Around The Block Blog

Compellent Technologies

Live from C-Drive 2010: Brian Bell Talks About C-Drive 2010

by Compellent Technologies, — May 06, 2010

Brian Bell, Compellent Vice President of SalesCompellent executives are excited about this year’s C-Drive 2010. Brian Bell, vice president of worldwide sales for Compellent, talks about the conference and highlights one of our presenters, Tom Dau, who presented on Oracle scripting.



Compellent Technologies

Live from C-Drive 2010: Why Unified Fabric?

by Compellent Technologies, — May 04, 2010

Mike Dreitlein, Senior Product Line Manager at Cisco, takes us through unified fabric with FCoE and how FCoE fits into the bigger picture for the data center.

1:23 pm – Our customers understand network consolidation intuitively – they’ve been doing it for a long time. 10Gb Ethernet is a ‘real’ technology that has market momentum.

1:26 pm – FCoE is dominant in the storage environment – it has tremendous vendor support; now we’ll talk about integrating it into the SAN.

1:18 pm – The ultimate goal for unified fabric is the ability for SAN and LAN traffic to be on the same line at the same time. In the long term, what we can see as goals is that we can apply the roadmap and scale of Ethernet to SAN. There are benefits to moving onto the Ethernet roadmap, including the ability to wire once in the data center – you can envision one port, one adapter, one way of doing things that will come out of the server that will do SAN and LAN. Having a standard I/O in that fashion will allow for faster applications.

1:31 pm – It might surprise you that we have a number of products on the market now for unified storage. First product to deliver FCoE was the Nexus 5000 10GbE LAN, first available in June 2008. A wildly successful product that’s used by about 2/3 of our customers for LAN, but about 1/3 of our customers also buy an FCoE license for SAN. Second is Nexus 2000, a remote line card for Nexus 5000. The latest version is shipping to customers this week, a 10GbE 2232, which supports FCoE. Third is Nexus 4000, a blade server switch for Blade Center. The fourth is the UCS system itself, which incorporates FCoE into the I/O subsystem.

1:39 pm – Why FCoE?

  • Because it integrates with today’s FC SANs
  • Because it enables unified fabric
  • Because it brings more customer advantage than FC
  • Because vendors can build better products with FCoE.

1:40 pm – How is FCoE enabled? They key is the Ethernet packet – a fully formed FC frame with no segmentation allows enabling of FCoE with a simple encapsulation/de-encapsulation enterprise. The people that standardized FCoE are the same folks that constructed FC standards. Some say the standardization isn’t done – but there’s always been an evolving element to standardization over the years.

1:43 pm – We also need something else to enable FCoE – CEE (aka DCB) provides FCoE Ethernet support. CEE provides industry-agreed features for: lossless and lossy Ethernet services on same link, bandwidth prioritization and management, and communication and feature negotiation between devices.

1:47 pm – Advantages of FCoE: easy to understand, same operational model as FC, same techniques of traffic management (in-order delivery, FSPF load balancing). Essentially, FCoE is an evolution of FC and they are the same thing: FCoE is managed like FC at initiator, target, and switch level.

1:52 pm – Many of our customers have a lot of connections coming out of servers – it’s typical to have 8 connections. This can cause problems with cabling, air flow.

1:53 pm – Cisco “eats its own cooking” – has implemented FCoE with Cisco Nexus 5000 – freed up rack space, enabling a 30% increase in server density. Also saving on cabling costs, reduced by 40%. Saving 1/3 on power and redeploying to what the data center is really all about – servers. This extends the life of the data center 12-18 months.

1:56 pm - FCoE doesn’t stand alone as just one idea – it integrates with a lot of other things we do in the data center. It’s a component in the thrust forward to improve data center efficiency.

1:57 pm – FcoE brings more to the table than FC – it invites more user choices, it enables FC to become more accessible, and offers better TCO for blade servers.

Compellent Technologies

Live from C-Drive 2010: Corporate Pitch - The Future is Fluid

by Compellent Technologies, — May 04, 2010

10:15 am – Bob Fine and Russ Taddiken asks the audience what they want to hear about in this session. Results include information about Data Progression, the development team, Fluid Data architecture and more.

10:23 am – Bob explains how Compellent came up with the figure that Fluid Data applications (thin provisioning, automated tiered storage and others ) slash storage costs by up to 80 percent. Thin provisioning can cut costs by 50 percent, and 80 percent of the remainder will sit on lower-tier drives. A combination of other technologies reduce costs even more.

10:25 am – Once VARs start exploring pain points with prospects, they may uncover difficulties they didn’t realize they had. Everything Compellent knows about modern compute technologies is being applied to the storage solution, which will result in better utilization and performance.

10:28 am – Bob explains that the majority of Compellent customers use automated tiered storage. Data Progression looks at metadata every day to determine which tier each individual block of data should be on.

10:30 am – Russ answers a question from the audience about whether the figure that 80 percent of data is inactive comes from Compellent or from the industry. Russ explains that it’s a Gartner figure and many of the VARs in attendance agree.

10:32 am – Bob says that about two-thirds of customers use Data Progression. Once customers go above 20TB of data, the percentage of customers using Data Progression goes way up!

10:35 am – An audience member asks about the performance hit after data is infrequently used and moved down to a lower tier. Russ explains that even though pages move between tiers, the volume stays where it is. With Compellent, the only changes in performance occur with those pages, and the performance of the rest of the volume is not affected.

10:38 am – An audience member asks if MLC flash is a good fit for lower tiers of storage. Russ answers by saying that SLC flash is perfect for tier-1 storage. There is an opportunity in the future for MLC to provide the ideal write capabilities for lower tiers of storage.

10:45 am – Bob points out that the PCI slots on the back of our SAN are available to be used. Competitive products bolt over those slots, and are more rigid about upgrade policies. In the end, those products have fewer ports than Compellent, and don’t let customers upgrade as much over time.

10:52 am – Units shipped in 2005 are still in production today using our latest technology. A lot of Compellent’s competitors have gone through multiple generations of products since then.

10:53 am – Customers should be sure to future-proof their technology by ensuring they implement solutions that are flexible, scalable, upgradeable and agile.

10:55 am – VMware users voted that Compellent was the easiest to use storage on the backend.

10:57 am – 100 percent of Compellent customers use thin provisioning.