SSD's RE: Final hardware config for E9 implementa

Here's another fun site for SSD info:

http://www.dvnation.com/SATA-SSD.html

Check out the photofast link for the 900MB/S Raid config. Total full retail cost: $2,500.

Michael

Michael Barry
Aspacia Systems Inc
866.566.9600
312.803.0730 fax
http://www.aspacia.com/

This email, and any attachments thereto, is intended only for use by the addressee(s) named herein and may contain legally privileged and/or confidential information. If you are not the intended recipient of this email, you are hereby notified that any dissemination, distribution or copying of this email, and any attachments thereto, is strictly prohibited. If you have received this email in error, please immediately notify me by telephone and permanently delete the original and any copy of any email and any printout thereof.




On Dec 29, 2009, at 8:05 AM, Len Hartka wrote:

> Good Day Jared:
>
> I rarely forward Utube items, but you may find this interesting.
> Using SSD's these guys achieved throughput of 2 GB per second.
> A little weird, but interesting
>
> http://www.youtube.com/watch?v=96dWOEa4Djs
>
> len.hartka@...
>
>
>
>
> ________________________________
>
> From: vantage@yahoogroups.com [mailto:vantage@yahoogroups.com] On Behalf Of k99ja04
> Sent: Tuesday, December 29, 2009 10:44 AM
> To: vantage@yahoogroups.com
> Subject: OT: SSD's RE: [Vantage] Re: Final hardware config for E9 implementation -- maybe
>
>
>
>
> Hi Len,
> Thanks for the compliment! I will email you my spreadsheet (and anyone else that asks). It's not the prettiest format, but has 2 examples. The first takes a 64 GB drive and writes continuosly at the inputted rate. For write endurance I used 2 million, but the figures I see are anywhere from 1-5 million. The decision to classify drives as 100,000 writes is more a quality assurance decision not a technical one. AMD was selling million lifetime flash back in 1998! The second example is more elaborate and has some details that will be drive and partition specific, but I used typical values. I have tried to bold the items you can change, and have shown most of the conversions (I was trained as a chemist so those that know what factor labeling is will appreciate the layout), but again, not the prettiest. These are pretty silly calculations as no database I know of, except maybe Google or Facebook, can sustain that level of throughput. For us lowly ERP users, we will never hammer on these drives that hard.
>
> It's a very technical subject, but the take-home lesson is that SSDs are not quite there yet. As in they don't belong as your OS drive. And I wouldn't look for them to replace conventional drives anytime soon. But in the right application, they are unbeatable and the pros outweigh the cons.
>
> Yes, the FusionIO is quite frankly ridiculous. I asked Santa for one, but he didn't deliver. I know MySpace uses these toys. I saw an article on Fusion's site with throughput of over 1 terabyte per second using a relatively small datacenter. The same throughput on a conventional SAN setup would have required something like 55,000 spindles. There's a snippet here http://www.tomshardware.com/news/fusion-io-iodrive-octal-1tb,9140.html <http://www.tomshardware.com/news/fusion-io-iodrive-octal-1tb,9140.html>
>
> Jared
> _______________________
> Jared Allmond
> IT Systems Administrator
> Wright Coating Technologies
> jallmond@... <mailto:jallmond%40wrightcoating.com>
>
> --- In vantage@yahoogroups.com <mailto:vantage%40yahoogroups.com> , "Len Hartka" <len.hartka@...> wrote:
> >
> > Good Day Jared:
> >
> > Thanks for taking the time to compose that astounding reply.
> >
> > You should submit it to a magazine.
> >
> > It agrees with what I have read elsewhere, but is much better
> > organized and more thorough.
> > My comment about Laptops having the fast drives is just what I was
> > told was available by the Dell Representative about two months ago.
> >
> > I had seen the FusionIO PCI card on their website. Its numbers are
> > astounding - a Consultant I was talking to said it was the equivalent of
> > 100 spindles.
> >
> > I would be interested in your Excell speed sheet. You discussion is
> > on the edge of my knowledge, but I think SSD's are the future and the
> > sooner the better, but I had to pull the trigger about a month ago on my
> > E9 implementation; but my NEXT server will be SSD's.
> >
> >
> > len.hartka@...
>
>
>
>
>
>
> This e-mail and any attachments may contain proprietary and/or confidential information. If you are not the intended recipient, please notify the sender immediately by reply e-mail or at 410-472-2900 and then delete the message without using, disseminating, or copying this message or any portion thereof. With e-mail communications you are urged to protect against viruses.
>
> [Non-text portions of this message have been removed]
>
>



[Non-text portions of this message have been removed]
Good Day:

M2K, 7.0 SP7, Informix ( not Unix) (Currently using).
Epicor 9, 9.05, SQL (live 7/1/2010)

FYI
This is what the implementation tech, from Epicor, claims he will do
on 1/11/2010 when he installs our E9 software.
Sized for 50-120 users - we were at 50 before last year - hope to be
over 50 in a few years - did NOT want to be slow.
All HDD's are 15K rpm - the SSD's are just not quite ready yet.
Unfortunate since the SSD's are 5 times faster and no noise and low
power.

Comments?

len.hartka@...
<BLOCKED::mailto:len.hartka@...>

*************************************************
Apps Server - (2-146) RAID1-OS, OS
24G ( 2 -146) RAID1 ,
Epicor 9 Aplication
(1-146) hot spare

SQL - (2-73GB) Raid1, OS
and SQL app
48G (4-73GB) Raid 10
Temp-db
(1-73) hot spare

Disc Array - (10-73GB) RAID10, Data-MDF

(DAS) (12-73GB) RAID10,
Log-LDF
(2-73GB) hot Spare

Misc app\Virtualized: (3 -300GB) RAID 5 1. Web
Server w/4 cores, 8 gig memory win 2008 64bit Enterprise
24G (1-300GB) hot spare 2.
Crystal reports, APM, Mgnt Reporter w/2 cores 4 gig mem win

2003 32 bit Std ( Add SQL std??)

3. E Portal w/2 cores 4 gig mem win 2008 64 bit Enterprise

( Add SQL std??)
Redundant Server: (3-300GB)- RAID5
24G (1-300GB) hot Spare
*************************************************************


**************************************************************

Sun Automation Group

Celebrating

25 Years of Service

to the Corrugated Industry

**************************************************************




This e-mail and any attachments may contain proprietary and/or confidential information. If you are not the intended recipient, please notify the sender immediately by reply e-mail or at 410-472-2900 and then delete the message without using, disseminating, or copying this message or any portion thereof. With e-mail communications you are urged to protect against viruses.


[Non-text portions of this message have been removed]
vantage@yahoogroups.com wrote:
> Good Day:
>
> M2K, 7.0 SP7, Informix ( not Unix) (Currently using).
> Epicor 9, 9.05, SQL (live 7/1/2010)
>
> FYI
> This is what the implementation tech, from Epicor, claims he will do
> on 1/11/2010 when he installs our E9 software.
> Sized for 50-120 users - we were at 50 before last year - hope to be
> over 50 in a few years - did NOT want to be slow.
> All HDD's are 15K rpm - the SSD's are just not quite ready yet.
> Unfortunate since the SSD's are 5 times faster and no noise and low
> power.
>
> Comments?

Are there any remaining issues running 64-bit OSs with any of the web-based
products? I seem to recall issues with one of the web-based technologies.

Mark W.
We had problems with Embedded Courses working but they finally worked on
9.04.505a and 9.04.505b. Epicor is on Windows Server 2008 R2 Enterprise
64-bit. We haven't tried the Web Client yet.

----- Original Message -----
From: "Mark Wonsil" <mark_wonsil@...>
To: <vantage@yahoogroups.com>
Sent: Wednesday, December 23, 2009 8:50 AM
Subject: RE: [Vantage] Final hardware config for E9 implementation -- maybe


> vantage@yahoogroups.com wrote:
>> Good Day:
>>
>> M2K, 7.0 SP7, Informix ( not Unix) (Currently using).
>> Epicor 9, 9.05, SQL (live 7/1/2010)
>>
>> FYI
>> This is what the implementation tech, from Epicor, claims he will do
>> on 1/11/2010 when he installs our E9 software.
>> Sized for 50-120 users - we were at 50 before last year - hope to be
>> over 50 in a few years - did NOT want to be slow.
>> All HDD's are 15K rpm - the SSD's are just not quite ready yet.
>> Unfortunate since the SSD's are 5 times faster and no noise and low
>> power.
>>
>> Comments?
>
> Are there any remaining issues running 64-bit OSs with any of the
> web-based
> products? I seem to recall issues with one of the web-based technologies.
>
> Mark W.
>
>
>
> ------------------------------------
>
> Useful links for the Yahoo!Groups Vantage Board are: ( Note: You must
> have already linked your email address to a yahoo id to enable access. )
> (1) To access the Files Section of our Yahoo!Group for Report Builder and
> Crystal Reports and other 'goodies', please goto:
> http://groups.yahoo.com/group/vantage/files/.
> (2) To search through old msg's goto:
> http://groups.yahoo.com/group/vantage/messages
> (3) To view links to Vendors that provide Vantage services goto:
> http://groups.yahoo.com/group/vantage/linksYahoo! Groups Links
>
>
>


--------------------------------------------------------------------------------



No virus found in this incoming message.
Checked by AVG - www.avg.com
Version: 9.0.722 / Virus Database: 270.14.117/2583 - Release Date: 12/23/09
00:28:00
Good Day Mark:


Thanks for reminding me of that. The answer is yes, and in my email
I did indicate that one of the Guest-Virtualized-servers is running
32bit even though the Host-Server itself is running 64. That is
something cool that Virtualization can do.

BUT, I thought web-portal also had to be 32 bit, so I have to double
check.


len.hartka@...



________________________________

From: vantage@yahoogroups.com [mailto:vantage@yahoogroups.com] On Behalf
Of Mark Wonsil
Sent: Wednesday, December 23, 2009 11:50 AM
To: vantage@yahoogroups.com
Subject: RE: [Vantage] Final hardware config for E9 implementation --
maybe




vantage@yahoogroups.com <mailto:vantage%40yahoogroups.com> wrote:
> Good Day:
>
> M2K, 7.0 SP7, Informix ( not Unix) (Currently using).
> Epicor 9, 9.05, SQL (live 7/1/2010)
>
> FYI
> This is what the implementation tech, from Epicor, claims he will do
> on 1/11/2010 when he installs our E9 software.
> Sized for 50-120 users - we were at 50 before last year - hope to be
> over 50 in a few years - did NOT want to be slow.
> All HDD's are 15K rpm - the SSD's are just not quite ready yet.
> Unfortunate since the SSD's are 5 times faster and no noise and low
> power.
>
> Comments?

Are there any remaining issues running 64-bit OSs with any of the
web-based
products? I seem to recall issues with one of the web-based
technologies.

Mark W.






This e-mail and any attachments may contain proprietary and/or confidential information. If you are not the intended recipient, please notify the sender immediately by reply e-mail or at 410-472-2900 and then delete the message without using, disseminating, or copying this message or any portion thereof. With e-mail communications you are urged to protect against viruses.


[Non-text portions of this message have been removed]
Good Day:

Attached is a copy\paste of the email text I sent originally.
A LOT of text is missing. Compare to text in mark's email.
I looked at the email I got from Yahoo, and it was also missing the
data, so it wasn't a "Reply" truncation.

So, first, the full text is attached.

Secondly, how do I stop it from loosing text? I did notice that some
emails have " some text deleted" but I always ignored it. Should I send
all my text as an attached Word doc?

I do not normally use Yahoo mail, so I may be missing something
obvious?


len.hartka@...

________________________________

From: vantage@yahoogroups.com [mailto:vantage@yahoogroups.com] On Behalf
Of Mark Wonsil
Sent: Wednesday, December 23, 2009 11:50 AM
To: vantage@yahoogroups.com
Subject: RE: [Vantage] Final hardware config for E9 implementation --
maybe




vantage@yahoogroups.com <mailto:vantage%40yahoogroups.com> wrote:
> Good Day:
>
> M2K, 7.0 SP7, Informix ( not Unix) (Currently using).
> Epicor 9, 9.05, SQL (live 7/1/2010)
>
> FYI
> This is what the implementation tech, from Epicor, claims he will do
> on 1/11/2010 when he installs our E9 software.
> Sized for 50-120 users - we were at 50 before last year - hope to be
> over 50 in a few years - did NOT want to be slow.
> All HDD's are 15K rpm - the SSD's are just not quite ready yet.
> Unfortunate since the SSD's are 5 times faster and no noise and low
> power.
>
> Comments?

Are there any remaining issues running 64-bit OSs with any of the
web-based
products? I seem to recall issues with one of the web-based
technologies.

Mark W.






This e-mail and any attachments may contain proprietary and/or confidential information. If you are not the intended recipient, please notify the sender immediately by reply e-mail or at 410-472-2900 and then delete the message without using, disseminating, or copying this message or any portion thereof. With e-mail communications you are urged to protect against viruses.


[Non-text portions of this message have been removed]
And why exactly aren't SSD's ready yet? I mean they have only been used in military applications for a decade or so...

Michael Barry
Aspacia Systems Inc
866.566.9600
312.803.0730 fax
http://www.aspacia.com/

This email, and any attachments thereto, is intended only for use by the addressee(s) named herein and may contain legally privileged and/or confidential information. If you are not the intended recipient of this email, you are hereby notified that any dissemination, distribution or copying of this email, and any attachments thereto, is strictly prohibited. If you have received this email in error, please immediately notify me by telephone and permanently delete the original and any copy of any email and any printout thereof.




On Dec 23, 2009, at 6:52 AM, Len Hartka wrote:

> Good Day:
>
> M2K, 7.0 SP7, Informix ( not Unix) (Currently using).
> Epicor 9, 9.05, SQL (live 7/1/2010)
>
> FYI
> This is what the implementation tech, from Epicor, claims he will do
> on 1/11/2010 when he installs our E9 software.
> Sized for 50-120 users - we were at 50 before last year - hope to be
> over 50 in a few years - did NOT want to be slow.
> All HDD's are 15K rpm - the SSD's are just not quite ready yet.
> Unfortunate since the SSD's are 5 times faster and no noise and low
> power.
>
> Comments?
>
> len.hartka@...
> <BLOCKED::mailto:len.hartka@...>
>
> *************************************************
> Apps Server - (2-146) RAID1-OS, OS
> 24G ( 2 -146) RAID1 ,
> Epicor 9 Aplication
> (1-146) hot spare
>
> SQL - (2-73GB) Raid1, OS
> and SQL app
> 48G (4-73GB) Raid 10
> Temp-db
> (1-73) hot spare
>
> Disc Array - (10-73GB) RAID10, Data-MDF
>
> (DAS) (12-73GB) RAID10,
> Log-LDF
> (2-73GB) hot Spare
>
> Misc app\Virtualized: (3 -300GB) RAID 5 1. Web
> Server w/4 cores, 8 gig memory win 2008 64bit Enterprise
> 24G (1-300GB) hot spare 2.
> Crystal reports, APM, Mgnt Reporter w/2 cores 4 gig mem win
>
> 2003 32 bit Std ( Add SQL std??)
>
> 3. E Portal w/2 cores 4 gig mem win 2008 64 bit Enterprise
>
> ( Add SQL std??)
> Redundant Server: (3-300GB)- RAID5
> 24G (1-300GB) hot Spare
> *************************************************************
>
>
> **************************************************************
>
> Sun Automation Group
>
> Celebrating
>
> 25 Years of Service
>
> to the Corrugated Industry
>
> **************************************************************
>
> This e-mail and any attachments may contain proprietary and/or confidential information. If you are not the intended recipient, please notify the sender immediately by reply e-mail or at 410-472-2900 and then delete the message without using, disseminating, or copying this message or any portion thereof. With e-mail communications you are urged to protect against viruses.
>
> [Non-text portions of this message have been removed]
>
>



[Non-text portions of this message have been removed]
All of the web stuff will work on a 64 bit server. Most of them must be
run in 32-bit mode, but that's a very simple configuration change in IIS.

Brian Johnson -- Epi-Center <http://epi-ctr.com>
Director of Technical Services | direct: 413-531-2859

On 12/23/2009 11:50 AM, Mark Wonsil wrote:
>
> vantage@yahoogroups.com <mailto:vantage%40yahoogroups.com> wrote:
> > Good Day:
> >
> > M2K, 7.0 SP7, Informix ( not Unix) (Currently using).
> > Epicor 9, 9.05, SQL (live 7/1/2010)
> >
> > FYI
> > This is what the implementation tech, from Epicor, claims he will do
> > on 1/11/2010 when he installs our E9 software.
> > Sized for 50-120 users - we were at 50 before last year - hope to be
> > over 50 in a few years - did NOT want to be slow.
> > All HDD's are 15K rpm - the SSD's are just not quite ready yet.
> > Unfortunate since the SSD's are 5 times faster and no noise and low
> > power.
> >
> > Comments?
>
> Are there any remaining issues running 64-bit OSs with any of the
> web-based
> products? I seem to recall issues with one of the web-based technologies.
>
> Mark W.
>
>


[Non-text portions of this message have been removed]
Good Day Michael:

The fastest SSD's are only available in laptops.
The Bigger size of SSD's needed for servers run slower then the
laptop ones. Still faster then HDD's, but they soon they will be as fast
as the laptop ones.
Also, they all still degrade with use. They have compensation
for it, but its still a little shaky to bet your ERP system on.
Cost is also still a factor.
Having software vendors "Certify" their use has also not
happened yet.


len.hartka@...

-----Original Message-----
From: vantage@yahoogroups.com [mailto:vantage@yahoogroups.com] On Behalf
Of Michael Barry
Sent: Wednesday, December 23, 2009 12:34 PM
To: vantage@yahoogroups.com
Subject: Re: [Vantage] Final hardware config for E9 implementation --
maybe

And why exactly aren't SSD's ready yet? I mean they have only been used
in military applications for a decade or so...

Michael Barry
Aspacia Systems Inc
866.566.9600
312.803.0730 fax
http://www.aspacia.com/

This email, and any attachments thereto, is intended only for use by the
addressee(s) named herein and may contain legally privileged and/or
confidential information. If you are not the intended recipient of this
email, you are hereby notified that any dissemination, distribution or
copying of this email, and any attachments thereto, is strictly
prohibited. If you have received this email in error, please
immediately notify me by telephone and permanently delete the original
and any copy of any email and any printout thereof.




On Dec 23, 2009, at 6:52 AM, Len Hartka wrote:

> Good Day:
>
> M2K, 7.0 SP7, Informix ( not Unix) (Currently using).
> Epicor 9, 9.05, SQL (live 7/1/2010)
>
> FYI
> This is what the implementation tech, from Epicor, claims he will do
> on 1/11/2010 when he installs our E9 software.
> Sized for 50-120 users - we were at 50 before last year - hope to be
> over 50 in a few years - did NOT want to be slow.
> All HDD's are 15K rpm - the SSD's are just not quite ready yet.
> Unfortunate since the SSD's are 5 times faster and no noise and low
> power.
>
> Comments?
>
> len.hartka@...
> <BLOCKED::mailto:len.hartka@...>
>
> *************************************************
> Apps Server - (2-146) RAID1-OS, OS
> 24G ( 2 -146) RAID1 ,
> Epicor 9 Aplication
> (1-146) hot spare
>
> SQL - (2-73GB) Raid1, OS
> and SQL app
> 48G (4-73GB) Raid 10
> Temp-db
> (1-73) hot spare
>
> Disc Array - (10-73GB) RAID10, Data-MDF
>
> (DAS) (12-73GB) RAID10,
> Log-LDF
> (2-73GB) hot Spare
>
> Misc app\Virtualized: (3 -300GB) RAID 5 1. Web Server w/4 cores, 8 gig

> memory win 2008 64bit Enterprise 24G (1-300GB) hot spare 2.
> Crystal reports, APM, Mgnt Reporter w/2 cores 4 gig mem win
>
> 2003 32 bit Std ( Add SQL std??)
>
> 3. E Portal w/2 cores 4 gig mem win 2008 64 bit Enterprise
>
> ( Add SQL std??)
> Redundant Server: (3-300GB)- RAID5
> 24G (1-300GB) hot Spare
> *************************************************************
>
>
> **************************************************************
>
> Sun Automation Group
>
> Celebrating
>
> 25 Years of Service
>
> to the Corrugated Industry
>
> **************************************************************
>
> This e-mail and any attachments may contain proprietary and/or
confidential information. If you are not the intended recipient, please
notify the sender immediately by reply e-mail or at 410-472-2900 and
then delete the message without using, disseminating, or copying this
message or any portion thereof. With e-mail communications you are urged
to protect against viruses.
>
> [Non-text portions of this message have been removed]
>
>



[Non-text portions of this message have been removed]



------------------------------------

Useful links for the Yahoo!Groups Vantage Board are: ( Note: You must
have already linked your email address to a yahoo id to enable access. )
(1) To access the Files Section of our Yahoo!Group for Report Builder
and Crystal Reports and other 'goodies', please goto:
http://groups.yahoo.com/group/vantage/files/.
(2) To search through old msg's goto:
http://groups.yahoo.com/group/vantage/messages
(3) To view links to Vendors that provide Vantage services goto:
http://groups.yahoo.com/group/vantage/linksYahoo! Groups Links





This e-mail and any attachments may contain proprietary and/or confidential information. If you are not the intended recipient, please notify the sender immediately by reply e-mail or at 410-472-2900 and then delete the message without using, disseminating, or copying this message or any portion thereof. With e-mail communications you are urged to protect against viruses.
Hi Len,

> Attached is a copy\paste of the email text I sent originally.
> A LOT of text is missing. Compare to text in mark's email.
> I looked at the email I got from Yahoo, and it was also missing the
> data, so it wasn't a "Reply" truncation.
>
> So, first, the full text is attached.

The first thing to know about this Yahoo! Group is that we don't allow
attachments - for good reason (security) - so your attachment is not here.
But your orignal post is still on the Yahoo! Group page in its entirety.

> Secondly, how do I stop it from loosing text? I did notice that some
> emails have " some text deleted" but I always ignored it.
> Should I send all my text as an attached Word doc?
>
> I do not normally use Yahoo mail, so I may be missing something
> obvious?

No, you're not missing anything obvious. I belong to several newsgroups and
I follow the posting policies outlined in the Web Design List:

http://webdesign-l.com/policies/

Specifically, I trim my posts. I only include the text relevant to my reply.
Some people get an email for every post but others get a daily digest. The
digest gets HUGE with the same text and all of the footers appearing again,
and again, and again for each reply. And if you ever try to search for
something at the Yahoo! website, good luck! The thread may break into
several topics and you'll be searching for hours going through false hits
because old text was included in a repsonse that has nothing to do with what
you're looking for.

Note that this group is not made up of Internet professionals or the typical
folk found in Internet technolody newsgroups like Web Design, Linux, or the
many programming groups. This is more "consumer orientend" so there's not
the same kind of peer-pressure to follow the old hard-liners' nettiquette.
The Web Design list is about twelve or thirteen years old - an ETERNITY in
Internet time. The "Jack-booted list Mom" sends out one or two of these
kinds of messages each year:

> [WD]: semi-periodic and vaguely furious reminder not to be a moron here
>
> Just as a quick reminder, people who reply to messages from this list
> without first trimming their reply so that it does not contain the
> previous message in its entirety will have their messages land in a
> silent hell called the "WD quarantine folder". This is also the case
> for those who post from accounts with stupid "intended for the
> recipient" boilerplate .sigs, which I assume are foisted on them by
> idiotic legal departments and compliant Exchange server admins. HTML
> email, fortunately, is rejected with the vague and unhelpful error
> that you've violated a rule against content types, which you must
> read as "MIME Content Type" rather than "words and such". Plain text
> only, people. And learn how to send email in a way that respects
> others' time and energies.
>
> Once in the quarantine, the messages will languish and not be posted
> to the rest of the list, which is what I can only assume is the
> intent of sending mail to the list. I occasionally check the folder
> just to see how many idiots are posting in a proscribed manner.
> Even more rarely, I will follow up with a mildly irritated taunt.
> Mostly, I just throw my hands in the air and wonder how you even
> feed yourselves.
>
> So, if you're so dumb and/or lazy you can't follow simple directions,
> we'll all be spared the horror of your presence here. It's a "win
> win"!
>
> Thank you.
>
> For those who are so dense as to never have bothered to find out what
> "top posting" is, despite the URL to a wikipedia article explaining
> same being included in EVERY POST to the list, I beg you - please,
> please leave. Let the rest of us all continue to exchange mail in a
> respectful manner, without the nightmare that is modern email
> etiquette (otherwise known as "ignorance").
>
> Yours sincerely,
> Steve
> jack-booted list mom

Fortunately, we're not that strict here. I still follow these guidelines
because of habit and because they do improve the quality of the newsgroup.

Happy Holidays!!!

Mark W.
Your SQL storage subsystem looks good in my opinion. My simple rules are to separate the transaction logs, tempDB and data files on separate RAID 10 arrays (controllers too!) if you can afford it. RAID 1 is fine for the OS. Hot spares are always a good idea. Buy enough RAM so you can load the entire database into memory if you can. The current bleeding edge processor from Intel is the Nehalem based quad core Xeon (X5570) @ 2.93Ghz. Too bad the Westmere 6 core bad boys are not due until March 2010. But the processor is not the usual bottleneck on a SQL box.

I maintain that SSDs are in fact ready for the enterprise in certain circumstances. We run Vantage 8.03.408B SQL and my production data files are presently on a RAID 10 array with 8 32 GB Intel X-25e SSDs off an Adaptec 5805. This setup screams! Now I did pay a pretty penny for the setup. Each SSD cost nearly $400 and list price on the card is $600 (but I picked up the controller for $200 off eBay). So while I do have $3400 invested in this setup I will take any $8000 Dell MD3000 fibre channel SAN for an absolute walk.

You could save a few bucks and only use 4 SSDs with a conventional drive as a hot spare, but chances are an SSD isn't going to fail and having 8 is way cooler.

The tradeoff with SSDs vs conventional disks in terms of economics is capacity. I paid $383 for 64 GB and could have bought a 300 GB SAS drive. A 2 TB SATA drive is even less than that. A 300 GB SAS drive My array setup gives only 128 GB of space which is paltry compared to a 2 TB drives. However, given this array only hosts my 8 GB .mdf SQL data file that is more than adequate.

Len, where did you get "the fastest SSDs come on laptops?" It is true that most SSDs are the smaller 2.5 inch form factor (with 1.8's inch on the way), which does make them ideal for laptops (plus they consume less power), but speed has nothing to do with what you can get in a laptop versus what's available in a desktop/server.

As for being "certified" by vendors... well, vendors are lazy! Validation testing takes time and money. Plus storage consolidation means vendors sell fewer servers so it's not in their interest to "certify." Screw certification anyway. Do your own homework and figure out what hardware plays nice together. If you only choose "certified" configurations, chances are you're already behind the times. There is a wealth of information on the web about SSDs and benchmarks etc, etc.

If you're not a techie person you may want to stop reading now. You've been warned.

Primer: SSDs are built using NAND flash memory which does have a limited life expectancy in that each cell can only be written to a finite number of times. NAND also comes in two flavors; single level cell (SLC) and multi level cell (MLC). Without getting into too much technical detail SLCs store 1 bit per cell and MLCs store 2 bits per cell. SLC based SSDs are more expensive and each bit has about a 2 million write cycle lifespan compared to MLCs 10,000 write cycle lifespan. (Note: You may read that SLC is 100,000 write cycles but that is not present day specs. Current SLC flash memory is between 1-5 million cycles.) Given the added cost of SLC flash memory chips, thes SSDs have capacities limited to about 64 GB whereas MLCs come in 100s of GB up to 1 TB (The MLC OCZ Colossus costs $4k though). The takeaway lesson here is MLC's are for the desktop, but get off your wallet for an SLC in the enterprise.

SSD performance compared to conventional hard drives is nothing short of stunning. Access latencies approach 0.1ms, with read throughputs at 250 MB/s and writes at up to 170 MB/s. The true benefit of SSDs is not in sustained throughput. Conventional disk based RAID arrays can still throw down 800MB/s. The key here is the access latency; which is near nonexistent. Where SSDs really shine is in random read performance, which is absolutely perfect for database applications. A typical DB I/O usage pattern is 75% reads and 25% writes. As I said, a snappy RAID array based on 15k rpm conventional disks can still nose ahead of SSDs in performance when it comes to sequential read/write performance. So store your SQL transaction logs there since they are written sequentially. Put your .mdf data file on a SSD array as that is most likely to experience random reads.

There is a price to that performance, not just in terms of dollars, but they do degrade with use over time. While SSD firmware is very sophisticated and employs techniques such as "wear leveling" to get the max lifecycle out of a drive, they also will suffer from something called write amplification which can prematurely kill a drive. Even more so if your partitions are not aligned! Flash cells can be written to in smaller 4k pages, but it must be erased in blocks of 128k (actual numbers depend on drive, worse with MLC ~512k blocks). An actual write cycle involves reading the block, erasing the block, and the writing the data. For small writes it is possible to burn two adjacent blocks even though you wrote a quantity that would sufficiently fit on a single physical block. Intel's firmware uses a caching technique called "write combining" to save up small writes until it has to wipe an entire block which helps.

However, write performance will degrade over time also due to that same wear leveling. When that SSD is shiny and new, all the cells are empty, so don't have to go through a read, erase, write cycle for that initial write. Erases are very much slower than writes, thus the performance degradation. But wear leveling basically creates "fragmentation". That being said you should never run a defragment utility as the "fragmentation" is at the physical block layer not the logical I/O block level. This is something only the drives firmware can correct. MLC drives are far more susceptible to this degradation than SLC drives. Current benchmarks suggest that there will be an initial dip in performance but the drive tends to adapt once every bit has been written to at least once. Also, this is not irreparable damage like the bit level write cycle lifespan. You can "wipe" the drive and return it to like new condition. Your mileage will vary on when this is necessary. Intel suggests to either "write big" or "write small" across the entire drive in order to synthetically defrag the drive, but empiric results show there is a tipping point and the drive should be erased. Again this is aimed more at MLC SSDs.

There is also a function called Trim that is only implemented at the OS level beginning with Windows 7 and Server 2008 R2, but firmwares have only recently become available that implement trim. This helps reduce the fragmentation described above, but only applies to individual files. This is of most benefit when using an SSD as the OS drive as Windows tends to do a lot of random writes in the background. If an SSD is hosting an SQL .mdf file then any writes to the database still occur within the same file, so trim is no help here. You will eventually have to resurrect the drive, but hey that's good practice to ensure backup integrity.

Most of this degradation talk is irrelevant as, and this part is really important, a degraded disk will still slaughter a conventional disk!

Now what about endurance in the enterprise over time. You can write 100 GB of data per day to a 128 GB SSD and the drive will last over 100 years. That assumes 100% efficienct write leveling, but still, I think we're ok here.

What about an absolute worst case scenario? And I mean something really obnoxiously super stupid! I did a quick back of the envelope calculation continuously writing to that same drive at 170 MB/sec 24/7. We assume 100% perfect wear leveling which means we need to fill the disk 1 million times to get to the write endurance limit. So making sure to convert all the appropriate units 1 million (write endurance) x 64GB (capacity) divided by 170M bytes / sec gives the endurance limited life in seconds. That's a meaningless number, which needs to be divided by seconds in an hour, hours in a day etc etc to give...

The end result is 24.4 years!

I made some other assumptions here like 64k I/O block size etc but you get the idea (I'd be happy to send anyone an excel with my math). And if you're not using 64k blocks for SQL partitions you should be.

There's another aspect of SSDs to consider which is power savings over time. Take a data center with tons over drives in SANs that each take 7 watts at idle. Over time that will add a huge chunk to the electric bill both in terms of direct use and indirect requirements to the cooling system. SSD's use less than half a watt and need little to no cooling.

The next evolution in SSDs is eliminating the bottleneck of the SATA 2 bus. There are a few companies out there (notably FusionIO) that have stuck NAND memory on a PCI express card. The I/O performance of those beasts is just jaw dropping. Sitting on the PCIe bus right next to the processor effectively makes the processor the bottleneck. Weird huh? But the price of this technology goes along the lines of "if you have to askÂ…"

The biggest reasons SSDs are not mainstream is cost and education. Most people haven't taken the time to educate themselves on what's out there and don't really understand SSDs. I hope this post has brought you a little closer.

Jared
_______________________
Jared Allmond
IT Systems Administrator
Wright Coating Technologies
jallmond@...
voice: 269.344.8195
direct: 269.341.4353
fax: 269.344.3007

--- In vantage@yahoogroups.com, "Len Hartka" <len.hartka@...> wrote:
>
> Good Day Michael:
>
> The fastest SSD's are only available in laptops.
> The Bigger size of SSD's needed for servers run slower then the
> laptop ones. Still faster then HDD's, but they soon they will be as fast
> as the laptop ones.
> Also, they all still degrade with use. They have compensation
> for it, but its still a little shaky to bet your ERP system on.
> Cost is also still a factor.
> Having software vendors "Certify" their use has also not
> happened yet.
>
>
> len.hartka@...
>
> -----Original Message-----
> From: vantage@yahoogroups.com [mailto:vantage@yahoogroups.com] On Behalf
> Of Michael Barry
> Sent: Wednesday, December 23, 2009 12:34 PM
> To: vantage@yahoogroups.com
> Subject: Re: [Vantage] Final hardware config for E9 implementation --
> maybe
>
> And why exactly aren't SSD's ready yet? I mean they have only been used
> in military applications for a decade or so...
>
> Michael Barry
> Aspacia Systems Inc
> 866.566.9600
> 312.803.0730 fax
> http://www.aspacia.com/
Good Day Jared:

Thanks for taking the time to compose that astounding reply.

You should submit it to a magazine.

It agrees with what I have read elsewhere, but is much better
organized and more thorough.
My comment about Laptops having the fast drives is just what I was
told was available by the Dell Representative about two months ago.

I had seen the FusionIO PCI card on their website. Its numbers are
astounding - a Consultant I was talking to said it was the equivalent of
100 spindles.

I would be interested in your Excell speed sheet. You discussion is
on the edge of my knowledge, but I think SSD's are the future and the
sooner the better, but I had to pull the trigger about a month ago on my
E9 implementation; but my NEXT server will be SSD's.


len.hartka@...

________________________________

From: vantage@yahoogroups.com [mailto:vantage@yahoogroups.com] On Behalf
Of k99ja04
Sent: Thursday, December 24, 2009 3:16 PM
To: vantage@yahoogroups.com
Subject: [Vantage] Re: Final hardware config for E9 implementation --
maybe




Your SQL storage subsystem looks good in my opinion. My simple rules are
to separate the transaction logs, tempDB and data files on separate RAID
10 arrays (controllers too!) if you can afford it. RAID 1 is fine for
the OS. Hot spares are always a good idea. Buy enough RAM so you can
load the entire database into memory if you can. The current bleeding
edge processor from Intel is the Nehalem based quad core Xeon (X5570) @
2.93Ghz. Too bad the Westmere 6 core bad boys are not due until March
2010. But the processor is not the usual bottleneck on a SQL box.

I maintain that SSDs are in fact ready for the enterprise in certain
circumstances. We run Vantage 8.03.408B SQL and my production data files
are presently on a RAID 10 array with 8 32 GB Intel X-25e SSDs off an
Adaptec 5805. This setup screams! Now I did pay a pretty penny for the
setup. Each SSD cost nearly $400 and list price on the card is $600 (but
I picked up the controller for $200 off eBay). So while I do have $3400
invested in this setup I will take any $8000 Dell MD3000 fibre channel
SAN for an absolute walk.

You could save a few bucks and only use 4 SSDs with a conventional drive
as a hot spare, but chances are an SSD isn't going to fail and having 8
is way cooler.

The tradeoff with SSDs vs conventional disks in terms of economics is
capacity. I paid $383 for 64 GB and could have bought a 300 GB SAS
drive. A 2 TB SATA drive is even less than that. A 300 GB SAS drive My
array setup gives only 128 GB of space which is paltry compared to a 2
TB drives. However, given this array only hosts my 8 GB .mdf SQL data
file that is more than adequate.

Len, where did you get "the fastest SSDs come on laptops?" It is true
that most SSDs are the smaller 2.5 inch form factor (with 1.8's inch on
the way), which does make them ideal for laptops (plus they consume less
power), but speed has nothing to do with what you can get in a laptop
versus what's available in a desktop/server.

As for being "certified" by vendors... well, vendors are lazy!
Validation testing takes time and money. Plus storage consolidation
means vendors sell fewer servers so it's not in their interest to
"certify." Screw certification anyway. Do your own homework and figure
out what hardware plays nice together. If you only choose "certified"
configurations, chances are you're already behind the times. There is a
wealth of information on the web about SSDs and benchmarks etc, etc.

If you're not a techie person you may want to stop reading now. You've
been warned.

Primer: SSDs are built using NAND flash memory which does have a limited
life expectancy in that each cell can only be written to a finite number
of times. NAND also comes in two flavors; single level cell (SLC) and
multi level cell (MLC). Without getting into too much technical detail
SLCs store 1 bit per cell and MLCs store 2 bits per cell. SLC based SSDs
are more expensive and each bit has about a 2 million write cycle
lifespan compared to MLCs 10,000 write cycle lifespan. (Note: You may
read that SLC is 100,000 write cycles but that is not present day specs.
Current SLC flash memory is between 1-5 million cycles.) Given the added
cost of SLC flash memory chips, thes SSDs have capacities limited to
about 64 GB whereas MLCs come in 100s of GB up to 1 TB (The MLC OCZ
Colossus costs $4k though). The takeaway lesson here is MLC's are for
the desktop, but get off your wallet for an SLC in the enterprise.

SSD performance compared to conventional hard drives is nothing short of
stunning. Access latencies approach 0.1ms, with read throughputs at 250
MB/s and writes at up to 170 MB/s. The true benefit of SSDs is not in
sustained throughput. Conventional disk based RAID arrays can still
throw down 800MB/s. The key here is the access latency; which is near
nonexistent. Where SSDs really shine is in random read performance,
which is absolutely perfect for database applications. A typical DB I/O
usage pattern is 75% reads and 25% writes. As I said, a snappy RAID
array based on 15k rpm conventional disks can still nose ahead of SSDs
in performance when it comes to sequential read/write performance. So
store your SQL transaction logs there since they are written
sequentially. Put your .mdf data file on a SSD array as that is most
likely to experience random reads.

There is a price to that performance, not just in terms of dollars, but
they do degrade with use over time. While SSD firmware is very
sophisticated and employs techniques such as "wear leveling" to get the
max lifecycle out of a drive, they also will suffer from something
called write amplification which can prematurely kill a drive. Even more
so if your partitions are not aligned! Flash cells can be written to in
smaller 4k pages, but it must be erased in blocks of 128k (actual
numbers depend on drive, worse with MLC ~512k blocks). An actual write
cycle involves reading the block, erasing the block, and the writing the
data. For small writes it is possible to burn two adjacent blocks even
though you wrote a quantity that would sufficiently fit on a single
physical block. Intel's firmware uses a caching technique called "write
combining" to save up small writes until it has to wipe an entire block
which helps.

However, write performance will degrade over time also due to that same
wear leveling. When that SSD is shiny and new, all the cells are empty,
so don't have to go through a read, erase, write cycle for that initial
write. Erases are very much slower than writes, thus the performance
degradation. But wear leveling basically creates "fragmentation". That
being said you should never run a defragment utility as the
"fragmentation" is at the physical block layer not the logical I/O block
level. This is something only the drives firmware can correct. MLC
drives are far more susceptible to this degradation than SLC drives.
Current benchmarks suggest that there will be an initial dip in
performance but the drive tends to adapt once every bit has been written
to at least once. Also, this is not irreparable damage like the bit
level write cycle lifespan. You can "wipe" the drive and return it to
like new condition. Your mileage will vary on when this is necessary.
Intel suggests to either "write big" or "write small" across the entire
drive in order to synthetically defrag the drive, but empiric results
show there is a tipping point and the drive should be erased. Again this
is aimed more at MLC SSDs.

There is also a function called Trim that is only implemented at the OS
level beginning with Windows 7 and Server 2008 R2, but firmwares have
only recently become available that implement trim. This helps reduce
the fragmentation described above, but only applies to individual files.
This is of most benefit when using an SSD as the OS drive as Windows
tends to do a lot of random writes in the background. If an SSD is
hosting an SQL .mdf file then any writes to the database still occur
within the same file, so trim is no help here. You will eventually have
to resurrect the drive, but hey that's good practice to ensure backup
integrity.

Most of this degradation talk is irrelevant as, and this part is really
important, a degraded disk will still slaughter a conventional disk!

Now what about endurance in the enterprise over time. You can write 100
GB of data per day to a 128 GB SSD and the drive will last over 100
years. That assumes 100% efficienct write leveling, but still, I think
we're ok here.

What about an absolute worst case scenario? And I mean something really
obnoxiously super stupid! I did a quick back of the envelope calculation
continuously writing to that same drive at 170 MB/sec 24/7. We assume
100% perfect wear leveling which means we need to fill the disk 1
million times to get to the write endurance limit. So making sure to
convert all the appropriate units 1 million (write endurance) x 64GB
(capacity) divided by 170M bytes / sec gives the endurance limited life
in seconds. That's a meaningless number, which needs to be divided by
seconds in an hour, hours in a day etc etc to give...

The end result is 24.4 years!

I made some other assumptions here like 64k I/O block size etc but you
get the idea (I'd be happy to send anyone an excel with my math). And if
you're not using 64k blocks for SQL partitions you should be.

There's another aspect of SSDs to consider which is power savings over
time. Take a data center with tons over drives in SANs that each take 7
watts at idle. Over time that will add a huge chunk to the electric bill
both in terms of direct use and indirect requirements to the cooling
system. SSD's use less than half a watt and need little to no cooling.

The next evolution in SSDs is eliminating the bottleneck of the SATA 2
bus. There are a few companies out there (notably FusionIO) that have
stuck NAND memory on a PCI express card. The I/O performance of those
beasts is just jaw dropping. Sitting on the PCIe bus right next to the
processor effectively makes the processor the bottleneck. Weird huh? But
the price of this technology goes along the lines of "if you have to
ask..."

The biggest reasons SSDs are not mainstream is cost and education. Most
people haven't taken the time to educate themselves on what's out there
and don't really understand SSDs. I hope this post has brought you a
little closer.

Jared
_______________________
Jared Allmond
IT Systems Administrator
Wright Coating Technologies
jallmond@... <mailto:jallmond%40wrightcoating.com>
voice: 269.344.8195
direct: 269.341.4353
fax: 269.344.3007

--- In vantage@yahoogroups.com <mailto:vantage%40yahoogroups.com> , "Len
Hartka" <len.hartka@...> wrote:
>
> Good Day Michael:
>
> The fastest SSD's are only available in laptops.
> The Bigger size of SSD's needed for servers run slower then the
> laptop ones. Still faster then HDD's, but they soon they will be as
fast
> as the laptop ones.
> Also, they all still degrade with use. They have compensation
> for it, but its still a little shaky to bet your ERP system on.
> Cost is also still a factor.
> Having software vendors "Certify" their use has also not
> happened yet.
>
>
> len.hartka@...
>
> -----Original Message-----
> From: vantage@yahoogroups.com <mailto:vantage%40yahoogroups.com>
[mailto:vantage@yahoogroups.com <mailto:vantage%40yahoogroups.com> ] On
Behalf
> Of Michael Barry
> Sent: Wednesday, December 23, 2009 12:34 PM
> To: vantage@yahoogroups.com <mailto:vantage%40yahoogroups.com>
> Subject: Re: [Vantage] Final hardware config for E9 implementation --
> maybe
>
> And why exactly aren't SSD's ready yet? I mean they have only been
used
> in military applications for a decade or so...
>
> Michael Barry
> Aspacia Systems Inc
> 866.566.9600
> 312.803.0730 fax
> http://www.aspacia.com/ <http://www.aspacia.com/>






This e-mail and any attachments may contain proprietary and/or confidential information. If you are not the intended recipient, please notify the sender immediately by reply e-mail or at 410-472-2900 and then delete the message without using, disseminating, or copying this message or any portion thereof. With e-mail communications you are urged to protect against viruses.


[Non-text portions of this message have been removed]
Hi Len,
Thanks for the compliment! I will email you my spreadsheet (and anyone else that asks). It's not the prettiest format, but has 2 examples. The first takes a 64 GB drive and writes continuosly at the inputted rate. For write endurance I used 2 million, but the figures I see are anywhere from 1-5 million. The decision to classify drives as 100,000 writes is more a quality assurance decision not a technical one. AMD was selling million lifetime flash back in 1998! The second example is more elaborate and has some details that will be drive and partition specific, but I used typical values. I have tried to bold the items you can change, and have shown most of the conversions (I was trained as a chemist so those that know what factor labeling is will appreciate the layout), but again, not the prettiest. These are pretty silly calculations as no database I know of, except maybe Google or Facebook, can sustain that level of throughput. For us lowly ERP users, we will never hammer on these drives that hard.

It's a very technical subject, but the take-home lesson is that SSDs are not quite there yet. As in they don't belong as your OS drive. And I wouldn't look for them to replace conventional drives anytime soon. But in the right application, they are unbeatable and the pros outweigh the cons.

Yes, the FusionIO is quite frankly ridiculous. I asked Santa for one, but he didn't deliver. I know MySpace uses these toys. I saw an article on Fusion's site with throughput of over 1 terabyte per second using a relatively small datacenter. The same throughput on a conventional SAN setup would have required something like 55,000 spindles. There's a snippet here http://www.tomshardware.com/news/fusion-io-iodrive-octal-1tb,9140.html

Jared
_______________________
Jared Allmond
IT Systems Administrator
Wright Coating Technologies
jallmond@...


--- In vantage@yahoogroups.com, "Len Hartka" <len.hartka@...> wrote:
>
> Good Day Jared:
>
> Thanks for taking the time to compose that astounding reply.
>
> You should submit it to a magazine.
>
> It agrees with what I have read elsewhere, but is much better
> organized and more thorough.
> My comment about Laptops having the fast drives is just what I was
> told was available by the Dell Representative about two months ago.
>
> I had seen the FusionIO PCI card on their website. Its numbers are
> astounding - a Consultant I was talking to said it was the equivalent of
> 100 spindles.
>
> I would be interested in your Excell speed sheet. You discussion is
> on the edge of my knowledge, but I think SSD's are the future and the
> sooner the better, but I had to pull the trigger about a month ago on my
> E9 implementation; but my NEXT server will be SSD's.
>
>
> len.hartka@...
Good Day Jared:

I rarely forward Utube items, but you may find this interesting.
Using SSD's these guys achieved throughput of 2 GB per second.
A little weird, but interesting

http://www.youtube.com/watch?v=96dWOEa4Djs

len.hartka@...




________________________________

From: vantage@yahoogroups.com [mailto:vantage@yahoogroups.com] On Behalf Of k99ja04
Sent: Tuesday, December 29, 2009 10:44 AM
To: vantage@yahoogroups.com
Subject: OT: SSD's RE: [Vantage] Re: Final hardware config for E9 implementation -- maybe




Hi Len,
Thanks for the compliment! I will email you my spreadsheet (and anyone else that asks). It's not the prettiest format, but has 2 examples. The first takes a 64 GB drive and writes continuosly at the inputted rate. For write endurance I used 2 million, but the figures I see are anywhere from 1-5 million. The decision to classify drives as 100,000 writes is more a quality assurance decision not a technical one. AMD was selling million lifetime flash back in 1998! The second example is more elaborate and has some details that will be drive and partition specific, but I used typical values. I have tried to bold the items you can change, and have shown most of the conversions (I was trained as a chemist so those that know what factor labeling is will appreciate the layout), but again, not the prettiest. These are pretty silly calculations as no database I know of, except maybe Google or Facebook, can sustain that level of throughput. For us lowly ERP users, we will never hammer on these drives that hard.

It's a very technical subject, but the take-home lesson is that SSDs are not quite there yet. As in they don't belong as your OS drive. And I wouldn't look for them to replace conventional drives anytime soon. But in the right application, they are unbeatable and the pros outweigh the cons.

Yes, the FusionIO is quite frankly ridiculous. I asked Santa for one, but he didn't deliver. I know MySpace uses these toys. I saw an article on Fusion's site with throughput of over 1 terabyte per second using a relatively small datacenter. The same throughput on a conventional SAN setup would have required something like 55,000 spindles. There's a snippet here http://www.tomshardware.com/news/fusion-io-iodrive-octal-1tb,9140.html <http://www.tomshardware.com/news/fusion-io-iodrive-octal-1tb,9140.html>

Jared
_______________________
Jared Allmond
IT Systems Administrator
Wright Coating Technologies
jallmond@... <mailto:jallmond%40wrightcoating.com>

--- In vantage@yahoogroups.com <mailto:vantage%40yahoogroups.com> , "Len Hartka" <len.hartka@...> wrote:
>
> Good Day Jared:
>
> Thanks for taking the time to compose that astounding reply.
>
> You should submit it to a magazine.
>
> It agrees with what I have read elsewhere, but is much better
> organized and more thorough.
> My comment about Laptops having the fast drives is just what I was
> told was available by the Dell Representative about two months ago.
>
> I had seen the FusionIO PCI card on their website. Its numbers are
> astounding - a Consultant I was talking to said it was the equivalent of
> 100 spindles.
>
> I would be interested in your Excell speed sheet. You discussion is
> on the edge of my knowledge, but I think SSD's are the future and the
> sooner the better, but I had to pull the trigger about a month ago on my
> E9 implementation; but my NEXT server will be SSD's.
>
>
> len.hartka@...






This e-mail and any attachments may contain proprietary and/or confidential information. If you are not the intended recipient, please notify the sender immediately by reply e-mail or at 410-472-2900 and then delete the message without using, disseminating, or copying this message or any portion thereof. With e-mail communications you are urged to protect against viruses.


[Non-text portions of this message have been removed]