This article summaries 2012 Chinese ATE activity: "Implementation of SRDF and Symmetrix- based remote replication technology solutions ". The original thread is https://community.emc.com/thread/165127.

Question 1：

How many types of SRDF are there?  

Answer：

SRDF provides a remote replication solution between two or more Symmetrix arrays. Overall, there are three types of SRDF:

(1) SRDF/S, synchronous mode. Updating the data between the production site and remote site in a synchronous mode. The production site acknowledges the front-end I/O only after receiving the acknowledgement from the remote site.

(2) SRDF/A, asynchronous mode. Updating the data between the production site and remote site in an asynchronous way. The front-end I/O on the production site will be acknowledged immediately. The data replicating process runs in the background asynchronously.

(3) SRDF/DM, data mobility mode. Combining SRDF with Timefindertechnology, which is used for data replication and disaster recovery.More specific types are derived from the three types of SRDF listed above. For example, SRDF/Star, Concurrent SRDF, Cascaded SRDF, SRDF/AR (Automated Replication), SRDF/CG (Consistency Groups), SRDF/CE (Cluster Enabler), etc.

---

Question 2：

Both SRDF/AR and Cascaded SRDF are multi-site replication solutions. What’s the difference between them?

Answer:

SRDF/AR（SAR）combines SRDF with the Timefinder technology to provide automatic and periodic remote replication. It provides data consistency and has low bandwidth requirement. There are two kinds of topologies in SRDF/AR, single-hop and multi-hop. In a single-hop environment, the copy path is from local R1/BCV pair (1) to the SRDF pair (2) to the remote BCV pair (3). In a multi-hop environment, the copy path is from local SRDF pair R1 (1) to the remote BCV pair (2), then to target SRDF pair R2 (3), finally to the target BCV (4). The BCV represents the local copy of the data. Cascaded SRDF (SRDF/CR) is derived from SRDF/AR. It introduces a new SRDF structure R1->R21->R2, which indicates the second site is the target of the first site as well as the source of the third site.  The benefit is that if the first site is down, the second site can still send data to the third site, which provides better RPO and RTO for the third site. It can provide zero data loss when R1 fails. The SRDF link between R1 and R21 usually runs in synchronous mode.   

---

Question 3:

What’s the difference between SRDF and RecoverPoint (RP) in theory?

Answer:

For RP, there are two types of disaster recovery: CDP and CRR. CDP is a kind of local synchronous solution and CRR is a kind of remote asynchronous solution. CRR works in the following way:

1. The Host requests a write I/O against the LUN protected by RP.

2. The Spliter (can be located on the host, switch or storage array) captures the I/O and makes a copy to the local RP appliance.

3. The local RP appliance receives the I/O and sends an acknowledgement to the splitter immediately.

4. After receiving acknowledgement from both the local RP appliance and the LUN, the splitter sends the acknowledgement to the Host.

5. The local RP appliance compresses the data and sends it to the remote RP appliance over the IP network. The local RP appliance receives and decompresses it and saves it in the log volume and remote volume.

In the RP solution, the source or the target storage array can be a non-EMC product. In SRDF, both arrays must be EMC Symmetrix. There is no need to install components on the host or the switch in SRDF.     

---

Question 4:

What are the things that you need to pay attention to when configuring the SRDF link?

Answer:

There are some best practice suggestion from the CONNECT GUIDE, that you can follow listed below:

• Since many SRDF/S configurations use Fibre Channel and DWDM technology, ensure that the customer is using quality fiber and connectors and that the DWDM and Fibre Channel switch optics are properly configured and well within optical link budgets. Also ensure that the latency is minimized by choosing fibre that runs carefully and by minimizing network equipment (switches, extenders, routers, etc) overhead and delay.

• For Metro SRDF/S implementations using DWDM or extended distance WAN connectivity, it is not recommended to extend Fibre Channel fabrics containing local host/storage traffic between data centers along with SRDF/S. The best practice is to isolate local host/storage fabrics within data centers so that link outages cannot cause disruption to the host connection to the SAN. The recommendation is to use Fibre Channel features, such VSANs or FC routing, to prevent merging of the local and remote fabrics.                                        

---

Question 5:

What are the SRDF management tools? 

Answer:

Both Solutions Enabler and SMC can configure and manage SRDF. Solutions Enabler is in the CLI and SMC is in the GUI. All of the documents about both tools can be found on EMC Online Support.

Author: Jun Tan