Scott LoweHigher Ed CIO, Blogger and Author

Read the full post at TechRepublic

Last week, I had the good fortune to be invited to attend Tech Field Day in Boston where a group of fifteen Tech Field Day delegates descended upon the likes of Data Robotics, vKernel, EMC, HP, and Cisco in order to be dazzled by the new storage and virtualization efforts being undertaken by these companies. In the last session of the event, the group of delegates met with representatives from Cisco and EMC (one of EMC’s representatives was “the other Scott Lowe” while I was simply a participant). These engineers were a part of the Virtual Computing Environment (VCE) program, a joint venture between VMware, Cisco, and EMC. The main product being sold and supported from this partnership is called a Vblock infrastructure package, or simply Vblock. In this article, I’m not going too deeply into technical particulars about Vblock except to say that Vblocks are intended to truly enable organizations to purchase, deploy, and support IT-as-a-service efforts without having to expend major resources tuning the environment.

Read the full post at TechRepublic

Read the full post at TechRepublic

Back in the days of the physical server, an administrator purchasing a server would size a server with plenty of RAM, disk, and processor for a server’s long life. In many cases, computing resources were over-provisioned in an attempt to make sure that potentially growing resource needs would not max out the resources of the new server. While over-provisioning resources cost a little extra money on the front end, upgrading a server mid-cycle tended to be pretty expensive.

How times have changed.

Read the full post at TechRepublic

Read the full article at SearchCIO-Midmarket

During the past couple of months, Westminster College has gone through two very different software selection processes. For many reasons — mostly related to budget, time constraints and integration needs — we decided not to go through the typical request for proposal-based procurement process.  In order to achieve our goals quickly and efficiently, we reworked our software selection process to guide us through two different scenarios.

Read the full article at SearchCIO-Midmarket

Session is starting… just met the other Scott Lowe

We’re actually going to start with lunch…

12:41 PM – Presentation is starting.  We’ll be talking about Unified Computing.

12:44 – Technology solutions to business problems.  Business focus, but there is a lot of back-end stuff that has to work in order for the front line stuff to work.  i.e. want BI?  You need underlying storage and architecture to enable that service.  Current fastest growing costs in IT are in the area of server management and administrative overhead.  Cisco is working on reducing these expenses.

12:46 – Power and cooling costs are also increasing; reducing servers = lower power and cooling costs.

12:47 – Virtualization takes it to one level, but also increases complexity, increasing operational expenses.  Other vendors simplify this challenge by adding a management layer and providing professional services.  Keep adding software on software to manage these challenges adds yet more complexity and cost.  Creates what Cisco calls an “accidental architecture”.

12:51 -Cisco sees an eventual migration from mainframe > mini > client/server > web > virtualization to cloud.  To get there, tie it all together in a way that’s easy to manage end to end.

12:52 – Today’s Datacenter = trusted, controlled, reliable, secure.  Or, go to a cloud provider to get flexibility, dynamic, on-demand and efficient services.  Instead, combine these goals so we have compute and control.

12:54 – A lot of the technology behind UCS is also found in other Cisco products, such as Nexus.  Unified fabric is one example.

12:59 – Cisco dismisses Gartner’s pessimistic outlook on convergence of data and storage networks as a controversial ploy to gain readership.

1:02 – Single server, very small environments aren’t going to benefit from this financially… only when getting to the rack level will this start to show promise.

1:04 – Legacy virtual switches – tied to the vswitch and not mobile.  So, some folks trunk every VLAN down to each server, thus negating a lot of policy decisions.  New: Cisco VN-Link.

1:06 – What does UCS solve?  UCS at first glance looks like a blade system.  Contains blades with mezzanine cards.  Servers slide in the front; connections in back.   There is a fabric interconnect “virtual backplane” ties to other networks – LAN, SAN (FC, etc).  This is also where UCS Manager runs.  Differs from other vendors with less cabling and simplified architecture.

1:12 – Wire once for bandwidth, not connectivity and then just allocate bandwidth to different services as needed.

1:16 – Virtualization scalability, reduced TCO.  I/O, CPU and memory are the three primary resource needs.  How do we take virtualization problems (i.e. app 1 needs to much I/O).  There becomes resource contention at the CPU to handle all of the needs (i.e. soft switch).  Make sure that the CPU is spending its time serving VMs, not infrastructure needs.  Solution: Hypervisor Bypass.

1:20 – Customers are running out of RAM in their virtual environments.  Xeon 5500 memory architecture – traditional: 12 to 18 DIMM slots (96GB to 144GB RAM per system).  Cisco developed a mux/demux technology supporting up to 48 DIMM slots, increasing RAM capability to 384GB.  Helps organizations better match resource needs to equipment purchases.  No more need to buy more complete servers just to get more RAM.

1:23 – Embedded unified management – Unified management domain with automatic discovery and dynamic provisioning of resources.  Traditional blades “are very stateful in nature” – i.e. BIOS revisions, etc.  Cisco’s blades are different.  All of it is centrally controlled… a “stateless server”.  You end up with a stateless bare metal server with no need to manage the hardware directly.

1:29 – The tool is a best effort tool… if it can’t manage a system for some reason, the management software lets you know.

1:40 – Quite a few questions in the room… been listening, so I haven’t been typing.  Sorry!

1:49 – Had to leave the room for a minute… back now.  Some specific questions are being answered about the impact of Cisco’s overall management philosophy.  Cisco is getting some good press on UCS.  Cisco believes in making sure that they can support UCS by hiring people from non-Cisco server-centric environments.  Where they don’t offer a service, they’ve partnered with others, including EMC, 3PAR, MS, VMware, HP, Oracle, SAP, CSC, and more.

1:53 – That was the UCS presentation…

1:57 – Cisco is seeing the most common replacement scenario be on the server replacement cycle.  The speaker was unwilling to identify the vendor they are displacing the most.

1:59 – Question: UCS is a lower margin market for Cisco than they are used to.  Can they sustain it and become am industry leader in the space?  Cisco has more than 2,000 people dedicated to UCS.

2:08 – It was pointed out that HP is happy to bypass a reseller and sell direct to a customer whereas Cisco won’t, so the VAR channel is happier with Cisco.

2:10 – Short break.

2:24 – Chris Naddeo is up.  Talking about multi-tenant environments.

2:25 – Challenges – individual tenants on a multitenant environment need to play nice with one another.

2:26 – Tenant defined: Was intentional ambiguous, but could be down to the VM guest level. SMT 1.0 is a reference design.  It’s an end-to-end design that ties together Cisco, NetApp and VMware products to ensure quality of service.  It’s a design – not a product with a SKU.

2:29 – Solution brief is 4 pages, overview is 25 pages, CVD is 90 pages, deployment guide is 100+ pages.

2:32 – There is no compliancy requirement… just a reference design.

2:34 – Now looking at the components of the SMT reference design.  Availability, secure separation, service assurance, management.  Material being shown in the screen is right from the design guide freely available for download from Cisco’s site. http://www.cisco.com/en/US/docs/solutions/Enterprise/Data_Center/Virtualization/securecldeployg.html

2:39 – UCS stateless stuff works better if you are booting from SAN rather than from local disk and this is included in the SMT reference design

2:44 – Cisco is targeting the service provider with SMT of which UCS is a key component. Was UCS developed for this/others/both?

2:49 – A key component of SMT is NetApp’s MultiStore.

3:00 – Ok.  I hate marketing lies.  Misleading, dishonest.

3:16 – VCE presentation – from “the other Scott Lowe” and Ed Saipetch

3:34 – Why VCE?  Help customers get to lower costs by moving to both internal and external clouds.

3:36 – VCE is integrated sales, engineering & support teams from Cisco and EMC/VMware.  This is their focus.  Engineer tested and validated packages.  Accelerated virtualization option.  They were asked if EMC has become more partner friendly… answer was yes.

3:40 – Scott Lowe indicates that he’s seen a major change in EMC’s stance on partnerships.

3:42 – Why would I buy a Vblock?  Answer is forthcoming.  Vblocks are actually being sold.

3:47 – The questions was asked: Who owns the Vblock from a task perspective?  Answer: Organizationally dependent and usually lands with the group that handles virtualization.

3:51 – Nuts and bolts of a Vblock – Design governance, compute, networking, storage, software

3:52 – Design governance – Three companies coming together to create a product.  Pretested, integrated, ready to go/grow.  Balanced configuration with regard to compute power, storage I/O, network capacity, virtualization, simplified deployment and management.

3:54 – Nuts and bolts – compute layer = Cisco UCS, VBlock Type 1 is 16 to 32 UCS B-series blades (128 to 256 cores, 960 to 1920 GB RAM) , Vblock Type 2 is 32 to 64 UCS B-series blades (256 to 512 cores, 3072 to 6144 GB RAM)

4:01 – Type 0 Vblock.  Smaller than a Vblock 1, but no specs beyond that at present.

4:10 – Network – 10GbE, FCoE and Fibre Channel, 2:1 oversubscription ratio for each chassis.  Vblock 1 (2) = 4 (8) GbE and 4 (8) FC uplinks per fabric; Predictable network oversubscription ratios min 4:1 to 8:1 max.  Need to have 10GbE upstream to connect the Vblock.

4:12 – Vblock type 1 uses either dual MDS 9222i or dual MDS 9506, 8 4 Gn N-port to each FI, 4 to 8GBE

4:13 – Storage – Type 1 uses CX4-480 with a blend of EFDs, FC and SATA sales to deliver from 43 TB, 41000 IOPS to 68 TB, 50000 IOPS.  Type 2, Symmetrix V-Max, 8 to 16 4Gbps fibre channel. 140 TB/92000 IOPS to 211 TB/140000 IOPS.

4:20 – Storage details are up on the screen.  I’ll post them later.

4:25 – Software = Vblock = VM container based on Enterprise Plus.  Nexus 1000V is used to manage virtual switching across the Vblock.  PowerPath/VE to enable path management.  Ionix Unified Infrastructure Manager (UIM) is strongly recommended.

4:43 – Ok… I really like this Vblock concept.

5:06 – Done, folks!  Listening too intently to keep typing!

8:48 AM – Day two of Tech Field Day kicks off with a presentation from HP.

8:49 – listening to Tom Joyce, VP of Marketing, Strategy and Operations for HP Storageworks.  Title of presentation: Storage and the converged infrastructure.

8:51 – Storage still has similar challenges as those faced ten years ago.  Today: Storage should be simple.  But being simple is really hard.  The market is tired to expensive, complex management tools.  The HP product line is still complicated.  Their challenge is to rationalize their storage and make it easier to understand.

8:54 – There are constant wars over who has the best storage for virtualization.  Tom feels that no other vendor has the end-to-end storage story that HP has, to include the network, servers, etc.

8:57 – Storage should be self-optimizing – capacity, cost, performance, protection, functionality.

8:55 – Storage should be infinitely scalable.  HP hits 16 petabytes now and they need/want to scale well beyond that.

8:57 – Storage should be software and very well integrated.  HP is taking a strong position in the networking space between their Procurve line and 3com acquisition.  Storage is a part of this.

8:59 – Storage should be converged.  Tom feels that the market is moving to HP’s strength.  Other vendors need to partner, but HP has all of the pieces.

9:00 – Tomorrow’s IT architecture – VROOM – virtualized, resilient, optimized, orchestrated and modular.

9:01 – HP will drive to their strategy through convergence across their product line.

9:03 – HP: Unified management for a single namespace.  Up to 16PB today. Non-disruptive growth.  Automated storage tiering.  Deployinh virtual storage with VMs.  HPP4000 SAN has integrated storage management for both physical and virtual Windows servers.  More usable space for apps with thin provisioning.  Cost reduction and async replication. Scales well and runs on industry standard architecture.

9:06 – Tom is now touching on EVA.  EVA was successful against others on architecture and simplicity.  The focus for EVA now is moving from storage virtualization in the box to outside the array.  HP is working with MS and VMware to create an ecosystem for VM customers.

9:09 – D2D backup and dedupe.  Tape is still out there and HP moves a ton of tape.  They just announced LTO5 and HP understands how people do backup.  They are going to do more with D2D and dedupe.  Saves money on networking and storage costs.

9:12 – Dedupe 1.0 to 2.0.  What does it mean?  How do you accomplish all parts of the dedupe effort with 3 products with a single product?  More details coming this year.  Will dedupe become a standard array feature?  No real answer from this on Tom.

9:15 – Key takeaways – HP is well-positioned to make the storage story work in a simpler way.  No surprise hearing this from the chief marketing guy.

9:18 – Concerns about the buzzword heavy presentation from Tom Joyce… all buzz, no meat.

9:20 – New speaker… agenda: trends, addressing today’s challenges, use cases, new features announced 3/29, HP X9000 portfolio.

9:24 – Former ibrix CEO is now speaking about where business is going… current scenario: data in unstructured silos.  More data is being shared with more people.  Applications are more granular.  Unstructured data will continue to grow.

9:28 – There is a method to the madness to HP’s recent acquisitions.  What the acquired companies have done for block and file level work is similar.

9:33 – Now jumping into the guts of the P9000.  Unstructured data needs and challenges: Massive scalability, performance that grows with capacity, management of large data sets, cost containment.

9:34 – Now up: Patrick Osbourne – ibrix deep-dive and we’re going to get a lab tour.  Slide title: A powerful combination: Scale-out and converged.

9:36 – Intent is to support HPC and grid-nodes.  Customers include Pixar, Dreamworks, investment houses, very large ISPs.  Can use any block storage as a data store.  X9000 supports single name space consisting of X9320s, X9300s, X9720s.  use allocation policies and data tiering to control data/file location.  Namespace can be added on the fly to include existing SAN/legacy storage with X9300 front ends.

9:41 – A new marketing term: Segment evacuation.  Migrate data from old storage to new storage.  A good marketing term that makes sense!

9:42 – Another decoder ring is being unveiled: ibrix fusion: filesystem product from ibrix.  Segments in a file system equate to a block device, usually coming from some kind of block-based RAID-enabled device.  File serving nodes/segment servers – responsible for making sure that single name space is mounted across all file serving nodes.

9:53 – More details about the IBRIX product.  We’re looking at a diagram of IBRIX’s non-hierarchical architecture, patent #6,782,389 right now.  Provides high aggregate performance from a single directory and single file.

9:58 – The data path is server to file serving node to storage… can this be geographically separate?  Not yet. but with CRR – Continuous Remote Replication – some semblance of this robustness can be achieved.  Latency has not yet been well tested.  IBRIX can tolerate some latency, but at some point, that latency affects HA.

9:59 – Where are the bottlenecks?  Depends on the workload… IOPS – pressure on back end disks; more storage and SSDs can help with this.  For streaming, it puts pressure on cluster interconnect or front-end network.  Infiniband can help.

10:01 – HP has a storage management framework that provides some performance details over time in order to determine event correlation.  They also have real-time stats (ibrix_stats) and ibrix top.

10:21 – Had to leave the room for a minute.

10:23 – Moving on to a tour of the lab

10:48 – We’re now seeing some use cases for HP’s X9000 storage after seeing a very, very cool lab.  Media and entertainment rendering is a major use case.  Health and life sciences (think DNA sequencing) is another use case.  Content depots (music, images, movies, books) are another major use case.  The goal is to keep cost of storage and its footprint down.  Manage everything from a single GUI to avoid siloing the data.

10:57 – Key takeaway: HP X9000 is a flexible, easy to manage solution purpose built for the markets they serve.

11:21 – End of presentation.

Read the full article at TechRepublic

How often are you told that you have to accept the lowest possible price on equipment that you need? If you have to work under this stipulation, are you really working to the best interests of the organization? Would a better tactic be to build a real relationship with a vendor that might include them becoming a trusted partner of your organization rather than just a reseller?

Read the full article at TechRepublic

3:40 EDT – starting a live blog from EMC headquarters

3:42 – Virtual storage discussion to help customers deliver IT as a service.  Virtual storage defined: Free physical media and physical location.  We started with individual servers, with VMware getting us to pools of virtual servers.  Once we get to servers and storage federation, we can seamlessly move applications between data centers without concern for underlying processing and storage resources.

3:47 – Next up: Data center virtualization.  True cloud stuff… who cares where the stuff is running as long as it has the resources it needs?  There is a major availability/DR component to this.  24/7 services.

3:49 – Question asked: In mobilty, are we talking about the mobility of the underlying data or are we talking about the LUN/storage block?  Answer was deferred to later.

3:52 – @storageanarchy is up now talking about how we can solve some challenges.  In cluster computing, write anywhere clusters is hard and is kept close by.  As we start talking about moving workloads, there are inherent technical limits.  Can’t write something simultaneously in NY and Singapore, so take that off the table.  So, instead, you can write to one place that knows about where all of the various blocks of data is stored.

3:56 – The solutions thus far has revolved around keeping people’s cache consistent.  What if the storage model instead was asking a central resource for write access and that resource handles the logistics.

3:59 – 95% of the I/O might be against 5% to 10% of the data.  Helps to answer the question of what is the most important data?

4:00 – The question was asked: Is this replication synchronous or asynchronous? The answer: It doesn’t really matter, but async presents more read challenges.

4:07 – Ok… the conversation has turned to what should be policies, not 100% technology-based… waiting for things to get back on track.

4:18 – I’ve lost the bubble a little… well beyond my normal use cases.

4:19 – The holy grail is a self-aware application that can define its own needs and locate adequate resources.

4:28 – Ok… this whole federation thing looks big, but we’ve been assured that it scales down.

4:30 – We’re now talking about education… They see a huge gap in skills in people coming out of college.

4:45 – Chad Sakac is up now.  He’s EMC’s VMware guy.

4:46 – Storage efficiency is a central topic in VMware-land, particularly given VMware’s focus.  Efficiency operates around capacity, flexibility and performance.  Do each of these for less dollars per X.  i.e. less dollars per gigabyte, etc.

4:48 – Changing stuff in storage is generally painful and expensive.  A technology that helps with one axis may or may not help with other axes.

• Capacity: SATA, Thin provisioning, density (i.e. 15 disks in 3U; future transition to 2.5″ disks, etc), automated tiering, data reduction (compression, data dedupe)
• Flexibility: Easy replication, IO modularity, Scale out, Multiprotocol, Local virtual storage, VMware-integrated, Global federation
• Performance: 10GbE/CEE (converged Ethernet models), solid state disks, mega caching, VMware hardware offload, auto-tiering

5:03 – I’m skeptical of some of the comparative information we’re getting with regard to the $/IOPS difference between solid state disks and more traditional disks.  I’ll follow up later.

5:08 – Storage systems should be responsible for hardware-assisting VMware, not the other way around.

5:12 – Today: What we see in VMware and traditional storage.  At the same time, people want storage as a service.

5:16 – On to a demonstration of another tool

5:23 – We just saw a fast clone demo.  Very cool.

5:24 – Now we’re looking at a compression demo that gains further advantage.

5:25 – Wait!  There’s more!

• Write Same/Zero – 10x less IO for common tasks
• Hardware offload locking – 10x more VMs per data store (stop locking LUNs and only lock blocks)

Day after tomorrow

• vStorage + FAST v2 integration – DRS and DPM for storage
• Long-distance and cloud migration – Vmotion between data centers at the VM level
• Core storage stack changes for storage virtual appliances – pNFS, VMDirectPath IO, changes in paravirtualized SCSI and vmkernel stack
• vDisk and VM array awareness – VM-object level awareness for block, file and add object-storage models to vSphere generation.

12:17 PM – arrived at Vkernel’s offices and got set up.

We’ll be meeting with Doug McNary, CEO; Alex Bakman, Founder & CTO; Ken Latimer, Senior Director of Program Management and Jonathan Klick, Systems Engineer.

12:21 EDT – Just told that there will be no PowerPoint presentation!  Wow!  This is good.  Vkernel folks are introducing themselves now.  A lot of experience in the room.

12:23 – Vkernel development happens in Russia; HQ is in Boston.  Business has 400 customers and grew 400% same quarter over last year.  Again, impressive growth.

12:24 – Vkernel is about optimizing resources.  Company allows free trial use of their tools.  Vkernel is usable within minutes of downloading and focuses on balance of cost and performance.  Struggle: How do you determine ROI when there are many applications competing with shared resources?

12:26 – VKernel’s CTO is up now and talking about the history behind the why of the VKernel product.  Multimillion dollar purchases were decided at the CIO level without much input from admins,.

12:28 – Framework-based management products were more difficult to implement than the problems they were trying to solve.  The idea is that “one app solves one problem”.  Download, deploy and begin seeing value immediately.  The Windows admin people grew up in a 1:1 world – 1 app, 1 server.  Led to server proliferation.  Capacity management in this arena wasn’t an issue.  As long as the box was sized correctly, all was well.  Now, Windows admins are  moving into a very shared world – VMware.  Now, Windows admins need to care about conflicting with others from both the resource and storage level.

12:32 – VDI increases this need. A desktop introduction requires immediate use of resources for that user.  Capacity management has become much more complex and even more critical than ever before.  The virtualization layer has changed the paradigm.

12:33 – Performance issues are linked directly to capacity constraints.  People have massive servers and can’t take advantage of them due to I/O constraints on the storage side of the equation.  Vkernel is providing tools that show administrators in easy-to-understand ways where they’re seeing bottlenecks – in a VM, between VMs, in a data center container.  There are also different kinds of bottlenecks – CPU, RAM, storage capacity, storage I/O.  VMs also grow legs – Vmotion.  That Vmotion action creates great fluidity in the capacity management game.

12:42 – Vkernel sees IT “swiss army knives” – people try to do a whole lot with just a few people.  It makes it hard to really manage capacity.

12:43 – Which VMs are the culprits creating all kinds of performance issues – particular related to storage I/O?

12:47 – An end user is sharing is 1,000 VM story – he uses Vkernel to help make decisions about what to do with their environment.  The ROI was very quick.  Once they proved the value, they immediately rolled it out company wide.

12:49 – Product can be installed in just 20 to 30 minutes.  It’s a 1 MB challenge.

12:50 – Don’t wait for end users to scream before addressing capacity issues.

1:05 – Just deployed the free Capacity View tool in my environment at Westminster College.  Very eye opening and actually validated what I had already assumed to be a problem in our vSphere environment.

1:11 – Capacity View gives the high level information.  Capacity Analyzer takes it deep to help show where problems are happening.

1:14 – Someone just asked if this was going to be extended to Hyper-V.  It already with with vSphere, Hyper-V and Citrix.  Not all metrics are available due to different collection methods, but everything available in System Center is exposed in Vkernel.  System Center is required as the collection agent.

1:17 – Now on to disk I/O… drill down by data store.  Backtrack… drill into a VM with a memory problem.  The tool provides specific recommendations including how much should be allocated to address the issue. Vkernel provides specific metrics that led them to make a particular recommendation.

1:25 – Some specific use case questions being asked right now.

1:31 – Vkernel is all about predictive analytics, too.  Based on historical/current consumption, what more is possible?  More importantly, why can’t I add more?  What’s the constraining resource?

1:36 – We’re taking a short break.

1:45 – Starting up again… the Vkernel optimization tool helps both upsize and downsize VMs as necessary in order to achieve overall goals.

1:47 – They will provide specific recommendation to help appropriately size VMs.  The company helps to reclaim underutilized resources but also makes recommendations on where to increase resources.

1:50 – A “safe assumption” – vKernel vNext will provide a way to click a button to implement optimization recommendations.

1:53 – The optimization tool also outlines VMs that have been powered off for a long period of time as a way to identify abandoned VMs.  Snapshots get the same treatment.  Tool also helps identify VMs that are just idling.

1:57 – Vkernel is licensed on a per-managed socket basis. Product scales to a point.  Largest implementation so far is about 2000 VMs.  Today, there is no single pane view beyond that limit, but that is being considered for a future revision.

2:02 – Limited support for Hyper-V in optimization pack.  Full supported is expected by the end of the summer.

2:04 – End of presentation… now on to roundtable for attendees.

Start posting @8:16 AM EDT

8:17 – Drobo CEO Tom Buiocchi is talking about Drobo’s history and current product set, which now consists of five products, Drobo, Drobo S, Drobo FS, DroboPro, DroboElite.

8:20 – Tom is talking about Drobo business customer feedback indicating that customer feedback has revealed more than $7500 average savings, better backup reliability and more.  Definite success.

8:21 – now discussing future focus. Will stick with the $500 to $15K market space where there is billions in opportunity since the big guys ignore the space.

8:24 – Tom was asked about his perception of their place in the market.  He cautions people to compare apples to apples and make sure to price a fully populated solution rather than just a bare chassis.

8:25 – He was just asked if the company had thought about targeting the VMware enthusiasts and asked the group for feedback on how to make that happen.

8:26 – Tom was pushed back a little on pricing for the VMware enthusiast.  It was pointed out that other companies provide a raw chassis at a much lower cost.  It was suggested that home enthusiasts already have a bunch of SATA disks laying around that they’d want to use.  It was also pointed out that VMware enthusiasts need higher end storage features in order to learn.

8:31 – Someone asked about HyperV support in Drobo products.  Win 2003 uses SCSI2 reservations while Server 2008 uses SCSI3 reservations not yet supported by Drobo, but it will come in a firmware update.  ESX uses SCSI2 reservations, so there is full support.  This primarily affects just Live Migration, so the Drobo is still perfectly usable for a single server HyperV lab environment.

8:34 – The new Drobo FS does support jumbo frames, but does not yet provide iSCSI support.  NFS is supported in the new Drobo FS through a free app.

8:35 – It was asked if the Drobo FS will eventually support iSCSI.  Drobo indicated that there are no plans to add iSCSI support to Drobo FS, but there was a suggestion to add it to the roadmap.  A Drobo engineer pointed out that the Drobo FS is intended to be an application platform.

8:37 – Drobo apps will have a management process to allow them to be installed and configured.

8:41 – Drobo was pushed back hard on lack of native NFS in Drobo FS.  NFS is available through an add-in app but there is concern about the quality.  After all, this is storage and a company’s lifeblood.  Drobo CEO Tom Buiocchi asked if some sort of certification process would help to alleviate the quality concern for more mission critical needs.

8:45 – Drobo was asked who their real target market is.  It’s the “prosumer”, SOHO, SMB, not the consumer.  It’s not intended to be for the consumer.  It’s intended to be the customer that doesn’t want to spend his time building white box RAID kits.

8:46 – It was asked if the Drobo Pro will be made VMware ready.  Answer:  It’s VMware ready on ESX 3.5, but not vSphere.  The Drobo Elite is vSphere ready and that’s where Drobo feels people should focus. As expected, the company was pushed back on this decision.

8:52 – We’re taking a short break while we wait for the next speaker to begin.

8:53 – The next part of the presentation is entitled “Enhancing vSphere with DroboElite”

8:55 – Some Drobo products have a great upgrade path – disk packs can be migrated between products, but not in every case.  Drobo has a knowledge base article on the topic.

8:56 – Demo is starting.  We’re going to be shown how to hot add storage and LUNs on vSphere.   We’re also going to do a BeyondRAID deep-dive.

8:59 – BeyondRAID virtualizes disks in a pool and is dependent on the number of disks.  If there are two 1 TB disks, they will simply be mirrored, but Drobo doesn’t preinitialize the whole disk, but only what’s needed as files are written.  When a third disk is hot added, nothing happens immediately.  The goal isn’t to just RAID 5 the third disk, but to protect the data, which is already taking place on the first two disks.  After a third disk is added, the next storage blocks are written across all three disks rather than just two.  This is how BeyondRAID can use disks of any size.

9:03 – Think of BeyondRAID as just a whole bunch of mini RAID sets.

9:04 – BeyondRAID helps users avoid unintentional damage by not allowing the full use of the first disk of larger size than all of the others, but as soon as a second disk of the larger size is added, that space can be used.  BeyondRAID won’t allow users to use space that can’t be protected.

9:08 – There is some misunderstanding of how BeyondRAID works.  The idea of BeyondRAID creating a bunch of mini RAID sets freaked people out just a little as it was considered a potential management nightmare, but BeyondRAID manages all of that on the user’s behalf so those mini RAID sets aren’t exposed.

9:11 – BeyondRAID knows about underlying data, such as NTFS volumes, EXT3, HFS+ and even FAT32.  They recognize the file system layout and as block free up, those blocks are placed back in the pool of available storage.

9:13 – Drobo can support other file systems beyond those discussed with the only impact being the loss of capacity lights on the unit.

9:14 – Healing – when a rebuild takes place, only the used disk space is rebuilt, not the whole disk (including empty space as in RAID).  Good deal!  Much faster rebuild.

9:18 – Questions were asked regarding the possibility of digital forensics and Drobo.  Stumped the presenter who will take the question back to the team.  Now moving on to a demo of a unit.

9:20 – Thin provisioning = good.  Drobo supports it.

9:21 – All Drobo products support BeyondRAID.  Simple LEDs identify drive and unit status.  Unit is managed via Drobo Dashboard.  Drive’s don’t need to be in any particular order or bays.  Can skip bays, etc.  Mainly means that the unit is more idiot-proof.

9:23 – In order to receive alerts from a Drobo, must run the Drobo dashboard, which runs as a service on a management PC.  There is internal discussion about including a mini web server.  There are also internal discussions regarding the addition of a second power supply to some of the Drobo products.

9:24 – For the current target market/price point, second power supply and other higher end features aren’t high priority, but are under consideration.

9:27 – Demo: When the Drobo hit 90% used space, LED turned yellow.  Since this is what is supposed to happen, that’s good!  A new disk was simply shoved in a slot and the space become available, thus eliminating the yellow light.

9:28 – They failed a disk, which placed the unit back into a yellow state.  Replaced the failed 320GB disk with a 750GB disk in the middle of a rebuild and everything worked.

9:23 – Must have misread my clock… either that, or went back/forward in time.  It’s now 9:23, not 9:28.

9:29 – We’re now getting a demo of a Drobo employee’s lab.

9:31 – DroboElite – each target = 1 LUN.  DroboPro – each target = up to 16 LUNs under the target.

9:35 – Updates will be less frequent for a bit… watching demos which aren’t that interesting unless watching.  Will post as interesting tidbits arise.

9:37 – Drobo Pro is not a NAS device – not a file sharing device, so no CIFS.  Block level storage only.

9:39 – Has Drobo considered making their dashboard a plug in to VI?  Yes they have.  Will Hyper-V enjoy a similar level of support in the future?  Yes.

9:40 – DroboElite does not support IPsec.  The iSCSI implementation will be improved over time.

9:43 – And now on to the Drobo FS demo.

9:46 – Drobo FS uses cases: Shared file storage, network backup, private cloud.

9:47 – DroboApps available at launch: iTunes media server, UPnP media server, BitTorrent client, web server FTP server and more.

9:49 – Drobo’s Oxygen Cloud turns Drobo FS into a secure, private cloud with ability to support office PCs, mobile devices and cloud-based backup.

9:53 – Drobo CEO just indicated that the best feedback he’s heard all day is that there needs to be a DroboApp certification process.

9:56 – A lot of discussion about the serious need for DroboApp certification process.

9:58 – Devices can remotely access the Drobo FS.  Drobo Share won’t get Oxygen Cloud.

10:01 – Currently, Drobo Elite internal architecture doesn’t allow DroboApps to work.  Yuck!  DroboElite would be a great use case and might be added in a firmware update.

10:02 – A blogger indicated that choosing an unknown partner organization for Oxygen Cloud didn’t help Drobo much.

10:19 – Q&A now.

Read the full article at TechRepublic

Westminster College’s IT infrastructure has grown throughout the years in a very organic way. In the beginning, it was only a single VAX terminal connected to the University of Missouri; now it’s a fiber-connected network with thousands of nodes, dozens of servers, and terabytes of storage.

Read the full article at TechRepublic