This is an excerpt of an email that was sent to the management of a company that I'm working with. They had spec'ed F85's for a fairly low performance 10-20Mbits/sec. environment. The main requirement is reliability and ease of use. How does one respond to blatant misrepresentation of a competitors product?
barry
A few highlights of the comparison:
- The F85 is a stripped down, single CPU, low end device with multiple points of failure and a very poor data protection. NetApps service is rated very low by industry experts and they offer next day shipment of parts that the user must install themselves.
- The IP4700 is fully redundant, multi-CPU, mid-range device with no single point of failure and hardware based RAID 5 data protection. EMC's world class customer service center has ranked #1 for 6 consecutive years by Gartner Group. Standard 2 year warranty guarantees 4 hours ON SITE w/Parts service by EMC technicians. Plus, our 'call home' proactive maintenance system monitors trends within the system and reports them automatically to our customer service center. Often, EMC technicians will repair a system BEFORE the component actually fails.
The 4700 will be configured with 8 drives usable, plus 1 drive for RAID 5 parity and 1 drive for hot swap redundancy. It is scalable all the way to 7000GB (7TB) vs. only 648GB for the F85. The 4700 as configured above is 8RU.
In general, the F85 does not scale sufficiently for growth, has poor customer service behind it, has no redundancy, has multiple points of failure, and utilizes a sub-par RAID 0 data protection scheme.
The F85 is a lower end solution, true. But poor data protection? Not hardly.
As for support, here's a real life example for ya....
Sometime over a weekend we had a disk fail (turned out to be Saturday afternoon around 4:00PM). The data was rebuilt on our hot spare and a replacement was at our S/R dock before I came in on Monday. No data loss, no issues.
I have had issues with tech support before, mostly costing me time before I got to the 2nd level engineer who understood the problem I was experiencing. But that will happen with anyone, guess who EMC has manning their first line of support? People who couldn't hack it at NetApp...)
The fact is, NetApp is the leader in NAS for a reason. Take what you will from that.
~JK
Barry Lustig wrote:
This is an excerpt of an email that was sent to the management of a company that I'm working with. They had spec'ed F85's for a fairly low performance 10-20Mbits/sec. environment. The main requirement is reliability and ease of use. How does one respond to blatant misrepresentation of a competitors product?
barry
A few highlights of the comparison:
- The F85 is a stripped down, single CPU, low end device with multiple
points of failure and a very poor data protection. NetApps service is rated very low by industry experts and they offer next day shipment of parts that the user must install themselves.
- The IP4700 is fully redundant, multi-CPU, mid-range device with no
single point of failure and hardware based RAID 5 data protection. EMC's world class customer service center has ranked #1 for 6 consecutive years by Gartner Group. Standard 2 year warranty guarantees 4 hours ON SITE w/Parts service by EMC technicians. Plus, our 'call home' proactive maintenance system monitors trends within the system and reports them automatically to our customer service center. Often, EMC technicians will repair a system BEFORE the component actually fails.
The 4700 will be configured with 8 drives usable, plus 1 drive for RAID 5 parity and 1 drive for hot swap redundancy. It is scalable all the way to 7000GB (7TB) vs. only 648GB for the F85. The 4700 as configured above is 8RU.
In general, the F85 does not scale sufficiently for growth, has poor customer service behind it, has no redundancy, has multiple points of failure, and utilizes a sub-par RAID 0 data protection scheme.
My story about EMC is rather different, I guess.
Some local EMC reps scheduled a visit with my boss and came to visit us a while back. I wore not one but _two_ NetApp shirts that day - the very cool engineering shirt (black pre-F630 shirt with Mothra on the back) and the snazzy denim Land's End shirt with the embroidered logo on it. The night before I printed a copy of the recent INPUT survey, highlighting the part where I'm quoted (anonymously), and left it on my boss's chair. He brought it to the meeting.
They arrived. I smiled and shook hands and did the business card thing. We sat down at the conference table. They asked their questions, and I politely answered them. It went something like this:
"What about reliability?" they asked.
"I've been running NetApps since 1995, and to the best of my recollection I have never once lost data due to hardware or software failure," I replied.
"Yes, but what about uptime?" they inquired.
"alexandria> uptime 3:32pm up 187 days, 14:10 1162740653 NFS ops, 129052586 CIFS ops, 0 HTTP ops colossus> uptime 3:32pm up 171 days, 8:48 1692312427 NFS ops, 0 CIFS ops, 0 HTTP ops", was my reply.
"Expandability?" they suggested.
"I have 21 free slots for more disks - I bought the box with extra shelves and a 2nd controller so that I could grow it on-line without any need for downtime. I can grow our box to > 1TB without even thinking hard."
"Performance!" they pleaded.
"My Oracle DBA is happy, and says he can't see a difference between the 760 and a locally-attached dual-loop A5200. Besides, with a head-swap we can trade up to the 800-series, or cluster, or trade-up and then cluster, and I've got all the slots I need to add extra Gigabit cards..."
"Yes, but our 'phone home' support feature and customer service are second to none!" they intoned.
"Actually," I said, "your story about the hardware tech showing up on a Saturday with a replacement disk for a demo machine reminds me of a similar incident... one day a box arrived unexpectedly from Netapp with a replacement drive in it, which puzzled us because we hadn't reported a failure. But sure enough, we walked into the server room and one of the disk canisters had a red light on. Five minutes later we swapped it out and that was that. Sure a lot cheaper than sending a field circus guy out to do a job that any basic systems operator can do."
(Ouch, that stung a bit, but hey, I _like_ playing with the hardware, and if I have one complaint about the reliability of my filers it's that after I bolt it into the rack and plug it in, I don't get to fool around with it much. That's a blessing, of course. :-)
"And as for service, a funny thing happened in the middle of the night a few months back, when I was innocently rearranging some power cords to balance the load on a split-phase UPS, oblivious to the fact that the 'autosupport' emails were being sent out. Imagine my surprise when an email arrived 20 minutes later from the Netherlands, asking if I needed to have a replacement power supply shipped..."
They got the point. Clearly, they weren't going to make a sale that day. :-) But we listened to their presentation, and had a pleasant enough chat, and that was that. I had done my homework about their product line, too, and I think they knew it. They didn't try to BS me, and they didn't try to badmouth Netapp. So I can't comment on whether or not this is typical behavior from EMC reps, but my experience wasn't at all negative.
Mostly, though, it's the shirts. If my vendors didn't clothe me I'd have nothing to wear! :-)
-- Chris (still waiting for NetApp boxer shorts)
-- Chris Lamb, Unix Guy MeasureCast, Inc. 503-241-1469 x247 skeezics@measurecast.com
"barry" == Barry Lustig barry@lustig.com writes:
Here's what my response would typically look like:
barry> - The F85 is a stripped down, single CPU, low end device with barry> multiple points of failure and a very poor data protection.
Plese provide examples. Are you getting this information from spec sheets, white papers and analysis of the hardware or from another vendor's propaganda machine ?
barry> NetApps service is rated very low by industry experts and they barry> offer next day shipment of parts that the user must install barry> themselves.
Personal experience has shown that EMC's "service and support" can be just as bad, if not worse, than NetApps. I have had case where both dropped the ball in a serious way. Neither really made up for it.
barry> - The IP4700 is fully redundant, multi-CPU, mid-range device barry> with no single point of failure and hardware based RAID 5 data barry> protection.
I have yet to see detailed information regarding the depth of their fully redudant design. Granted, my gripe is with the all-important cache of the Symms that no one at EMC will talk about other than to say "it's fully redundant." The info I've seen on the IP4700 look good, though. But so do clustered NetApps.
barry> EMC's world class customer service center has ranked #1 for 6 barry> consecutive years by Gartner Group.
Personal experience has shown that EMC's "world class customer service center" still sucks. "Gosh, you mean the reason you haven't been able to fix my problem is because your engineers AREN'T EVENING RUNNING THE SAME CODE (ie, timefinder) ON THEIR TEST MACHINE?"
barry> Standard 2 year warranty guarantees 4 hours ON SITE w/Parts barry> service by EMC technicians.
NetApp offers comparable service.
barry> Plus, our 'call home' proactive maintenance system monitors barry> trends within the system and reports them automatically to our barry> customer service center. Often, EMC technicians will repair a barry> system BEFORE the component actually fails.
I can't actually rate this one, because we never got the system into a working state before I moved on. They may very well hold up to this one.
barry> The 4700 will be configured with 8 drives usable, plus 1 drive barry> for RAID 5 parity and 1 drive for hot swap redundancy. It is barry> scalable all the way to 7000GB (7TB) vs. only 648GB for the barry> F85. The 4700 as configured above is 8RU.
Are you actually going to ever put more than 650GB on it? Is that *really* an issue?
barry> In general, the F85 does not scale sufficiently for growth, has
Perhaps. That depends on what your expected growth is. It might very well be that the F85 is not the right choice. How about an 820 or maybe even an 840c?
barry> poor customer service behind it, has no redundancy, has barry> multiple points of failure, and utilizes a sub-par RAID 0 data barry> protection scheme.
The F85's use RAID0 ? Not having seen one, I suppose that's true, but why the break from RAID4 ? If it is RAID0, how is that "very poor data protection" as stated above ? EMC told me once that RAID0 was better than RAID5.
Anyways... lest people think I'm *completely* on NetApp's side, they both like to sling mud at each other and make up numbers and generally turn into a bunch of children. They both screw the pooch on service and then fail to take responsibilty for it. At least for the most part, NetApp is cheaper. ;-)
K.
This is, unfortunately, the way the EMC chooses to do business and the main reason that I do not like them. I have, to this point, not been required to purchase any of their equipment and prefer NetApp. As for NetApp support, they are outstanding in North America and are working on their global support model. Quite frankly I would rather buy an Auspex over an EMC -- is Asupex still in business?????
EMC has a God complex. They will perform maintenance on any EMC anywhere whether you want them to or not. I suggest that anyone having an EMC disconnect the phone line until you need assistance from tech support and then disconnect it immediately afterwards. EMC bills itself as number one in storage solution, which is true -- for direct attached storage. It is not, however, the number one in NAS. They are new to the NAS market and are undergoing some growing pains but they are improving.
Attacking the competition with untruths is a poor way to do business and one should really take that into consideration before purchasing a product from someone that does so. If they are lying to you now, when are they telling you the truth?
My very humble opinion on EMC.
Sincerely,
Glen D. Geen
Barry Lustig wrote:
This is an excerpt of an email that was sent to the management of a company that I'm working with. They had spec'ed F85's for a fairly low performance 10-20Mbits/sec. environment. The main requirement is reliability and ease of use. How does one respond to blatant misrepresentation of a competitors product?
barry
A few highlights of the comparison:
- The F85 is a stripped down, single CPU, low end device with multiple
points of failure and a very poor data protection. NetApps service is rated very low by industry experts and they offer next day shipment of parts that the user must install themselves.
- The IP4700 is fully redundant, multi-CPU, mid-range device with no
single point of failure and hardware based RAID 5 data protection. EMC's world class customer service center has ranked #1 for 6 consecutive years by Gartner Group. Standard 2 year warranty guarantees 4 hours ON SITE w/Parts service by EMC technicians. Plus, our 'call home' proactive maintenance system monitors trends within the system and reports them automatically to our customer service center. Often, EMC technicians will repair a system BEFORE the component actually fails.
The 4700 will be configured with 8 drives usable, plus 1 drive for RAID 5 parity and 1 drive for hot swap redundancy. It is scalable all the way to 7000GB (7TB) vs. only 648GB for the F85. The 4700 as configured above is 8RU.
In general, the F85 does not scale sufficiently for growth, has poor customer service behind it, has no redundancy, has multiple points of failure, and utilizes a sub-par RAID 0 data protection scheme.
-
Barry Lustig -
Chris Lamb -
fubar＠pir.net -
Jeff Kennedy -
Mr. G. D. Geen