Recently I did a deployment where a great and awesome longtime customer added a new site in Canada. I flew up to Canuckland to install new VMware servers, new Netapp 2240, and a Riverbed Steelhead. There was a partner Riverbed Steelhead here in PA at the customers office. This was a great deployment.
To digress a little… The bad news is, on this trip, the car rental place gave me a Yaris. The worse news is, there was a massive snowstorm. Have you ever tried to drive a little Yaris in a snowstorm? I thought I was going to die. Luckily, there was a giant parking lot across from the hotel which was about 4-5 football fields in size, and I got to spend a few hours doing donuts in the snow and ice. You have to always find a way to have fun when traveling for business.
Back to the real story.
The Netapps are hosting all the VMware environment. Top to bottom virtualization. Everything old or still physical P2V’d, everything already virtualized moved to new gear, on ESXi 5.1 and sitting on the Netapp FAS2240. Replicating back from Canuckland to Pennsylvania. I love fully virtualized environments.
We will be setting up VMware SRM (Site Recover Manager) in the CA office, with single button click failover automation back to PA.
I quickly was able to setup replication between the NetApps at each of the customer sites with Snapshots and SnapMirror. With NetApp deduplication technology, the customer was seeing a great reduction in used space on disk. Though, hundreds of GB of deduped data, with heavy daily change on the email and SQL servers crunching away is mucho data-o to send over the WAN.
I am snapshotting the customers SAN volumes and replicating on schedules ranging from hourly to every 15 minutes (depending on the RPO/RTO needed by the customer).
Well, the customer has a good connection between sites.. but just not good enough to send all change from all VMs every 15 minutes, and have the rest of the WAN be undisturbed. Enter Riverbed Steelhead WAN Accelerators…
Riverbed Steelheads dedupe, rewrite chatty protocols, fix stuff within the TCP stream, and add in some magic mojo juice in order to accelerate traffic between your sites, be it HQ and branch offices, or even HQ and remote users.
Once I got everything worked out and configured, their WAN got a major injection of Honey BooBoos GoGoJuice.
Yes Faster than that!
How much faster? Well, the cold optimization quickly jumped up to 97% WAN reduction and a 23.6X increase in throughput!!!!
23.6X increase. Say it with me. 23.6x increase. And again 23.6x!
Imagine if for just 25MB of pointers to data in the dedupe datastore, you could send 1GB of data across the wan!
OMFG it’s a beautiful thing! Wow!
Look at this —>>>
Makes me all tingly!
<— a peak reduction of 98%!
Plus, imagine your laptop users. IF you used a WAN accelerator like the Steelhead with the Steelhead Mobile client, your VPN users would pre-cache and dedupe, and CIFS/SMB shares would be less chatty because the units clean up the protocols, and only sends blocks of data that change and are not in the dedupe pool. Huge performance jump, and happy users!
This is just a small part of how I help my customers every day!
If you are interested into taking your company into 2013, let me know. This is just a day in the life of a Professional Services Engineer! Working with the most awesome technologies like Riverbed, Netapp, and VMware!
[asa]1118407911[/asa]
Comments are closed.