Kudos to Symantec for publishing an API for Symantec Enterprise Vault (EV), and indeed for rigorously certifying vendor products that want to use it. In fact, we were thrilled to be the first migration vendor to be STEP (Symantec Technology Enabled Program) approved to use the API some 5 years ago.
But, in addition to our API connector we also have a direct connector into EV which some of our partners prefer to use.
The reason? Customer projects have shown that the direct connector can deliver performance up to 10 times faster than using the API when the EV archive is ‘busy’. It also has a lower impact on the EV environment (as it does not go via its servers) and it can also access a non-operational EV archive store/database.
Sounds great – but what are the down-sides of going direct?
Well first and foremost, Symantec – quite rightly – is unable to guarantee support for any migration that uses a non-API approach. Whether you’re upgrading or restructuring EV, migrating to Enterprise Vault.cloud, or even moving away from EV, the API is the only way to go to have that guaranteed support. Next, the API is a sure-fire way to successfully access items that reside on a long list of storage devices other than generic CIFS/SMB storage. The most popular devices such as EMC Centera, we can access directly, but there are many others (from vendors including Amazon, Rackspace, Dell and Hitachi) which rely on the Enterprise Vault Streamer API to act as a go-between.
The API also insulates data access from other ‘un-predictable’ scenarios, such as unusual database schemas.
So a Direct-only approach cannot be the only way to ‘roll’.
The problem our partners had was that (prior to 6.5) they had to decide which connector to configure: Direct or API. Unpicking a migration that started with a direct connector, but which ran into problems and needed to use the API was do-able, but challenging.
A New Hybrid Approach
Our new hybrid connector introduced with Transvault Migrator 6.5 released this week delivers the best of both worlds: It automatically switches from using the direct connector to the API connector whenever it encounters an item it can’t access properly.
The net result is that your migration goes at full velocity whenever it can, but without risking data loss for any items that need the API to access them.
But we hasten to add, we always encourage our migration partners to err on the side of caution, so you can still elect to use a fully API connector. It’s also worth noting that our API connector scales in parallel across multiple servers to give the industry-leading API performance too.
Better safe (and slightly slower) than sorry!