Yesterday SearchStorage ANZ published excerpts from a document we are almost certain is an EMC competitive intelligence assessment of 3Par and its products. Today, 3Par's Marc Farley responds with its analysis of EMC's assessment.
As far as ease of use goes, we are happy letting customers decide which is easier.
EMC says everybody has thin provisioning and that these offerings are equivalent to 3PAR’s.
Yes, most vendors have a Thin Provisioning feature, but they are not equivalent to 3PAR’s. 3PAR TP does not require capacity reservations or special pools – including the snapshot and remote copy space that customers need. Creating thin or thick volumes is identical, including using the exact same disk drives.
EMC says 3Par’s ‘Thin-Built-In’ messaging “is actually not ‘built-in.’ It requires careful coordination with host software (Windows Server 2008, VERITAS File System) to work,” then points out that EMC can do the same thing with its PowerPath Migration Enabler and Open Replicator / Sparse Volumes.
Thin, zero detection technology is designed into 3PAR’s storage co-processor. Host software can be tightly integrated as it is with VMware’s new VAAI and Symantec’s File System APIs – which makes it very easy to use, but 3PAR’s thin tech was also designed to work with native host software commands, such as those found that natively exist in Windows, Server, Linux and even Oracle ASM. In most cases, it involves a single host command. The ability to thin volumes live, online is something EMC does not have and the performance of their functions pale next to ours (the advantage of HW integration). Not close.
“No flash or 1TB SATA drive support”
Ignorant on both counts: 3PAR has both.
“3PAR’s “tierless” storage approach leads to performance plateaus. This means that 3PAR hits a certain performance point and stays flat.”
All storage has limits and bottlenecks. 3PAR systems do too, but at much higher utilization levels than EMC’s. Where are their benchmarks to substantiate this anyway?
“Forces customers to add more drives which in turn undermines 3PAR’s TCO claims”
Most customers need to add drives at some point – with any vendor. 3PAR customers take longer to reach that point because of our high utilization rates and thin technologies. When drives are needed, it’s relatively easy with 3PAR and awful with EMC due to the need to rebalance pool resources. They say it’s not necessary, but that’s only if you like running everything at the same RAID level – like IBM’s XIV.
“3PAR designed for large pools / wide-striping using no or few tiers”
3PAR doesn’t have disk pools – period. But we can have many tiers defined by drive class and RAID type. Try mixing RAID types in an EMC pool.
“EMC can configure large pools for wide-striping as well—no advantage for 3PAR”
EMC pools form a single tier. If you want all your storage – including snaps and remote copy to have the same RAID level then a single pool can work. Most EMC customers split their drives into multiple disk pools for different purposes – that makes them significantly “less-wide”.
“No pro-active sparing to avoid performance penalty and risk during re-build”.
Wrong, our arrays (and most everybody else’s enterprise arrays) do this every day somewhere in the world.
“3PAR claims faster rebuilds but EMC avoids rebuilds altogether in most cases”.
Compared to what? Not using predictive failure technology? EMC customers have to deal with rebuilds too.
“No RAID6. 3PAR only offers RAID 0, 1, and 5. How will 3PAR handle multiple drive failure?”
Wrong. Not only do we have RAID6, but it’s implemented in our co-processor so it’s much faster.
“Increased risk of data loss and performance penalty during rebuild”
Nope - and the risk of data loss is lessened by the performance of rebuilds, which is done with hardware assistance over more disk drives involving smaller amounts of data on each disk.
“3-lane PCI-X controller node architecture. Prevents ability to upgrade to faster front-end connectivity (8Gb FC, 10GbE, 10Gb FCoE)”
Yes, and yes. But our controllers can be “supersized” with ports – up to 16 per controller. That’s 64 in a 4 controller array and they are all full active/active. There is bandwidth to spare.
“No 5-9s Availability”
Wrong again. Designed in and field data to prove it.
EMC says 3Par possesses only 170 engineers and is therefore slow to develop products, citing the fact that it took them six years to upgrade its platform from S-Models to T-Models as a sign of 3Par’s poor ability to develop new products.
And just when is that FAST tiering technology that EMC talked about all last year going to be released? 3PAR already delivered theirs this spring. We will also beat them out the door with VAAI integration. To be fair, EMC has many products to support, which dilutes their focus and adds time to their release cycles.
EMC also feels 3Par's decision to outsource support may have backfired, saying the results are “Spotty … in terms of guaranteeing consistent support coverage & quality ...”
Stating that 3PAR “outsources support” is just plain silly, especially coming from a company that keeps most of the worlds’ largest offshore outsourcing companies in business. Like EMC, 3PAR uses Third Party Maintenance suppliers (TPM’s) for break-fix field activities. In some geographies, EMC and 3PAR even use the “same” TPM. But EMC also outsources most of their volume Call Center and Level-1 Technical Support to offshore suppliers. Not so at 3PAR. Everyone that touches a 3PAR support case is a 3PAR-badged employee. I challenge EMC to identify a single outsourcing company that handles 3PAR technical support. EMC’s outsourced technical support sub-contractors could be listed alphabetically, by geography, or by technology category … but you’d need a couple of sheet of paper to do it.