My organization wants to use Microsoft Text Analytics API for sentiment analysis. But my employer concern is that MS will be using that data for the live training of their sentiment engine. Is this the case?
See the Microsoft Trust Center for your answer.
How we manage your data
With Microsoft, you are the owner of your customer data.
Microsoft will use your customer data only to provide the services we
have agreed upon, and for purposes that are compatible with providing
those services. We do not share your data with our
advertiser-supported services, nor do we mine it for marketing or
advertising. If you leave the service, we take the necessary steps to
ensure the continued ownership of your data.
Related
Question: Does Unity Catalog in Azure Databricks have the feature of classifying assets? If so, can someone please provide links to online documentation on this feature in Unity Catalog? Please see the context below:
Unity Catalog is the Azure Databricks data governance solution for the Lakehouse. And Microsoft Purview is a data governance solution for on-premises, multicloud, and software-as-a-service (SaaS) data.
The Data classification in the Microsoft Purview governance portal is a way of categorizing data assets by assigning unique logical tags or classes to the data assets. Classification is based on the business context of the data. For example, you might classify assets by Passport Number, Driver's License Number, Credit Card Number, SWIFT Code, Person’s Name, and so on.
When Purview scans your data storage it identifies/classifies the data assets based on the data that matches the classification (logical tags such as Passport Number, Credit Card number etc.)
When you classify data assets, you make them easier to understand, search, and govern. Classifying data assets also helps you understand the risks associated with them.
Unity Catalog right now doesn't provide such classifications. There were roadmap presentations that included the notion of attribute classes that could be also used to classify objects and assign permissions for accessing data marked as specific attribute (attribute-based access control). But right now there is no specific timeline for that future.
Although it's now possible to use SQL to add comments on both table & column level, although it may not be very convenient.
P.S. Potentially Purview <-> Unity Catalog integration will allow use use Purview's UI to classify data in UC, but I haven't seen such integration in
I have a general question related to the Azure FHIR server. Specifically, are there "technical" reasons to use it to store electronic health records (EHR)?
People told me major healthcare providers are using FHIR server to store EHR. I understand the business reason. But, technically, Azure cloud is HIPAA compliant. Anything stored and run in Azure is HIPAA compliant. So, if there are other ways such as SQL Server and Data Lake to store EHR, why do we have to store them in FHIR server?
Another reason I ask this question is that my team feels it is not efficient to store EHR via API calls, and it is difficult to query data in JSON format for reporting. So, they feel if Azure is HIPAA compliant they would prefer to store EHR in SQL Server or Data Lake for ease of data management. Both also provide role based security to limit unauthorized access with encryption.
No sure if I miss anything. Can anyone tell me the "technical" differences / advantages of storing EHR in FHIR server? Thanks in advance for your replies.
FHIR is primarily for health data interchange using more modern formats such as JSON, so you don't necessarily have to store the EHR data in FHIR format; you can instead have a layer that transforms your EHR data (in SQL or any other format you prefer) to FHIR format (JSON/XML/RDF) as and when needed.
Without a modern standard like FHIR, for the most part, it is very costly and time consuming to design and develop "interfaces" for healthcare data interchange;
Does that help clarify the "technical" need for FHIR?
Since ADC is provided by MS as SaaS to customers, is MS taking backups of the dataset and business glossary? If yes, how often and how can a customer get access to the backups for recovery purposes?
Unfortunately, there is no explicit backup/restore feature available for catalogs.
I would suggest you to vote up an idea submitted by another Azure customer.
https://feedback.azure.com/forums/906052-data-catalog/suggestions/33125845-azure-data-catalog-backup-feature
All of the feedback you share in these forums will be monitored and reviewed by the Microsoft engineering teams responsible for building Azure.
The closest way to achieve this with current functionality is to use the Azure Data Catalog REST API to extract all assets and persist them locally (and re import them manually later).
There is a sample application available that demonstrates this technique: Data Catalog Import/Export sample tool.
I've been researching the new unified API for office365 (graph.microsoft.com). Currently one can authenticate to it using an organisationalID and it will return information from emails-from-exchange/OneDrive4Business/Skype4Business(future)/calendar-from-exchange/exchange-contacts/etc...
The ability to surface such a plethora of information from one endpoint strikes me as being hugely valuable. In the future I'd really like to see information relating to Azure be surfaced in here also. For example, I'd love to use graph.microsoft.com to access lists of:
Azure storage accounts
Azure SQL DB servers and databases
Azure SQL DB firewall rules
Azure ML experiments
Countless other things
Are such things on the roadmap?
That is a great question, and it's great to see such enthusiasm around a single endpoint surfacing this information. We're not sure when Azure resource providers would be a part of the unified API, but it would be great to have them. There are a number of other teams who are very interested, and in general we've been pretty focused on user or person centric developer experiences here as you should be able to see from Yina Arena's //Build or Ignite presentations.
I'm just learning about Azure so forgive me for my naivety. I work for a federal government that would be very hesitant to have their applications and data hosted in another country. Could a local company offer "Azure" services? i.e. could software developers in a government department build their applications and deploy them to the Azure cloud, ensuring that their data stays within the country? Or would they have to look at a non-Microsoft cloud provider?
Data and Compute will reside in the datacenter you specify. Blobs, Tables and Queues are also backed up automatically to a paired data center:
San Antonio <--> Chicago
Dublin <--> Amsterdam
Hong Kong <--> Singapore
You can opt-out of cross-datacenter data backup if data sovereignty becomes an issue. Once opted-out, data would only be in the specified data center, and you'd need to handle DR on your own (by possibly backing up data to on-premises storage).
Aside from those 6 datacenters, Fujitsu runs a Windows Azure data center in Japan. See this press release for more info.
Yes, when you create your Azure service you can specify what region (of the country) it runs in.
I'm not sure if you know this, but the Federal CIO (Vivek Kundra) is really pushing hard for Agencies to move to the cloud. You might want to check out Info.Apps.Gov for guidelines on the Federal Cloud initiative and resources for what you can and can't do.
To answer your immediate question: No. Only MS hosts Azure to my knowledge. I do know that Amazon is bending over backwards however to accommodate Government clients and you can control which datacenters are used on that service. MS appears to have a similar capability per the other answer to this question.
As far as I can tell, these are the only locations:
http://en.wikipedia.org/wiki/Azure_Services_Platform#Datacenters
If they're that concerned about data security though, they should deal directly with Microsoft, not buy Azure services that same way a client usually would. Microsoft may be able to arrange something depending on budget (but probably not).
Edit: What I'm basically saying is, Microsoft is not going to arbitrarily do special licensing. Meaning you either need a large enough budget to convince MS to build a data center in your country, or you need some other way of convincing MS to allow Azure services hosted in your country. Also, I hate to sound paranoid, but if you're worried about America seeing your data, you likely should avoid Ameican companies.
If there isn't a Windows Azure Data Centre in the relevant country, but you still want to use Azure, you'll need to look at a hybrid cloud model where data remains resident in a private cloud. However, in-flight data can still present complications for some organisations and Azure may not be the right answer in all cases.
If you like, I can talk about it some more using Chat. The company I work for specialises in just these cases and has the only production Windows Azure data centre that isn't owned by Microsoft (and isn't in the US). Probably best not go into further specifics here, though, for fear of my answer looking like pure spam!