Possible to have 5Mbps point to point leased circuit with ping latency of <5ms end to end

EMC SAN sales person told me that as long as the leased circuit
WAN (point to point link) has end to end ping latency of less than
5ms, then it's possible to have SAN replication.

Based on our current log shipping largest file of 50MB & our MS
SQL log ship file is created every 15mins, I deduced that the
peak write IO is   50x1024x1024x8 / 15x60 = 466kbps.

Giving the benefit of doubt that there could be surges in write
IO to the database occasionally, I reckon 5Mbps should be able
to cover the occasional surge in IO.

EMC SAN sales person told me as long as we can get a ping latency
of less than 5ms, this 5Mbps WAN circuit should be able to support
EMC's SRDF (sort of EMC's SAN replication)

Q1:
Anyone has any experience whether a normal IP link (copper)
5Mbps can achieve this ping latency time?

Q2:
Would a "Dark Fiber" 5Mbps link give a much better ping latency
time?

Q3:
If I were to use compression modules on the WAN router, would
it help with the ping latency & help improve the bandwidth utilization?
sunhuxAsked:
Who is Participating?

[Product update] Infrastructure Analysis Tool is now available with Business Accounts.Learn More

x
I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

sunhuxAuthor Commented:
Does the link / url below provides a good way to size the bandwidth
needed for SAN replication:

http://searchstoragechannel.techtarget.com/feature/Network-bandwidth-Sizing-for-long-distance-storage-networking-applications
0
giltjrCommented:
Total bandwidth is the the major impact on latency.  Distance and link utilization is what effects latency.

The base RTT between Washington D.C. London, England is roughly 90-100ms no matter what.  Could be a 9.6 Kbps or a DS3.

Now the more traffic you put on the link, the more it will interfere with other traffic.  The higher the latency.

Dark Fiber vs. a dedicate P-to-P circuit will not be that much difference in latency.  It would allow greater bandwidth.

WAN compression modules may or may not help.  Some SAN replication products already do compression, some don't.  However, it will not help with latency or bandwidth utilization.  Well not directly.  What it will do is change how long the link utilized.

If you have 100MB of data to send over a 5 Mbps link,  it would take roughly 160 seconds at 100% link utilization.  Now if you compressed the data down to 20MB it would take 1/5 of the time or 32 seconds at 100% utilization.  These are roughly est., the time it would take would be a bit longer for both.  First because I did not include any overhead from TCP/IP, the overhead of the replication software, or the time it takes to compress/decompress the data.

In both cases the link is 100% utilized, but in one case just for a longer period of time.
0

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
kevinhsiehCommented:
My first question is, do you need SRDF/S, because you probably don't. SRDF/S is really target at financial institutions that can not, under any circumstance, lose a single transaction. Think Wall Street banks and brokerages. The reason why you need low latency is because your local write transaction isn't considered complete until that same write gets written to on the remote array. This is called synchronous replication. Most businesses don't have that kind of requirement. synchronous replication requires that both locations are in the same general metropolitan area because of latency requirements. Log shipping is an asynchronous method, where your second copy can be seconds to minutes behind, or even more. Asynchronous replication doesn't care much about latency, and your replication array can be half way around the globe.

Here is some more information comparing the two.

http://blogs.msdn.com/b/douggowans/archive/2008/12/12/asynchronous-or-synchronous-replication.aspx
0
sunhuxAuthor Commented:
The current customer's setup is log shipping & even with log ship that's
created every 5 minutes, the customer is not satisfied.  I guess I'll have
to achieve something like 1 minute RPO or lower.  Unless we can have
an asynchronous SAN replication or log ship that will give lower than 1
minute RPO, the customer is not satisfied
0
kevinhsiehCommented:
Aysnchronous replication would typically only be a second or two behind as long as you have sufficient bandwidth, and 5 Mb/s should be more than enough. To get a feel for it, you can trial Double-Take, which installs on machine that you are replicating. It doesn't care about what type of storage you have on both ends, or the latency. It can also do compression to lower the amount of bandwidth required. You can do replication on the host level, or at the storage level.
0
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today
Network Architecture

From novice to tech pro — start learning today.