srdf

Read:

NAS
\\who\ash\ING\Storage Migrations\SRDF\General Ref\Symmetrix Remote Data Facility (SRDF) Connectivity Guide.pdf

\\who\ash\ING\Storage Migrations\SRDF\General Ref\Symmetrix Remote Data Facility (SRDF) for VMAX 40K, VMAX 20K-VMAX, DMX Series.pdf

 

External Sites

https://powerwindows.wordpress.com/tag/srdfce/
https://msmvps.com/blogs/jtoner/
 

  • The actual SRDF work queue (I/O jobs) is distributed to individual SRDF directors' local cache (queues) by the front-end host adapters (HA) for SRDF/S and SRDF/A, and by the disk adapters (DA) for adaptive copy and SRDF/AR.
  • The distribution of the work queues to the SRDF directors eliminates multiple memory accesses and lock/unlock cache slot contention.
  • Each SRDF link will therefore operate independently at its own speed and latency. If the SRDF links are on a balanced network (speed and latency), then the SRDF workload will be well-balanced between directors.


SRDF director hardware
This section contains the following information on the Symmetrix DMX Series, and earlier, systems

SRDF director hardware consists of a director and adapter board set. This hardware provides the communications physical layer for SRDF data and information exchanges between Symmetrix systems. SRDF director hardware includes any of four types of director/adapter board sets, depending on the protocol:

  • ESCON remote adapter (RA)

        Note: ESCON is not supported with Enginuity version 5874.

  • Fibre Channel remote adapter (RF)
  • GbE remote adapter (RE)
  • Multiprotocol Channel Director (MPCD), supported with Symmetrix DMX series, and configurable for SRDF over Gigabit Ethernet (GbE). With DMX-3, Fibre Channel was added. The MPCD is also configurable for iSCSI, FICON, or combinations of these protocols.



Fibre Channel director (RF)
Enginuity version 5x66 and later supports SRDF Fibre Channel emulation.

Note: All Fibre Channel director ports have seven (7) Buffer-to-Buffer Credits (BB_Credits).

BB_Credit management occurs between any two Fibre Channel ports that are connected. For example:
◆ One N_Port and one F_Port
◆ Two E_Ports
◆ Two N_Ports in a point-to-point topology

The standard provides a frame-acknowledgement mechanism in which an R_RDY (Receiver Ready) primitive is sent from the receiving port to the transmitting port for every available buffer on the receiving side. The transmitting port maintains a count of free receiver buffers, and will continue to send frames if the count is greater than zero.


Fibre Channel link initialization (FC timeouts)
For link initialization of a Fibre Channel port, Fibre Channel specifications state that the maximum tolerable response time for a response is 100 milliseconds round trip time. This timeframe coincides with the limited timeframe of the Receiver-Transmitter Timeout Value (R_T_TOV), which is how long an FC-port listens for a link response to a link service before an error is detected.

For all Fibre Channel-only ISL configurations, R_T_TOV is a mandatory limitation.

SRDF with SiRT
SiRT (single round trip) for Fibre Channel SRDF directors (RFs) was introduced in Enginuity version 5772 for SRDF/S mode only. It is dynamically enabled for SRDF/S links > 12 Km for block sizes up to 32K on DMX and 64K on VMAX. SiRT is compatible with fast write/write acceleration switches and extenders, as it will measure link latency and disable automatically if connected to these devices. As a best practice, it is recommended that either the EMC SiRT feature or the third-party fast write feature should be used. Both should not be enabled simultaneously.

The Fibre Channel SiRT feature for the Fibre Channel director can be set to Off or Automatic. When set to Automatic, this feature will only accelerate write I/Os using criteria based on latency and I/O size.

Note: SiRT defaults to Automatic (Enabled) in VMAX for SRDF/S. It automatically disables for Adaptive Copy and SRDF/A.


Fibre Channel director considerations
SRDF static flow control

In addition to the Fibre Channel mechanisms and restrictions, SRDF static (legacy) flow control can be governed by the SRDF Flow Control Algorithm which allocates the number of I/Os that can be sent from the SRDF layer. This feature, introduced for SRDF over Fibre Channel in Enginuity version 5671, should be enabled for all extended distance asynchronous links (SRDF/A, SRDF/AR, and Adaptive Copy).It should not be used for SRDF/S and is disabled by default.


The SRDF algorithm first tracks, analyzes, and monitors for certain I/O conditions. It then determines the optimal number of I/Os (jobs) to send out on the SRDF links, based on certain RTT windows: 0-40ms; 40-120ms; and over 120ms. The algorithm maintains a manageable number of I/Os (job queues) while making efforts to avoid time-out scenarios.

Network quality
The network quality of an SRDF link is now tested with Enginuity version 5875 ucode and later. VMAX systems on both sides of the link must be running 5875 or later to take advantage of this feature. This feature takes the following actions:
◆ Qualify a questionable SRDF connection before allowing traffic to be sent
◆ Gradually slows down intervals between unsuccessful SRDF link re-establishment attempts
◆ Puts bouncing SRDF links out of service

A qualification test for any SRDF connection can be manually triggered. This feature is supported for both GbE and Fibre Channel SRDF links.

Compression
Beginning with Enginuity version 5874, software compression is available with the Fibre Channel port. Software compression can be enabled at the SRDF group level and SRDF mode of operation.

Both sides of the link must support software compression.

FCIP
FCIP is a tunneling protocol that transports Fibre Channel ISL traffic.FCIP uses IP as the transport protocol. An FCIP link tunnels all SRDF traffic between locations and may have one or more TCP connections between a pair of IP nodes for the tunnel end-points. From the Fibre Channel fabric view, an FCIP link is an ISL transporting all Fibre Channel control and data frames between switches, with the IP network and protocols transparent. One can configure one or more ISLs between Fibre Channel switches using FCIP links.


Proper network design is required given the speed and bandwidth differences between Fibre Channel and a typical IP network. This design should take into consideration the management of congestion and overload conditions. Fibre Channel fabrics connected via FCIP links are not isolated unless isolation is achieved using one of the fabric isolation methods mention earlier.

N/W Topology

Network quality
Latency should be as low as possible, given the distance separating the sites, with a minimum number of hops. SRDF has been tested against what is referred to as the Distance Envelope. This involves a series of tests to make sure that SRDF will continue to function as long as the latency does not exceed 200ms RTT and packet loss does not exceed 1%.

A quality link should have packet loss no greater than 0.01%. SRDF can tolerate packet loss up to 1%. If the packet loss is greater than 0.01% it is not considered a quality link and will typically result in throughput that is severally impacted. Packets should not be delivered out of order since this can cause the SRDF link to bounce or result in a major drop in performance. The rapid variation in the latency on the network (jitter) should also be very low. It is recommended that all of these aspects be covered in an SLA (Service Level Agreement). An SLA is a contract from the ISP (Internet Service Provider) guaranteeing the quality of the link.