This is a multi-part message in MIME format.
------=_NextPart_000_0042_01C4FFBF.B8FF1DE0
Content-Type: text/plain;
charset="iso-8859-7"
Content-Transfer-Encoding: 7bit
Following Phil's reply on the "ESCON Distance" thread, I would like to have
any suggestions or knowledge anyone may have on the setup we're going to use
to implement our disaster recovery plan.
We have two cpus. One (z/800) for online production (07:30am - 18:00pm)
which then runs half of the batch load, and a second one (z/890) for
development (07:30am-14:30pm) which then runs the other half of the batch
load. We use a SHARK model F20 using ESCON, and have ordered a second SHARK
model 800 with FICON.
The management wants us to design a setup whereby the two cpus will be in
different locations. The development cpu and the F20 on the primary site and
the production cpu with the 800 on the disaster recovery site. We will
replace ESCON with FICON on all machines for IO.
The 800 SHARK will be our primary storage and the F20 will be our mirror
(using PPRC).
The problem is that they want us to access the 800 Shark directly from the
development machine. To accomplish this we are going to use dark fibre and
DWDMs to eliminate the distance.
The main question is whether we can really use the FICON and DWDM technology
to run our batch on the development cpu with the 800 Shark being 15km away!
Does anyone have experience on that ?
We know that we're introducing other unwanted factors here (such as multiple
failure points), but I want to concentrate the discussion on IO throughput.
Our F20 is most probably be replaced with another 800 in order to do PPRC
with FICON.
Thanks
Yiannakis Vakis
Systems Support Group, I.T.Division
Tel. 22-848523, 99-414788, Fax. 22-337770
------=_NextPart_000_0042_01C4FFBF.B8FF1DE0
Content-Type: text/html;
charset="iso-8859-7"
Content-Transfer-Encoding: quoted-printable
charset=3Diso-8859-7">
size=3D2>Following Phil's=20
reply on the "ESCON Distance" thread, I would like to have any =
suggestions or=20
knowledge anyone may have on the setup we're going to use to implement =
our=20
disaster recovery plan.
size=3D2>
We =
have two cpus.=20
One (z/800) for online production (07:30am - 18:00pm) which then runs =
half of=20
the batch load, and a second one (z/890) for development =
(07:30am-14:30pm) which=20
then runs the other half of the batch load. We use a SHARK model F20 =
using=20
ESCON, and have ordered a second SHARK model 800 with =
FICON.
The =
management wants=20
us to design a setup whereby the two cpus will be in different =
locations. The=20
development cpu and the F20 on the primary site and the production cpu =
with the=20
800 on the disaster recovery site. We will replace ESCON with FICON =
on all=20
machines for IO.
The =
800 SHARK will=20
be our primary storage and the F20 will be our mirror (using=20
PPRC).
The =
problem is that=20
they want us to access the 800 Shark directly from the development =
machine.=20
To accomplish this we are going to use dark fibre and DWDMs to eliminate =
the=20
distance.
size=3D2>
The =
main question is=20
whether we can really use the FICON and DWDM technology to run our batch =
on the=20
development cpu with the 800 Shark being 15km away!
Does =
anyone have=20
experience on that ?
We =
know that we're=20
introducing other unwanted factors here (such as multiple failure =
points), but I=20
want to concentrate the discussion on IO throughput.
size=3D2>
Our =
F20 is most=20
probably be replaced with another 800 in order to do PPRC with=20
FICON.
size=3D2>Thanks
Yiannakis =
Vakis
face=3DArial size=3D2>Systems Support Group, I.T.Division =
face=3DArial size=3D2>Tel. 22-848523, 99-414788, Fax. 22-337770 =
------=_NextPart_000_0042_01C4FFBF.B8FF1DE0--
| 
