Storage Networks

EMC World 2015: Take Control with IP Extension

by mdetrick on ‎05-04-2015 07:30 AM (7,810 Views)

Good morning EMC World 2015! Here we are again in Las Vegas pursuing the latest technological advancements for our storage and cloud environments.

 

This year EMC and Brocade are proud to present something entirely new: IPEX (IPEX = IP Extension) on the Brocade 7840 (EMC Connectrix MP-7840B). You say, Hummm… Why do I need extension for IP? I’m sure you’re thinking, can’t I just handoff my IP storage flows to the network? Of course, I hope the Network Admins meet their SLAs (Service Level Agreements) so I can meet mine! Who out there wants to have SLA responsibility that in turn relies on the network SLA, of which you have no control? Yep, that’s what I thought… No One!

 

Actually, none of this is conjecture! Brocade has been researching this very technical, cultural and political issue between Storage & Mainframe Admins and their cross functional Network Admins for a couple of years now. The statement above has been reinforced over and over again and validated by 90% of customers throughout the world at enterprises of every type. Working with the networking guys can be challenging, to say the least. OK, so what is Brocade and EMC doing about it? How can Brocade and EMC make Storage and Mainframe Administrators’ jobs easier and more efficient? We all know the adage, “do more with less.” Good news, the answer is IPEX on the Brocade 7840 Extension Switch!

 

IPEX is not a mere luxury, although, after implementing it in your environment you will find it luxurious. IPEX meets real needs and requirements. There are 5 pillars unto which IPEX stands:

 

  • High Performance (Tuesday May 5, Part 2 of 3 in this series)
  • Operational Excellence (Wednesday May 6, Part 3 of 3 in this series)
  • Superior Availability (Wednesday May 6, Part 3 of 3 in this series)
  • Strong Security (Wednesday May 6, Part 3 of 3 in this series)
  • Easy Deployment (Wednesday May 6, Part 3 of 3 in this series)

Envision you have various IP storage end-devices that handoff an IP connection to the black-hole that is your IP network. These devices may include a wide swath of EMC products such as SRDF, Data Domain, RecoverPoint, database host replication, iSCSI, other host-based replication, VM replication like SRM, IP backup applications, NAS replication, IP tape applications (i.e., DLm and others) and more… There can be hundreds of IP storage flows between data centers across the WAN. User based traffic is not in scope here. Your traffic is on the IP network and things can break. 

 

Image1.png

 

IP storage applications can benefit from the same network quality as provided by FC fabrics. What do Brocade fabrics offer? Visibility, flow management, auto-corrections, flow rate limiting, QoS, encryption, compression, SLA reporting and end-to-end diagnostics, and much more… To achieve the same quality local IP network (Ethernet Fabric), use the EMC Connectrix VDX 6740 switches as an isolated backend Ethernet Fabric. Configured and managed the same as your FC fabric. You don’t have to be a Network Admin to build and operate Connectrix VDX Ethernet Fabric; however, that’s a different blog for a different time. Here, we are focused on extension between data centers over the WAN.

 

Back to the IP storage handoff problem, in most cases what you get from the IP network after handing off a connection is:

 

The Problem is…

  • No visibility
  • No bandwidth management
  • No flow management
  • No flow rate limiting
  • No QoS
  • No encryption
  • No compression
  • No SLA reporting
  • No end-to-end diagnostics 

Can the IP network do these things? Well, yes to a small extent. Is that going to happen for you? No, probably not. Most IP networks are not flexible, exceedingly complicated and operationally a nightmare; hence, the pushback you will receive when asking for such things. Now, what usually comes to mind is, “well we can’t just leave the IP guys on the side of the road…”, and that’s true and we certainly don’t advocate that. We still have to use their physical IP infrastructure; nonetheless, we can create a logical storage-based network within. That logical network is then managed by the Storage or Mainframe Administrators and provides everything on the above problem list.

 

Here is some pertinent and common customer feedback… When it comes to the IP network, from a Storage or Mainframe Administrator’s point of view, the network is either “sick” or “dead”. Dead, we all know what that is and equals hard-down/offline. Sick is much tougher though. When is it sick? Degraded to a varying degree? You think to yourself, “seems we should be getting better performance but we’re not?” Frustrating!

 

You ask yourself, “Where’s the problem?” Local end-device? Remote end-device? Local IP network? Remote IP network? Local Brocade box? Remote Brocade box? The carrier? So frustrating! We must know where the problem is before we can start to address the problem. Time is of the essence.

 

Now you think, “What do I do when I believe there’s an IP network problem?” You call someone in networking and it often goes like this... “Hey Joe, I think I should be getting better performance on my IP storage app X. Can you please take a look into that for me?” Joe, “Sure, no problem. Just give me a bit and I’ll call you back in a few.” Days go by and there is no response, so you follow-up. “So Joe, what did you find out?” Joe, “Find out about what?” pause… “Oh, yeah… It’s fine, I didn’t see anything.” Now, what? Who are you going to call? Very frustrating!

 

You must take control and arm yourself with some visibility and reporting so that you can see exactly what aberrant behavior is occurring across the network and exactly when. For example, Brocade Network Advisor has a TiVo like recording feature in which you can playback network events to actually watch failures as they occur. Bottom-line, hold the IP network guys to their SLA and document every time an SLA hasn’t been met. If there are numerous SLA violations, who knows, maybe you can justify your own Connectrix VDX 6740 backend IP storage network, which will connect IPEX as well. Enjoy the highest availability and performance possible.

 

With IPEX you can create an underlying logical IP storage network that you (Storage or Mainframe Admins) manage across the IP infrastructure between two data centers. If you want trunking, QoS, encryption, compression and bandwidth management, that is completely under your control. Oh, and you don’t have to be a Network Administrator to configure such things. It’s all very simple to configure by just adding attributes to the tunnel.

 

Image2.png

 

Brocade and EMC are redefining what it means to Take Control. So, do it!

 

Tomorrow we’ll explore how to gain huge performance improvements by using Brocade IPEX. Learn about these advanced technologies: TCP Acceleration, WAN Optimized TCP, vWindows, Adaptive Rate Limiting (ARL) and others…

 

To do list: Read this blog tomorrow, go to the Brocade booth, have fun at EMC World and stop by my technical session:

Storage over Distance from SRDF to SyncIQ

Date: Wednesday, May 6, 2015

Time: 12 - 1 pm (Pacific)

Location: Lando 4205

Labels