Failover Behavior: Secondary CME Gateway
Secondary CME Gateways offer traders the opportunity to continue trading should one or more routes to the CME Group fail. When a CME Gateway fails in a failover environment during trading hours, all current orders remain as working orders in the market
You can use secondary, redundant CME Gateways under the following conditions:
When you set up secondary CME Gateways on a network, traders can log into another CME Gateway to continue trading if the CME Gateway fails.
After reconnecting to the CME Group in this manner, traders have no access to their current working state or position in the market (i.e., historical information). This information remains on the failed TT Gateway in its log and table files.
If you intend on installing secondary CME Gateways for disaster recovery purposes, traders must use unique Member IDs when logging into the different TT Gateways. This is a CME Group restriction. For example, if trader Jim wants access to the exchange through CME -A and CME-B, he must have a Member ID for CME-A and a different Member ID for CME-B. In this manner, his login on each TT Gateway is different.
Gateway Failover Behavior
After a trader closes and then reopens X_TRADER® and then logs into the alternate CME Gateway, the TT system exhibits the following behavior:
- Fills: X_TRADER® does not have access to any fills that occurred previous to the trader connecting to the secondary CME Gateway because the *.bof files do not hold the fills for the new trader.
- Orders: After traders reconnect, they can submit orders. However, they do not have access to any active orders that were in the market at the time of the previous connection being lost. This information is on the failed TT Gateway.
Secondary Gateway Recovery: Software
Refer to Single Gateway Failure for details on CME Gateway software recovery.
Secondary Gateway Recovery: Hardware
Refer to Single Gateway Failure for details on CME Gateway hardware recovery.
Example 2: Unsuccessful Failover
Having recovered from the hard drive crash in the previous Example Scenario 1 – Single CME Gateway, Quick Financial Services (QFS) now has two CME Gateways (referred to as A and B) conforming to the setup criteria for Redundant CME Gateways. During the day, the exchange connections on gateway A fail. Gateway A immediately loses its connection to the exchange and shuts down. All traders on gateway A lose their connection to the exchange and gateway A displays as red in Guardian across the network. Prices become stale, traders stop receiving fills, and they are unable to trade.
Jim first documents his position. With his documented position in hand, Jim calls CME Group and cancels all outstanding orders. He then closes X_TRADER®, reopens it, and logs into gateway B to continue trading. Jim already has previously saved workspace on this gateway and is not prompted to convert. When Jim logs into gateway B, none of his fills or orders refresh. He does not have access to his historical fills. He resumes trading.
The IT staff at QFS restarts gateway A. Traders can begin logging into gateway A, and trading continues as normal. However, if a trader submits orders on gateway B, then logs off B and into gateway A, that trader does not have access to trading activity (orders and fills) managed by gateway B.
Example 3: Unsuccessful Failover
QFS has two TT Gateways (called A and B) with 40 traders – 20 traders on each gateway. During trading hours, software on gateway A fails. All 20 traders using that gateway begin logging into gateway B. However, trading is particularly heavy that day and the software on gateway B begins experiencing noticeable latency.
To resolve this situation (latency caused by heavy volume) from occurring in the future, QFS calls CME Group and orders three more SLE configuration connections. QFS also calls TT for installation of three additional TT Gateways. This provides better failover and supports future growth of the company.