Differences

This shows you the differences between two versions of the page.

--- workloadpatterns [2024_02_28 01:03] – external edit 127.0.0.1
+++ workloadpatterns [2024_03_14 18:13] (current) – removed steven
@@ Line 1: / Line 1: @@
-#Workload Patterns
-##A useful precursor
-Access Anywhere does not provide storage directly, it provides the ability to connect various on-premises and on-cloud storage solutions, unifying them into a 'single pane of glass' access.
-The storage solutions can include, but is not limited to, CIFS/ SMB, Windows Filers, NAS / SAN, Object Storage, Amazon S3, Azure, Google Storage, Dropbox, Office 365, OneDrive etc.
-When Access Anywhere connects to these solutions it does not replicate files, it indexes the remote file store and stored the metadata of items such as filename, size, path, date etc.
-When end users connect to Access Anywhere file listing and navigation is quick because the listing does not need to be pulled from the back end data store. This can have tremendous advantages when the remote listing is very large (we have seen billions of files / objects) and when latency is very high. It can also be used to provide high speed unified search for files across storage repositories.
-The Access Anywhere is non-proprietary and therefore bi-modal which means data can be added directly to the storage solution, it does not need to be processed directly through Access Anywhere. To ensure the metadata that is stored is up to date Access Anywhere can, at intervals, continually index the storage or it can be setup to asynchronously pull new items for the current view the user is at, on-demand.
-The Access Anywhere can also do deeper metadata indexes in which the content of a file is indexed providing much richer metadata and indeed this is what we use for our [[https://storagemadeeasy.com/GDPR/|content discovery PII/PHI]] features.
-## Workload Planning
-In general many customers use a single Access Anywhere deployed in a HA fashion, that maybe GEO deployed and/or use the [[sitelink|Access Anywhere's SiteLink]] feature.
-We do have some customers in the governmental, Medical / Genomics, and Research verticals that work with extremely large data sets / data lakes and have many different use cases.
-The way in which Access Anywhere cache data in these scenarios can be very advantageous because access and search  by any other means can be very slow or in some extreme cases when Object Storage is used and prefixes are shallow, almost impossible due to timeouts.
-Quite often such cases are read heavy often with files being added directly to the underlying stores and users requiring direct desktop access through Access Anywhere's direct desktop drives or sync.
- In parallel it is not unusual for such companies to also have document repositories whose access characteristics are different, often very write heavy. In these circumstances we recommend system administrators and architects of Access Anywhere consider the tenant partitioning of Access Anywhere to segment / partition metadata or consider deploying a separate Access Anywhere node.