, stitch together a cohesive mapping and summation of the application's journey across IT infrastructure, and generate an alerting drill-down when application performance slows.

Sounds great, but the problem is that often these solutions are really meant for developers who can implement and understand deep app instrumentation. Otherwise they require significant transaction profiling, both upfront and as applications and systems change. Neither of these approaches is really suitable for dynamic production environments. It goes without saying that production is hardly the place to install developer level debugging tools, and with the size and scope of today's data centers there are few applications that can justify the cost and effort to continually maintain hardcoded transaction profiles (excepting perhaps trading applications in the exchanges).

BlueStripe's FactFinder has broken away from those approaches and figured out that their real APM value is to IT operators. IT ops folks just aren't up for instrumenting apps or modifying operating systems or maintaining transaction profiles. They just want to know whose throat to choke when the light goes red on performance - i.e. which level 2 admin in which domain should they call? FactFinder v6 adds Java app JVMs, Websphere, and Weblogic to its automatic application transaction discovery, mapping, and analysis. With these additions BlueStripe provides a fairly broad coverage of application components inside an enterprise data center, which might obviate the need for the other types of APM tools in production. IT operators should take a good look now that BlueStripe is aiming squarely for their production APM needs.

Citrix and the Mobile Cloud

2012-05-11T18:16:09+00:00

Citrix’s cloud announcements are coming in hot and heavy. One major outreach is providing cloud-based application delivery, file sharing and IT access control for corporations.

Mobile device end-users and IT are often on a collision course. On one hand corporate users are married to their mobile devices, particularly the iPad tablet, iPhone and Blackberries. Many end-users buy their own devices and intend to use them for work as well as personally. These users want what they want: application availability and file sharing, freedom to use their devices at will, and IT support even if the devices are private purchases. At the same time IT has to control application and data access, and to provide security against digital attack and data leakage.

Citrix is trying to create an environment that benefits both communities. Citrix has focused development around an integration between Receiver, ShareFile and CloudGateway 2.

Citrix Receiver software virtualizes and delivers applications to mobile devices and ShareFile lets users share files from those devices. This is good news for end-users who can access web- or network-based applications, and can collaborate and share data. CloudGateway 2 provides access control to IT needs by delivering approved Windows, SaaS and mobile applications to Receiver and supporting single sign-on. By controlling application and file delivery at the source, IT has a better chance of managing sensitive corporate data.

There were additional product announcements at Synergy, such as upgrades to video conferencing over the iPhone, better mobile support for GoToAssist, and launching of the social collaboration platform Citrix acquired with Podio. But the real news here is Citrix attempting to provide efficient access AND access control to the corporation and its mobile end-users.

AlphaLit Legal Review and Ease-of-Use

2012-05-11T17:55:12+00:00

eDiscovery review platforms aren’t exactly uncommon. It’s not surprising: there is a lot of money – a LOT of money – to be made in review. This also means a lot of competition so failing review platform makers can lose their shirts, but the risk is worth it if you are already established in the marketplace, have strong product features, sport a big differentiator that matters to legal users, and partner with complementary companies with major market reach.

AlphaLit is one of the oldest document review makers out there and is a solid provider for large corporations and law firms. Once a vendor is in these environments they may be solidly entrenched for some years. However, eDiscovery is moving and changing fast and there are some big review makers that are aggressively going after their competition. AlphaLit has responded by upgrading and differentiating their E-Direct 3.0 review platform. They have also made some strategic partnerships with complementary vendors Nuix and Content Analyst.

Their review differentiator is deceptively simple: ease-of-use. For technical audiences such as IT, ease of use is a good feature but rarely makes it to the top of the must-have charts. But for legal review users, ease-of-use is a vital feature and can make the difference between winning and losing sales opportunities. In fact, the larger the opportunity the more important ease-of-use becomes. When a purchaser has a large group of legal reviewers to train on a platform, then ease-of-use becomes tremendously attractive.

Simplicity doesn’t exist in a vacuum and E-Direct 3.0 offers the features we expect to see in a full review platform. Some upgraded offerings include batching and workflow modules, privilege log creation, powerful search, and analytics in concert with partner Content Analyst. But even these features are characterized by ease of use with one-click access, batch-based organization for better search, and simple interfaces.

Like we said – ease-of-use may sound deceptively simple, but simplicity is a big selling factor in the crowded review marketplace.

Citrix Accelerates Users’ Path to the Cloud

2012-05-10T18:26:51+00:00

Many of the end users we speak with on a regular basis have not yet embraced the cloud. When we ask what’s holding them back, we often hear users’ concern that the cloud will force them to turn their IT infrastructures and operations upside down, causing them to lose control and opening their environments up to security threats and all manner of inefficiencies.

Citrix recognizes this, and based on a series of announcements at Synergy this week, is bringing to market a series of offerings that enables users to evolve to the cloud at their own pace and in a non-disruptive way.

Let’s take a quick look at key pieces of Citrix’s cloud portfolio:

Citrix CloudStack IaaS enables customers to build and manage their own public or private compute and storage clouds. By providing CloudStack as an open-source offering under an Apache Software Foundation license, Citrix has broadened its appeal – with more than 60 industry partners announcing support – and has increased the rate of user adoption at the same time.
The new Citrix CloudPlatform, also under Apache distribution, brings an orchestration capability to CloudStack, allowing users to deliver secure and reliable cloud services.
CloudBridge2 enables users to transparently connect their private clouds and data centers to third-party clouds such as Amazon AWS. This lets private cloud administrators leverage low-cost public clouds as additional data storage tiers. CloudBridge2 offers pre-configured cloud connections and simple provisioning for first-time connections.
The NetScaler 10 cloud-enabled application delivery platform allows users to cost-effectively scale their clouds and meet service commitments as their business grows. NetScaler’s TriScale technology further simplifies the scaling process for large enterprise and carrier networks. Citrix also plans to customize NetScaler 10 for popular networks, starting with AWS.
XenDesktop with integrated XenApp eases the painful process of running existing Windows apps and desktops alongside new mobile and web-delivered services. XenDesktop decouples Windows applications and desktops from local networks and transitions them into web-delivered services. In addition, Project Avalon integrates CloudPlatform, Apache CloudStack, XenDesktop and XenApp to quickly deploy personalized Windows apps and desktops via a private cloud, with easy access to public clouds for data retention and sharing.

Citrix also revamped its mobile strategy in a big way, which we’ll discuss in a separate post. Collectively, this portfolio of cloud offerings elevates Citrix into the big league of leading cloud technology providers while providing a platform that perfectly complements its newly announced mobile enterprise strategy. Granted, not all the pieces are in place yet and Citrix is trailing competitors like VMware in this segment of the market. Nonetheless, we’re excited to see Citrix’s cloud story coming together, and hope that both IT and service providers will stand up and take notice at what the company has to offer.

FalconStor Dedupe Redux

2012-05-09T15:42:25+00:00

We talked with Darrell Riddle of FalconStor on their Dedupe Redux, more familiarly known as VTL 7.5 with deduplication. Lots of VTLs have deduplication; it’s rare these days to do any kind of storage on disk without deduping the data. The challenge is where does dedupe happen and how fast does it happen? Concurrent, Inline, Post-Processing? They all have advantages and disadvantages. Deduping at source (concurrent with the backup process) means smaller data is moving through the pipes; an advantage on busy LANs and a really important deal over WAN bandwidth. Issue: processing for dedupe at the source can slow down backup, which is already overrunning backup windows. The last thing it needs is another factor to make it take longer. Deduping inline takes processing out of the source so it’s not impacting backup process performance, an efficient move for environments with heavy backup requirements. Issue: Inline can suck up bandwidth and you need good-sized pipes; it also needs to be fast so it does not retard backup speeds. Post-process deduping lets data reach the storage target at top speeds for the fastest possible transport speeds and performance; dedup product may be a third-party software or running at the storage array level. Issue: Back-end disk storage doesn’t have to be huge because data is still getting deduped, but front-end cache and throughput should be very fast so it won’t retard incoming performance while deduping.

Dedupe vendors of course talk about how source, inline or post is the BEST – the “best” depending on what they make, of course. But there isn’t one “best,” there are different environments and architectures that will benefit the most from different dedupe choices, often within the same enterprise.

FalconStor’s VTL 7.5 is stepping up the game for everyone because they have made a way to do all three, and do them well. The FalconStor VTL could already flexibly dedup at the current and post positions, and now they have added inline capabilities to serve all three dedupe choices. FalconStor Turbo also lowers cycles and boosts speeds at the post and current dedupe locations. We would like to see Turbo added to inline; but even now flexible choices with good –to-exceptional speeds allow customers to tailor dedupe processes to differing systems and infrastructure. We hope to put their ingest and recovery speeds through their paces in our company labs soon, and salute FalconStor for this breakthrough flexibility in VTL-based deduplication.

Whiptail Invicta, Not Your Grandfather’s SSD

2012-05-08T22:19:16+00:00

SSDs aren’t big news anymore. They’re a critical part of high storage performance, no doubt about that. But to qualify for an announcement, I suggest that SSD has to offer something beyond the mere fact of its presence in a storage system.

Whiptail offers a lot more. After 3 years in the SSD business, Whiptail upgraded its XLR8r flash storage array and renamed it Accela. No, that’s not the big news – the news is that the revamped Accela can become a component node of the new high availability, high performance Invicta appliance. Invicta is a scale-out, fully SSD array that yields exceptionally high speeds for high storage IO environments.

The Invicta uses Silicon Storage Routers to manage data protection (including redundancy) and connectivity. Scaling all the way up to 6 Accela nodes boosts capacity to 72 TB of NAND flash and, more importantly, boosts IO speeds to nearly half a million read and write IOPs. Even at 2 nodes read/write IOPs come in at 300,000 and 250,000 respectively. Bandwidth is equally robust at over 2 GB/s for double nodes and up to 7 GB/s read, 5 GB/s write at the top scalability. The multi-protocol system supports both block and file: iSCSI, Fibre Channel and InfiniBand, still common in HPC environments, and file protocols NFS and CIFS.

A fully SSD system is meant to store highly active data in heavy streaming and transactional environments. (Word to the wise: make Invicta the lead act in a multi-tier data retention environment. This kind of array isn't meant for long-term data storage; don't tell me that doesn't happen with expensive primary arrays. It does.) When you truly need next-gen fast storage processing – at speeds we didn’t expect to see quite this soon – the Invicta is your man. Or rather your machine.

Storage Interface Throw-Down

2012-05-03T17:18:05+00:00

These days it seems that the hubbub around the differences between the tech giants has died down to steady sailing. But comparing and contrasting Apple and Google has often been an entertaining subject in the past, and it occurs to me from time to time that the contrasts have increasing relevance to the storage industry. Google historically has had a lot of offerings and a lot of successes, but they have always suffered from a lack of consistent user experience across products. Today they still lack this experiential element. This is a sharp contrast to Apple, who took a "button" and amplified it into a singular thematic element of user experience. They’ve had an almost immeasurable amount of effectiveness because of it.

What does this have to do with storage? Across a huge swath of the market, from SMB to enterprise, we’re seeing an increasing number of vendors amongst the latest generations of storage offerings take this experience factor to heart, and I frankly think the current generation of IT buyers will determine the success of IT products based on user experience and well built interfaces more than ever before. In fact, I’d even suggest that “user experience” is the backseat driver in all those vendor cars racing after unified storage nirvana.

So who would I call out? Among new companies, StorSimple has created a serious interface polish throw-down. Among big companies, Dell EqualLogic has led the charge and their level of integration here is largely responsible for their SMB success. In point of fact, the Dell guys have a uniquely disciplined orientation toward this - I know from my interaction with them and my hands-on experience with some of their pre-release products, Dell often considers things unacceptable or insufficient to mention when from my perspective they've already shamed most of the competition. I'll leave the specifics out to protect the innocent, but there's more functionality in Group Manager and SANHQ than most recognize, and that power extends to other products that Dell is mashing up with EqualLogic gear. Meanwhile, IBM has some serious polish happening in their SVC and V7000 interfaces, and products like HP's scale-out NAS are showing some serious capability as well. Think I'm missing a notable example that does better than these? Vendor or customer, let me know. But if you're a vendor, you better show it off - my place or yours, I have a few U's of rack open

Excuse me, but I think your cache is showing…

2012-05-03T13:13:27+00:00

Everybody these days is adding flash-based SSD to their storage arrays. Some are offering all flash storage for ultra-high performance. And a few are popping flash storage right into the server as a very large, persistent cache. But taking advantage of flash in these ways requires either hardware refresh or significant service disruption - or both.

GridIron offers a drop-in, non-disruptive way to immediately super-charge existing infrastructure. Their TurboCharger appliances logically plug into the middle of the SAN fabric where they can be installed (and removed) non-disruptively by taking advantage of I/O multi-pathing. Once installed, they jump in to the data path as a virtual LUN fronting the real LUN on the back-end, providing a massive amount of SSD write-through cache that automatically adjusts to multiple workloads. Because it's in the SAN, TurboCharger can virtually "front" any underlying storage - even storage that is in turn further virtualized.

GridIron customers have generally faced serious data access challenges with large databases and in consolidated and virtualized environments that benefit from read-intensive IO acceleration. GridIron is now expanding its product line to help accelerate structured and unstructured "big data" access. The OneAppliance all-Flash product line includes the FlashCube for offloading temp, log, and scratch space write-intensive workloads, and an iNode that combines massive flash and compute together for building high-performance compute clusters.

GridIron is clearly differentiating from other flash solutions in its direct and practical approach to bringing the power of flash to bear directly on the extreme data access and movement problems with big data.

GridIron and the Performance Race

2012-05-02T23:27:32+00:00

The performance race – racing toward somewhere, versus just racing.

Lately we’ve been having an increasing number of conversations and witnessing an increasing number of announcements around how vendors are chasing extreme storage performance challenges. My 2 bits is that we’re at the cusp of a big split between solutions. In fact Som Sikdar, the CTO at GridIron, echoed some of my sentiments in a conversation just yesterday.

I’ve long been a bit disappointed by the industry’s collective first generation take on SSD integration into arrays, which amounted to little more than inserting a replacement solid state device. For those following this space for a bit, I don’t need to rehash any story about array controllers falling apart, performance being disappointing, etc. etc. Yawn, yawn.

But lately, the game looks like it is on. Some vendors are integrating SSD in unique ways in the array and coming up with seriously elegant architectures for general purpose storage. Kudos to these vendors and more on this in a later post -- because where they are applying their energy, they are changing the world for business whether their focus is SMB or enterprise.

On another hand, we’re seeing other vendors charge hard towards extreme performance environments. The market here is early, and there’s a battle being waged between a few different approaches. Extreme performance issues can be solved with point products applied at the server to enable high performance. In fact, that seems to be some of the dominant flavor of the day, in part because it seems lower risk and lower upfront cost. And clearly Fusion IO has beaten a fierce marketing and messaging drum in the market here. They have consequently claimed mind share and continued to introduce some pretty cutting edge features such as their recent application APIs.

But even more recent innovators are tackling extreme performance with shared appliances that sit in front of shared storage environments. And as the GridIron guys laid out, the appliance guys can be just as innovative in integrating APIs and application-level tech. Now if you’ve looked around the storage market long enough, chasing a hook here and there isn’t all that break-through. The IBM guys have long done it with DB2 and AIX cache hints being passed to a DS8000 series storage array. You might even remember the Panasas guys chasing some Hadoop integration right before HDS picked up their remains.

But with high speed storage accelerators, clearly APIs make tremendous sense. They can in effect transform the application interaction path by letting more elegant transactions pass back and forth without all of the overhead of the IO path, and thereby massively amplify the performance effect.

I still put my bets on the shared guys in this market. Eventually it will become pretty clear that localized is more complex over time, as it has some serious overhead. But regardless, as this direction continues we may see acceleration technologies turn into something entirely different.

Amazon Defines Big Data As Big Opportunity

2012-04-27T17:21:15+00:00

Today I had the opportunity to hear about Amazon's Big Data solutions at an Amazon Web Services Big Data Summit here in Boston. At this event crowded with local tech talent from the hot bio, research, financial, and web industries, AWS showcased their EC2 compute cluster instances for HPC and their Elastic Map Reduce service that runs Hadoop in the cloud - trotting out several interesting real-world users.

John Rauser, an Amazon big data scientist, presented a thought provoking session on how Amazon uses and views Big Data. I wouldn't want to shamelessly steal his material but I just have to relate his definition of Big Data. He said that you are really dealing with Big Data at the point when you have data that needs distributed processing. In other words, it's Big because it's more than one node or a single monolithic application can handle today, or even can be expected to handle "forever" as the dataset grows. Once you have to cross the threshold to "distributed" you have effectively entered the land of the Big.

In this view the effective Big Data market isn't just the "extremely Big Data" of petabyte sized datasets that others are talking about. Those petabyte apps get a lot of news but they are out on the long tail of datasets. Rather, the Big Data opportunity is every potential analysis and app that is just out of reach of current single node IT systems, developers, and operators, up to and including the petabyte monsters. That covers a really broad swath of datasets that isn't limited by a hard threshold on dataset size.

I really like that Big Data definition. It's a practical and useful way to think about when you might and should get into Big Data technologies. And it clearly drives into Amazon's strategies to enable both cost-effective analytical development and ongoing cluster operations for any indefinitely scalable dataset, even if those datasets are only a few hundred GB today. With all datasets conceivably growing large eventually, eventually all datasets will be Big Data. If you are developing any new data analytics apps, the time to develop it as a Big Data app might be now.

NetApp E-Series and Big Potential

2012-04-27T15:48:39+00:00

NetApp recently announced some new features for their E-Series technology that’s based on the block storage piece of their LSI Engenio acquisition. The features announced here revolve around DDP – loosely a new disk aggregation, parity protection, and disk group virtualization technology set – and SANtricity, the management layer. Block storage systems and features are seemingly non-remarkable on the surface; I don’t know that I’ve seen all that much chatter about it.

But frankly, NetApp is demonstrating some interesting evolution in the E-Series technology through the intersection of their traditional expertise with this new block technology. Some pundits initially didn't see it as having a clear fit in the portfolio, but in a relatively short term, NetApp has taken a very well honed area of expertise -- data integrity and homogeneous volume virtualization -- and have used it to create DDP. On the surface it looks like a unique differentiator for the E-Series platform. Note that volume virtualization is so often labeled with WAFL (Wide Area File Layout) that most don't see it as in array volume virtualization. But it is exactly that.

There are more tools in the NetApp toolshed from acquisitions like Bycast and others. We suspect these other technologies will foment future, deeper innovations on top of these two NetApp storage systems, which at present look like they are lining up with how workload segregation commonly happens in the data center, mostly around primary IO (FAS) versus deeper, bigger content (E-Series). This announcement supports this potential and we’re looking forward to seeing more from NetApp.

NetApp E-Series and the Dynamic Disk Pool

2012-04-27T15:21:30+00:00

NetApp just announced some major enhancements to the E-Series storage system that optimizes the platform for big data usage. We were particularly interested in Dynamic Disk Pool (DDP) because it is running counter to some of today’s big data trends that include RAIN architectures for replicating data across nodes, and scale-out storage arrays. Instead, NetApp's DDP heightens performance and disk protection by distributing parity and data across all disks in the pool. NetApp also claims to rebuild a pool following disk failure up to 8x faster over traditional RAID.

DDP offers 64 TB per pool. Ever-growing data volumes require more disks per LUN, which compounds the probability of disk failure. DDP, with 64 TB per pool, seems to counter this increased probability of failure head-on. This provides a significant advantage to environments storing larger and larger data sets.

Eventually, we think this kind of approach could mean the end of business risk due to disk failure. We hope that arrays continue to get smarter and self-healing to the point where end users aren't affected under expected, common component failure. This will go a long way towards protecting active big data environments, which require high levels of performance, capacity and business continuity.

Virtual Instruments and Data Centers, and Why it Matters. A Lot.

2012-04-26T17:39:26+00:00

In March I published my article on meaningful visibility into the data center. I mentioned several vendors who are doing good work in this challenging field including Virtual Instruments (VI).

There are lots of monitoring, capacity planning and general trending products out there. These vendors market their products to virtualization and cloud markets hoping to catch the big wave of investment in those fields. There is nothing wrong with these tools and IT needs them in discrete settings, but they don’t go nearly far enough in managing new levels of data center complexity. They offer some visibility, but dynamic data centers require not only information but also the ability to correlate information across multiple systems and to automatically act on it. We call this crucial piece of the puzzle “instrumentation.”

So why is instrumentation more important than ever before? Because virtualization and cloud computing add layers of complexity to the IT infrastructure, while simultaneously generating high expectations and SLAs to match. This is an impossible state of affairs if you have little visibility into your data center. There is just no way to track performance and SLAs, let alone optimize them. You need instrumentation that lets you see what’s going on and to efficiently act on it.

Your instrumentation product needs to have some critical capabilities. First off it needs to provide insight starting at the lowest physical layer, not just surveying across the logical or virtual layers. Data centers are both physical and logical entities and each element contributes to its health (or lack thereof). Purely logical visibility will ignore the base physical layer, even though those problems will cascade into big ones in a complex infrastructure. You need to optimize your lowly wiring just as much as you do your hulking SAN, because if your wiring or optics or any other physical element is compromised, your whole data center suffers. In addition to vertical movement up and down the stack, cross-correlation allows for horizontal movement across storage, server and networking domains to further optimize data center operations. Combining these two movement axes is powerful stuff.

The solution also needs to be low impact and real-time. If your instrumentation product is sucking cycles from your data center, then you are shooting yourself in the foot. Get an efficient package with minimal impact on the infrastructure. And make sure it’s as dynamic as your data center is – you need real-time observation, alerting and action across multiple domains.

Sure this level of control is a tall order, but this capability is what separates vendors like VI from the pack. The products are out there and they work, and we strongly urge you to take advantage of them.

White Servers from Intel Now Available with Tightly Integrated InMage Software

2012-04-26T13:52:04+00:00

On March 19, 2012, Intel and InMage introduced a packaged solution at Intel Solution Summit that intrigued me. InMage, as we know, is a data protection software vendor that has been in the industry for 6+ years. Their software suite called Scout is essentially a server based CDP that eliminates backup windows and allows for very rapid (in most cases instant) recovery. Recovery snapshot capability is built in, as is replication, failover and failback. One product to solve both backup/recovery and replication issues. Over the past few years they extended the solution to include Windows, Linux, Solaris, AIX, VMware, Hyper-V and XenServer support. So what is a data protection vendor doing with Intel?

Before we look deeper into this Intel/InMage relationship let's spend a moment on InMage. The company has mostly focused on the SME space, even though over the years it has built a good enterprise level clientele as well. They survived the 2008-2009 collapse that killed many smaller storage companies and continue on a steady growth path. But then I see this Intel/ InMage partnership which stands out. So I looked deeper into what it is and whether it is important.

Essentially, InMage has tightly integrated their Scout software with Intel's server management and storage management software. That means anytime an Intel server chassis or a board ships to one of its partners (thousands around the world but not the ones with their own brand, such as Dell, HP, IBM), data protection is bundled in. This is the first time we have seen this level of integration between servers and data protection software coming from a white box server vendor. Frankly, it is even rare from the name brand server vendors since their VARs generally want the option to choose which data protection software is best suited for the customer. In this case, the server ships with the integrated software.
I think the implications are serious. First of all, I think it is a coup for InMage. In a flash, their software will permeate tens of thousands of servers worldwide. Intel will do all the marketing while InMage gets to focus on building the best possible SMB data protection solution. While the per-server margin is probably small there are near-zero marketing and selling costs, and at expected volumes, net margins would likely be in tens of millions, in our estimate.

I expect this will have a strong net positive effect on both Intel partners and InMage's other selling efforts. Integration by Intel at this level is proof point that the product has been put through the paces and works. A lot of these Intel servers and boards end up at SMB/SME type customers who don't have expert IT capability. They will likely welcome such integration, especially when it is supported by Intel. We also understand that the request for this integration originated from Intel Technology Partners (ITPs) who felt this deep integration would allow them to compete more effectively with name brand server vendors and improve their margins to boot. So we see Intel, ITP, InMage and most importantly, the end user winning out with this strategy.

And, of course, InMage suddenly became a serious competitor in a market space that is littered with a ton of riff-raff.

Hitachi Data Systems Unifies Block, File and Object

2012-04-25T16:04:37+00:00

Following vendor announcements is an important part of what we do as analysts. Product introductions and new technology developments help us keep our fingers on the pulse of the industry.

But there is rarely an announcement that shakes up the industry, especially when the announcement is centered on mid-market – traditionally a less well-heeled segment than the enterprise. However, mid-market spend is increasing as businesses increasingly adopt virtualization and databases, and need storage to go along with it – yet do not have the big dollars to spend that the specialized enterprise storage groups might have.

In the past, vendors have taken successful enterprise models and stripped them down for mid-market. That approach is no longer working as mid-market has important requirements of its own. Vendors woke up to this fact several years ago and started to develop systems optimized for transactional processing and virtualization network storage, located in simple-to-mange systems that packed a lot of power and intelligence under the covers. The systems are optimized both for mid-market and the cloud storage industry, which has similar issues and needs. Vendors offering cloud-based storage needed to find a way to offer economical and powerful storage to their customers. Without that, they risked squeezing their profit margins.

Some vendors have introduced very attractive products to mid-range and/or cloud-based systems. Nexsan introduced a very nice E5000 series a few months ago that applies storage intelligence and fast SSD-based processing to file and block data. We also like GreenBytes Solidarity for its fast SSD, single tier storage system for the mid-market, and IBM PureSystems for a really interesting approach to integrating storage systems and cloud optimization in a single appliance.

Now Hitachi Data Systems (HDS) has stepped in with a new level of unified performance for block, file, and object storage system that is priced for the mid-range. HDS built the system on recently acquired file architecture from BlueArc, and wisely spent the necessary development time and dollars to maximize the platform for cost-effective unified storage with a small footprint and high performance. The system is based on a single clustered namespace and an object-based file system that enables high performance search, copying for replication and snapshots, and automated tiering.

The platform supports object storage and key block and file protocols including Fibre Channel, iSCSI, NFS and CIFS. Management software simplifies the process for mid-range IT, which often lacks the specialized experience of high-end enterprise storage administrators. Block and file data may share the same storage pools for consolidated storage and management.

System capacity scales to nearly 3PB without impacting performance. Its dense architecture is packed into small footprint. Depending on the model, HUS block systems house 2.5 and 3.5 drives or front-end port modules. Customers may add single node or clustered file modules for NFS, CIFS and FTP file sharing access, and dense expansion trays add extra capacity.

Dynamic virtual controllers automatically correct performance issues and enable fast provisioning. Multi Level Cell (MLC) SSDs. MLCs and 400GB SSDs lower cost per capacity over single level cells while HDS made more performance improvements to ports, ASICs and processors, and memory management. Metadata development increases snapshots capacity to 1024 and also improves snapshot, replication and dynamic tiering speeds.

All of these improvements and more were made with an eye to efficiency, which allows HDS to offer a highly advanced system at a reduced price per GB – a terrific advantage to mid-market and cloud-based storage needs. We believe HDS is the first major storage manufacturer to effectively integrate block, file and object storage in a single cost-effective system, and in so doing has raised the bar for other unified mid-market storage offerings that follow.

GreenBytes Solidarity Offers Better than Solid Performance

2012-04-24T15:21:39+00:00

GreenBytes Solidarity is a high availability primary storage system with a single, highly optimized SSD storage tier. It’s targeted to the SME market where virtualization is growing by leaps and bounds, which presents its own set of challenges around storing fast-growing data.

GreenBytes GO OS is Solidarity’s snappily named operating system that provides inline, real-time deduplication and compression. The OS also provides a technology called Smart Write that coalesces writes and the transactional IO associated with databases. This helps to optimize system endurance and performance, and lets databases take advantage of the efficiencies of single tier storage.

The dual controller unit offers 3.5 TB (15 TB deduped) to 13.5 TB (60 TB deduped) along with high IOPs for very fast throughput, a must for transactional applications. Virtual storage pools turn single tier storage into virtual tiers without the expense and overhead of physical tiering products and data movement.

The overall value proposition is optimizing data for greater capacity and longer term retention on disk, while simultaneously optimizing every stored byte for maximum performance. The product stretches the delivered capacity and performance of storage, particularly storage behind virtual infrastructures. This is where fewer datastores ease management and increase the benefits of dedupe.

Solidarity offers strong features and high availability to a market that doesn’t usually get these benefits from their storage systems. This is certainly not something we see every day. It’s a promising technology; we hope to put it through its paces very soon.

DH2i: Getting Mission Critical Apps Virtualized

2012-04-18T19:15:49+00:00

Last week we found ourselves in a discussion with a new vendor in the virtualization space. It’s a crowded space to get into but their target seems right on – virtualizing mission-critical applications. That’s a hot topic these days as folks try to squeeze even more benefits out of enterprise virtualization efforts by tackling apps that have high availability, performance, and integrated storage requirements.

DH2i presented their approach to virtualizing the application at a distinct level above the hypervisor-based virtual machine. In fact their approach to application virtualization is agnostic as to whether the underlying O/S is running on a virtual or physical instance.

It’s interesting to consider the limits of server virtualization when it comes to managing critical applications. For example, can server hypervisors ever provide the same virtualization benefits to individual database instances that they provide to machine instances? The primary database owner, usually a DBA, tends to work at the instance level and may have hundreds of instances to manage. Sure, each database instance could be run in its own virtual machine, but that could lead to terrific VM sprawl, not to mention huge licensing costs and perhaps unnecessary strain on the VM hosts. And then the DBA effectively needs to administer each VM instance rather than focusing on the database layer.

Ideally the database application itself would be abstracted from the O/S, enabling the DBA to work more effectively with idealized “virtualized” database instances. IT admins could then host and shift those virtual app instances around the data center (or cloud), as cost, capacity, performance, and availability needs dictate.

What DH2i is offering with their first release of DxConsole, database virtualization above the O/S for Microsoft SQL Server, sounds promising:

· High workload mobility – migrate database instances seamlessly across VMs and physical hosts as needed

· Incredibly simplified database app version maintenance and upgrade with golden images not of whole VM’s, but of database installs

· Avoiding machine sprawl – virtual and physical – since effectively isolated database instances can easily be stacked on the same O/S

· And for some databases like SQL Server, stacking in this way can save big licensing costs

Application virtualization appears to be the necessary next step in bringing virtualization to mission-critical applications. By layering application virtualization on top of server and storage virtualization, the benefits multiply. And it’s not just for layering on server virtualization - we think application virtualization could soon play an important role in the more agile enterprise in the areas of DevOps, quick test and dev cycling, hybrid cloud hosting, and big data analysis.

IBM Puresystems, or, How Many Heads Does Convergence Have?

2012-04-18T16:23:08+00:00

I just wrote a column for Storage Magazine on the nature of convergence in the IT industry and specifically in the storage industry. The column will run in the June issue. “Convergence” may seem like a market-speak term that has fallen out of favor, but the reality is that convergence is alive and well, and a deeper influence than we often think it is. The convergence value proposition only continues to grow as convergence trends have become an on-going design element. Think unified storage, virtual data protection, continuous data protection (even though they are often not called that in their latest incarnation), or cloud-integrated enterprise storage. What do they have in common? Convergence.

IBM has just proved my point by announcing IBM PureSystems, which seems to be a major convergence play. Let’s take a quick look at convergence history to see why IBM might be a very different take on a long-term industry development.

For a while convergence was driven by lock-in fears. Pundits were crying lock-in when Oracle and HP were first duking it out, though I frankly think lock-in is less a given than many make out. Maybe those two made convergence a little more obvious by turning it into a soap opera, but convergence has been a much longer and broader running trend.

Next to these two vendors, Cisco’s UCS probably made the biggest splash over the past few years. Cisco paved the way by proving some interesting things can be done, and new approaches can come with success. It isn't that they pulled off raw, clever innovation, but rather took some good ideas and made it successful - many of those ideas are probably in part attributable to pioneers like Fabric7.

That led to folks like Pivot3 being able to demonstrate some success with something a little more innovative, shifting the focus to storage + compute. Next up, Nutanix. And soon to be added to the roster of current vendors will be Scale Computing. You might call this something along the lines of super convergence, because it gets pretty clever when both the storage and the compute can scale. And we certainly shouldn't overlook Dell. HP has furiously waved the "Converged Infrastructure" and "Converged Storage" banner but to be fair, Dell has long been selling product that fits into a rack-based converged data center vision. So convergence certainly isn't new, and it is easy to see it is a long running influence that is way too often overlooked.

The IBM PureSystems convergence takes this development a step further, and at the same time strikes back at a shift that happened many years ago. Distributed applications and Microsoft once swept in and pulled the IP and much of the value away from system vendors. Now in 2012 comes IBM’s convergence take with a twist: they built it with more intelligence than we commonly see. PureSystems is a packaging of hardware and application that is fully optimized and tightly integrated. That type of system used to be a synonym for “complicated” and struck fear in customer's hearts. But times have changed and the market with it. The largest part of a system cost is no longer CAPEX, but its integration with other infrastructure elements and on-going management requirements and costs. Meanwhile, increasing power and better architectures mean a right-sized system can be acquired fairly simply, and then that system can be scaled.

Now with trends toward open source, cloud, virtualized abstraction, fewer applications and infrastructure dependencies, and heightened compatibility; the system vendor's play is back. The differentiator is integration up the stack. Is IBM taking the convergence equation even further? Are we seeing a paradigm shift? Time will tell, and we’ll be watching.

IceWEB: Nailing Unified Storage

2012-04-18T16:20:30+00:00

IceWEB is a unified storage player that has been in the market for a number of years, with a big footprint in federal government. But the market is awash in new primary storage offerings, many of them focused on unified storage. NetApp obviously held the reigns for a long time, but a few representative recent contenders include Scale Computing, Nexsan’s E5000, Dell's Fluid File System offerings, companies like Nimbus, and more. This newly crowded market means that even established unified storage companies have to have concrete differentiators from their competitors, or they’ll lose out to more attractive new offerings.

IceWEB seems to have some clear differentiators including an enterprise-grade storage subsystem and a user interface that is more user-friendly than many we’ve seen. These two differentiators should help IceWEB to go further than competitors in delivering both infrastructure and operational efficiency.

The underlying storage subsystem seems to be built for enterprise-class primary use, and is architected for better workload isolation and more granular performance tiering (including across SSD) than many products in the market.
A highly polished interface that looks more sophisticated than a lot of the other stuff that has been coming to the market. At first glance an interface may not seem like a real thrill. But a highly useable interface is conducive to better user adoption and comfort, which in turn makes a unified storage offering even more attractive and marketable.

It seems like IceWEB has something to show off, and I'm personally looking forward to putting their product under the magnifying glass to see if the efficiency factors really surface. Perhaps we'll see them in the Taneja Group Lab in the near future.

IceWEB appliance screenshot

VMware Cloud Foundry: Linux of the Cloud?

2012-04-14T07:40:08+00:00

Cloud Foundry celebrated its first anniversary this week, and it’s looking mature beyond its youthful age. Tens of thousands of developers are using the platform-as-a-service to design, build and test cloud-enabled applications. Cloud Foundry is supported on major public clouds such as Amazon Web Services, and can run on private clouds enabled by infrastructure stacks such as Nimbula and Eucalyptus. Many developers also run Micro Cloud Foundry, which provides full development services on a user’s laptop.

What attracts so many developers to Cloud Foundry?

First, its openness, which has enabled developers to contribute and share source code under a lenient Apache 2 software license. With this announcement, VMware has introduced cloudfoundry.org, which provides a new system for managing open source software contributions.

Second, the richness of its environment. Cloud Foundry and SpringSource together have spawned a vibrant ecosystem, providing developers with the choice of a large number of languages, tools and frameworks. Several new partners announced support for Cloud Foundry this week, including X.commerce, which is using Cloud Foundry to build eBay’s next-generation, end-to-end commerce platform.

Third, the support for Cloud Foundry across multiple clouds and service providers, ensuring that developers will not be locked in to a single cloud. This week, VMware launched Cloud Foundry BOSH, an open-source toolset for engineering, deploying and managing large-scale, distributed services. BOSH already supports AWS and vSphere, and includes a Cloud Provider Interface (CPI) that will enable it to run on other public and private cloud infrastructures. By facilitating the operation of production instances of Cloud Foundry, BOSH will help VMware attract new service providers to its platform, who will in turn benefit from the pull of Cloud Foundry’s rich ecosystem.

So what’s next for Cloud Foundry? VMware would like to make the platform the “Linux of the Cloud”, based on an application portability layer that enables fluid movement of applications across multiple clouds. We see Cloud Foundry as a key enabler of VMware’s hybrid cloud strategy, which has progressed significantly since VMware first debuted its cloud initiatives three years ago.

We believe that to stand out from the crowd, a PaaS must be truly open, supporting a large number of developers, a rich ecosystem, and a wide choice of clouds and service providers. By this measure, Cloud Foundry is well on its way to achieving success.

EMC VSPEX: It’s All About the Channel

2012-04-13T22:54:28+00:00

I’ve been reading with some interest the various takes on EMC’s VSPEX announcement in San Francisco yesterday, many of which are putting VSPEX into the category of a converged infrastructure offering, something akin to EMC’s own VCE. But I think these characterizations miss the point. Yes, the initial launch of VSPEX features 14 pre-qualified hardware/software configurations designed to provide a full server/storage/networking and virtualization solution. And yes, the launch stage featured the backdrop of a row of VSPEX hardware chassis under the labels of large distribution partners. So the launch event had some of the look and feel of a VCE-like offering.

But I believe that VSPEX is really about enabling the channel. EMC Velocity partners – from large distributors to small regional resellers and integrators – that up until now have had to assemble and qualify converged hardware/software solutions on their own, can now start with a series of reference architectures and pre-validated multi-vendor solutions. There’s not a lot of choice at each layer yet, but it’s a start, and over time EMC has pledged to provide further options. EMC will be making its lab facilities available for partners to qualify and test additional configurations, so the means are in place to expand VSPEX offerings. The initial comments and “background chatter” I heard from EMC partners – including some international ones – were quite positive.

EMC has no doubt seen the success that its subsidiary VMware is having in building a highly productive, multi-tiered channel, and VSPEX demonstrates to me that EMC is taking its investment in channel development more seriously than ever before. The reported investment of 600 engineers and some $3 billion in VSPEX is a testament to EMC’s channel commitment.

The sweet spot for VSPEX is mid-market customers, not big enough to afford VCE, but savvy enough to know that it’s a heck of a lot easier and probably more cost effective in the long run to buy an integrated solution that’s been pre-qualified and tested, versus trying to buy and assemble the pieces themselves. In effect, VSPEX gives these smaller customers a faster and more sure-footed path to virtualization and the cloud.

VSPEX will help ease the complexity of the pre-sales cycle for EMC technology and channel partners, while in most cases providing an opportunity for a larger sale. Channel partners like to recommend integrated solutions that they know will work, without forcing customers into buying only a single flavor of each component, so from that standpoint, VSPEX is a step in the right direction.

VSPEX will certainly help EMC to compete more effectively with NetApp in channel-driven sales to small and medium enterprises, and in the process, accelerate sales of VNX storage. So all in all, the VSPEX initiative feels like a “win” for all participants. We’ll see how it all shakes out in the next few months.

EMC and Private Cloud/VMware Security

2012-04-04T22:52:52+00:00

Cloud computing desperately needs secure access, near-perfect availability and data transparency. But security is not the Golden Child of most cloud announcements. Many cloud customers simply assume security as a given... which it's not. I've spoken with many a Tier 1 cloud provider who bemoans the fact that outside of an IT sub-set, customers don't want to hear about the tough layers of physical and digital security in a top-notch data center. Yet this is precisely the message that they need to know.

EMC has stepped into this gap by insisting on the critical role of secure data movement and management in VMware-based virtual datecenters, or private clouds. They are quite right. Data is growing very quickly and access points and remote devices are growing right along with it. End-users demand pervasive access to data, and virtualized data helps to meet that demand. But along with that demand comes some very real -- and sometimes very dangerous -- security vulnerabilities.

EMC's Security Division RSA has made some strong strides in this area by combining security consulting, data management technology, and virtualization management. Their Cloud Security and Compliance consultants assess the existing security of VMware infrastructures and create plans to fill the compliance holes without sacrificing cloud flexibility and performance. (This includes the central infrastructure and also virtualized endpoints/desktops, which offer highly flexible computing but present potential security challenges.) Consultants use the RSA Archer eGRC Platform, which offers a central console for VMware security management.

This approach to security combines professional services with EMC, RSA and VMware technology Security integrates with compliance controls and data management tools, delivering a well-governed and highly trusted VMware private cloud. This is a security message the enterprise should find quite atttractive.

Acronis vmProtect – virtual edition of Acronis’s successful disk backup and recovery software

2012-04-02T16:02:40+00:00

Acronis debuted the vmProtect™ version 6 last year at VMworld in August 2011. Earlier this month, Acronis released version 7 of vmProtect further extending the capabilities of this Recovery Time Objective (RTO) centric OS-level image recovery Disaster Recovery (DR) software that takes timely image-based consistent snapshots with file system specific data and replication. The vmProtect is optimized for vmWare environments providing VADP-based agentless backup with Changed Block Tracking (CBT), hot-plug and direct SAN access.

Acronis’s rationale for introducing a solution that only protects vmWare environments is very simple; most small companies have only physical or virtual environments. The switch from physical to virtual happens very fast and once it takes place, these customers are looking for data protection and fast recovery their virtual environments.

The vmProtect can be installed as a virtual appliance into an existing ESX infrastructure or as a Windows agent on any physical machine to offload processing from your ESXi servers. It can be managed using any web browser or via the vCenter. Having seen and played around with the interface of vmProtect, I can tell you that a very simple yet intuitive GUI controlling this disk-to-disk-to-cloud solution.

The vmProtect simplifies the restore process for VMware environments. Typically, in case of a data loss event, one has to purchase or spin up a new VM host, re-install the ESX server on that host, then restore the VMs on that host. The vmProtect simplifies that into a one-step Bare Metal Restore (BMR) of the whole ESXi host, including binaries and configuration settings. Since most of the Small and Medium Businesses (SMBs) or branch offices of large organizations lack the IT staff or technical knowhow, it is imperative that the restores are simple. vmProtect’s ease of use and intuitive GUI that requires minimal learning curve is designed for these novice users in mind.

The vmProtect features an intuitive automated DR Plan which is a quick step one-page guide that tells the user step-by-step how to conduct restores. It is very easy to setup with just a few mouse clicks. Another neat feature is the built-in VM conversion which performs Physical to Virtual (P2V) migrations, saving tremendous amount of time when compared to building a VM from ground up.

Restores can be to the same or dissimilar hardware and are either image based i.e. running a VM from the backup (ability to boot a VM from its backup), Exchange recovery (entire data base, individual mail boxes, or email level), etc. or file based (file / email level). Restores can also be setup for Microsoft Exchange.

With the enhancements to vmProtect, Acronis has instilled itself as a formidable player among other data protection solutions in the marketplace. Their new offering takes timely image-based consistent snapshots with file system specific data improving efficiencies of backups and restores and data protection system performance while lowering bandwidth utilization costs.

Helping MSPs deliver better business outcomes and become more profitable

2012-03-27T18:11:35+00:00

In most companies, nightly data backups are often delegated to junior or entry level personnel who lack data protection knowledge or experience. Due to thankless nature of nightly backups, experienced IT personnel avoid volunteering for this job like the plague. Additionally, the high failure rate makes nightly backups responsibility unpopular. The lack of experienced personnel increases probability of failures for restores. This has become a persistent and troubling IT problem. It is but one of many when it comes to in-house data protection.
While in-house IT is protecting your data 365 times a year, there is a big difference between protecting the data and making sure it’s recoverable. Managed Service Providers (MSPs) offer data protection expertise built around effective best practices. Often times, the MSPs offer teeth in their contracts for failure or non-compliance as well as written guarantees for Recovery Time Objectives (RTOs) and Recovery Point Objectives (RPOs).

In today’s environment of economic uncertainty, more-and more organizations are choosing and opting for MSPs cloud-powered offerings. Other factors fueling the adoption of cloud-based data protection offered by the MSPs include data growth, longer data retention requirements, regulatory compliance mandates, and stringent data security requirements.
Cloud data protection services are experiencing phenomenal growth. A large majority (>75%) of cloud usage today is for data protection and archiving. Large players like Amazon & Google have demonstrated early success and have been instrumental in validating this business model. However, one of the critical things missing from their offerings is the Service Level Agreements (SLAs) on service availability. That requires higher level of knowledge and service delivery capabilities of the in-house IT.
In this stage of mass adoption of the cloud-based services and solutions, MSPs are looking to achieve better business outcomes by driving self- branded, high value, sticky, and profitable data protection services offering at a lower business risk and capital expense and, above all, shorter time to market.

Enter the realm of Modular Data Protection Services (MDPS), a service provider to MSPs/VARs, trusted advisors experienced in delivering better business outcomes and profitable data protection service offerings at a lower business risk, capital expense, and shorter time to market for MSPs. They drive cost and expense out of the MSP’s model as they grow their data vaults. They teach MSPs the business and technical knowledge required for successful service delivery.
The Modular 360° Partner Program™ provides business / service plan development, infrastructure, sales / marketing process / tools, deployment / support services, and flexible financial options to MSPs.
MDPS provides a variety of professional service options and skills to their MSP partners to help them deliver best in class data protection services and drive profitable quickly.

MDPS markets its services under the AiR™ (Assured Information Recovery) and DPaaS™ (Data Protection as a Service) branding. Their partners are free to extend this branding to their clients or have full freedom to market it under their own services branding. The MSPs retain their client ownership. The MSP’s end customer data resides in the MSPs vault within their custody and domain. Data access and security policies are set by the MSP in consultation with their customers. MDPS cannot access, view or manipulate the data at any time.
MDPS has the technology partnerships and vendor relationships to custom tailor the right cloud infrastructure as well as the data protection software and appliance solution for their MSPs / VARs.

Quantum introduces cloud-powered data protection

2012-03-26T23:54:05+00:00

Quantum announced its foray into cloud-powered data protection to complement the company's existing on-premise data protection portfolio, which includes DXi-Series disk backup and deduplication, vmPRO virtual machine protection and Scalar tape automation products. For a company delivering effective and efficient premise-based data protection for a while, the cloud-based data protection is the logical next step.

San Jose, California based Quantum provides data protection and big data management solutions. Their cloud-based data protection platform is all software and based on the Quantum vmPRO™ technology paired with a new virtual deduplication appliance, DXi V1000, which also debuted a couple of days ago. Additionally, the all-virtual platform powers Xerox’s recently announced cloud backup and disaster recovery (DR) services.

During the year, Quantum will offer cloud-based data protection to other cloud storage providers. The company also intends to provide its own branded cloud-based data protection service and a white-label version to be offered by Quantum’s channel partners under their own brands. Quantum also plans to develop software and hardware that can be used in a cloud-based archive.

Current off-site data protection options in the market force customers to choose between:

Very fast but very expensive storage array-based replication
Cloud-only solutions that can take days to deliver restored data.

Both of the above approaches also require additional hardware as data grows.

Quantum has taken a different path, reflecting the reality that disk and tape are both prevalent in typical enterprises, with a mix of physical and virtual servers, and that most customers prefer the hybrid cloud approach. Hence, Quantum’s cloud data protection platform provides:

Local copies of virtual machines (VM) and their data resulting in LAN-speed restores and allowing companies to meet their Recovery Time Objectives (RTOs)
Fast backup and restore in the cloud, resulting from up to 95 % data reduction and WAN-optimized bandwidth utilization for replication.
The ability to perform Virtual Disaster Recovery (VDR) by bringing up full VM images in the cloud, resulting in efficient and cost-effective Disaster Recovery (DR) and Business Continuity (BC).

Quantum’s new DXi V1000 provides target-based deduplication in a simple and affordable virtual appliance that protects physical as well as virtual environments. Quantum claims an impressive 95% data reduction.

With these new offerings, Quantum has differentiated itself from other data protection solutions in the marketplace by enabling a flexible public, private, or hybrid cloud solution. Their new offerings eliminate the drawbacks of the legacy cloud-based data protection solutions that are either cost prohibitive to most small enterprises or take too long to restore the data, thus missing critical RTOs and jeopardizing BC.

DDN Delivers Big Data in Small Packages

2012-03-26T18:29:31+00:00

DataDirect Networks (DDN) has introduced two new entry-level offerings in its SFA line of big data appliances, reducing the form factor by 50% and price by 40% versus current models. Based on the company’s Storage Fusion Architecture (SFA) technology, which has become a foundation for extreme-storage applications, the new SFA10K-M and 10K-ME offerings are notable for bringing big data management capabilities within the reach of smaller players in high performance computing, data analytics and other content-rich application markets. It’s what DDN aptly refers to as one more step in the “democratization” of big data, which is key to making the technology more accessible and affordable for companies in industries as diverse as media production and biotechnology, who may not even realize they have a big data challenge in front of them.

The news of dramatically smaller footprints and reduced-cost SFA entry points is not what we’re used to hearing from a company that is accustomed to extending the scalability and performance envelopes of big data applications. But in the long run, we believe this initiative will be just as important to DDN’s mission of enabling users of all shapes and sizes to extract significantly greater value from their information.

Integreon eView: Adaptive Coding with Review Expertise

2012-03-21T17:35:47+00:00

Integreon is offering adaptive coding with their eView 4.0 review platform. Adaptive coding and predictive coding (and other related terms) are in essence the same technology. Pairing the technology with Integreon’s process expertise is an excellent way to plumb the depths of both offerings.

Integreon’s approach to adaptive coding uses Latent Semantic Indexing (LSI) to automate pattern identification between text terms and concepts. LSI algorithms identify patterns and relationships existing between terms and concepts residing in unstructured document sets. Keyword searching requires a spelling similarity between query and keyword, but LSI operates on the principle that different words used in similar contexts might mean similar things. So if an attorney enters the word “stock” in a keyword search and allows for wildcard matches, he’ll get back files containing the word “stock” and also “stocks,” “stockbroker,” and stockings. (We’ll assume that “stockings” is non-responsive.) He’ll have to do additional keyword searches for “broker” or “trading.” In fact, meet-and-confer meetings frequently consist of brainstorming lists of keywords for exactly this process.

If another attorney uses a search technology founded on LSI and searches for “stock,” she can return results containing “stock” and variations that actually have to with financials and not lingerie. She will also get back “broker,” “dividends,” “trading,” and more that appear within the same context as stocks.
Let’s look at Integreon’s offering based on this technology. Adaptive coding lets hosted eView users quickly classify large document sets by concepts, prioritize the results and batch-code review documents. This highly automated approach enables highly accurate and very fast document review.

But just as importantly, adaptive coding is not the only quality that sets this review platform apart. eView is by no means the first or last eDiscovery platform to adopt adaptive or predictive coding. As valuable as adaptive coding is, Integreon has clearly differentiated itself with their Managed Review service option. Integreon offers expert eDiscovery workflow service throughout the entire EDRM, and their review consultancy is one of the best in the business. By combining eView and adaptive coding with Integreon’s deep workflow and review expertise, users can experience the best of review worlds.

Think All Public Clouds Are The Same? Think Again.

2012-03-16T13:12:51+00:00

One would think that the rationale to place data in a public cloud is universal. Not so. The variety of reasons chosen by users also points out the varying capabilities of public cloud services. The industry has a tendency to lump all public clouds in the same bucket. In fact, I’ve seen some presentations putting Box.net next to Terremark next to Rackspace. But, as the cloud market matures it is becoming clear that each offering is different, as some are designed with consumer-level features and are trying to move upstream, while others are designed with enterprise-level features and working to scale downstream.

On the topic of enterprise clouds, I spoke to two of Nirvanix’s customers recently, one just before the holiday season and one yesterday. Both conversations exemplified what was important to them and why they chose the Nirvanix public cloud service, the Cloud Storage Network.

The National Geographic Society HQ is located in Washington, DC and they just shifted to the Nirvanix public cloud. Until about three months ago the corporate system archive was kept on optical disk systems. This included all the source files for National Geographic issues, going back to 1888, source photographs, PDF files and a variety of other data used by writers, photographers and editors. The vast library of video files is kept elsewhere. Just a few years ago the IT group had migrated all this data from a tape-based system to optical disks and the migration had taken about one year. Then the optical disk drive vendor had gone out of business. Perhaps this was the right time to take a more strategic view and not buy another physical system to manage and maintain on their data center floor. IT was no longer interested in doing rip and replace of systems every three to five years and then spending a year migrating the data to the next system. They felt they were in the constant business of doing updates and data migrations, leaving them with less time for doing strategic work. This time they decided to choose a strategic solution that would free them from this IT bondage.

Dan Backer, Director of Infrastructure Systems at National Geographic, liked the idea of going to the public cloud so that his users could access the data from anywhere in the world. He would never have to do another data migration; would only pay for the storage he actually used; could free up space in the data center for other strategic initiatives, and never had to worry about data protection again. His research led him to Amazon S3, Google Cloud Storage, Microsoft Azure and several other public cloud offerings, besides Nirvanix. He chose Nirvanix for several reasons.

First, Dan highlighted that he liked the fact that the Nirvanix sales and support team behaved and interacted with him much in the same way as the enterprise sales team from his previous storage vendor. It was not a credit card oriented, self-service, “visit our support forum” type of business discussion.

The Nirvanix team understood his storage needs, the sensitivity of his business data and the importance of a true geo-diverse, global namespace (which will become more important to him in the future). He liked the fact that Nirvanix could create a private cloud on his premises, if and when he needed it, and that it could be federated seamlessly with the Nirvanix public cloud, to deliver a true hybrid cloud. He could pay by the drink. He didn’t have to pay additional charges for downloading data, bandwidth usage and enterprise support, as he would for AWS S3, for instance. He liked the fact that he could do a robo copy of terabytes of data and send the physical disks and tapes to Nirvanix for uploading via their White Glove Data Transfer service. And while he was paying for storage, his data would be protected automatically by Nirvanix, by keeping a live copy (not a DR copy) in another node, in another state. He also liked the fact that the Nirvanix cloud automatically grooms archived data so any silent data corruption over the years is corrected and he never has to worry about it. He knew that all data is encrypted in flight and at rest. He also knew that at some point in time he could enable the subscribers of National Geographic to access select files, without having to change or expand any system footprint in his data center. He would simply ask Nirvanix to provide global access through up to eight data centers spread around the world and that these eight nodes would, without any ifs ands or buts, look like one namespace. No other public cloud service delivered on all these requirements.

This is Dan’s first venture into the cloud. He is very optimistic and enthusiastic about the cloud and taking a measured and methodical approach to migrating archival data to Nirvanix. National Geographic is similar to one of Nirvanix’s other big media customers, NBCUniversal, who started with a 40TB archive and grew that to over 2 petabytes over the course of 24 months, uploading north of 100TB per month. We will do another pass at Dan in a year or so to gauge his experience. For now he is definitely one of the leaders in the enterprise shift to the cloud.

We also spoke with DRFortress, a managed service provider who operates out of a former large-scale Equinix data center in Honolulu, Hawaii. Fred Rodi, President of DRFortress, needed to expand his portfolio of managed services to include cloud storage. He had the choice of buying equipment from a major storage vendor and building a cloud on his data center premises but this would require a huge investment, hiring people to manage the infrastructure and training them, and dealing with all the maintenance issues. If he went down the public cloud path he would need to keep his customer’s data in CA or Oregon, given current offerings from Amazon and other cloud providers. But that would introduce huge amounts of latency for his Hawaiian customers and send his bandwidth charges through the roof. From Fred’s view, he was located in the middle of the Pacific ocean and ideally wanted to provide a local cloud on his premises but still have access to a public cloud on the mainland, but only if it all looked like a single cloud with a single namespace.

DRFortress is located very strategically approximately halfway between Asia Pacific and the West Coast of the US. For a number of customers in APAC, Hawaii could offer a lower latency alternative than having to go all the way to the West Coast. This midpoint location would be even more strategic if the cloud they used all shared a global namespacethat stretches across countries and data centers and pools all data into a single globally accessible repository.

As he researched the alternative solutions available in the marketplace he narrowed his choice quickly to Nirvanix. Why? Because Nirvanix is the only one that offered a true hybrid cloud that was fully managed as a service with usage-based pricing. In essence, this means he could create a private cloud in his data center using an exact architectural equivalent of the Nirvanix CSN and tie that to all eight locations of the Nirvanix public cloud. This would give him instant global presence and all nine locations would present a single global namespace for his customers. But the best part for him was that Nirvanix would manage the cloud instance in his data center and that he would continue to pay by the drink. To him this definition of “hybrid cloud” made infinite sense. Fred indicated that Amazon provided “security by ambiguity” and didn’t allow third party data center inspections and audits which he grew concerned about. Fred also took a look at other companies claiming to have a cloud storage service, such as Savvis, but found that they had no actual web-based object store, only raw disk coupled with servers in tightly bound cloud compute nodes.

Public clouds have come a long way in the past three years. An increasing number of enterprises have some of their data in one or more public clouds today. As the industry starts to mature it is becoming clear that all public clouds are not created equal. In fact, beyond the obvious there are vast differences between public cloud offerings. They differ in how they deal with multi-tenancy, what they mean by global namespace, whether or not they charge for bandwidth and data transfer, how they define a hybrid cloud, and whether they provide a fully managed service or just support a “cloud framework” that you deploy and maintain yourself.

Are you looking to include public cloud access in your strategic IT plans for 2012? Do the homework. While it all looks similar on top, the icebergs are right under the surface.

Scale-Out, DR and Data Protection for Small to Medium Virtualization Users

2012-03-14T06:02:08+00:00

With today’s announcement of a Veeam partnership, Scale Computing’s iconic scale-out primary storage solution takes on a whole new dimension. The joint solution gives Scale Computing customers access to one of the best and most widely used virtual infrastructure data protection and DR offerings – Veeam backup and replication.

The partnership itself looks great on paper: both companies target the already large and growing base of virtualization (primarily VMware and Microsoft) customers, and both are 100% focused on selling through the channel. Scale Computing provides hardware-based primary storage, while Veeam offers a complementary software-based data protection solution.

But the collaboration looks even better in practice. This partnership gives Scale Computing channel-driven access to thousands of Veeam backup and recovery customers, many of which will be looking to expand their storage capacity and/or performance to support a growing virtualized environment. On the flip side, Veeam will have an inside track at becoming the data protection provider of choice to Scale’s existing SMB customers, the majority of which are running VMware vSphere or Microsoft Hyper-V along with Scale Computing primary storage.

Most importantly, end users gain access to a well-rounded virtual infrastructure storage solution, including both primary and secondary storage functionality. This solution extends to disaster recovery as well, including VM-level replication to remote sites and failover control at a full VM level.

In my experience, not too many technology partnerships deliver a win-win-win scenario, but this one certainly does. Small to medium-sized virtualization users should take a closer look.

Zmanda’s Amanda Enterprise now supports Google’s Cloud Storage APIs

2012-03-09T17:17:34+00:00

Amanda Enterprise, Zmanda’s flagship enterprise-class cloud backup solution that has supported the Amazon Cloud Storage API for the last 2 – 3 years is now integrated with Google’s Cloud Storage’s open RESTful API. Zmanda’s code is engineered to support other REST-based APIs.

Sunnyvale, California based Zmanda provides affordable enterprise backup and disaster recovery in more than 60 countries. Zmanda’s technology is based on Amanda open source code which is the foundation of more than half a million data backup and recovery installations worldwide.

Google’s Cloud Storage debuted last year and already many technology providers / providers of enterprise storage solutions have integrated with Google Cloud Storage including Panzura, StorSimple, TwinStrata, Zmanda, and Gladinet. Google’s advanced storage infrastructure and global footprint of their network infrastructure provides a flexible, efficient, and economical cloud backup solution for offsite backups.

Organizations of all sizes can protect their corporate servers, desktops and applications by backing up their business critical data from these machines to the Google Cloud Storage offsite. Amanda Enterprise encapsulates each backup archive as an object and stores it on Google Cloud Storage. Data compression is done before converting the backup image to objects. Each image is reusable on its own from the cloud.

Each backup set can have its completely independent policy and backup location e.g. One backup set could be backed up to the east coast data center and another one to west coast data center.

Backup archives are created in open formats (e.g. ZIP64 for Windows and tar for Linux). This open architecture prevents vendor lock-in.

The backup data and archives can be stored on disks, tape, optical devices, or storage clouds such as Google Cloud Storage.

In addition to utilizing Zmanda’s recovery interface, users can also browse, read or delete their data objects through Google Storage Manager.

Today, regulatory compliance mandates and cost efficiency are top of mind for every for profit organization. Amanda Enterprise provides advanced cloud backup features such as location control, bandwidth throttling, and parallel streams for backups and restores.

Zmanda has differentiated themselves from other data backup offerings in the marketplace by adopting open source technologies at an attractive price. They sell their solution direct as well as through Service Providers / Resellers. Their price list is very simple and is posted on their web site. In this day and age, most vendors don't even want to discuss price early on in the sales cycle. Zmanda's customers know exactly what they're getting and for how much right from the start.