Our application (using a SQL Server 2008 R2 back-end) stores data about remote hardware devices reporting back to our servers over the Internet. There are a few "families" of information we have about each device, each stored by a different server application into a shared database:
static configuration information stored by users using our web app. e.g. Physical Location, Friendly Name, etc.
logged information about device behavior, e.g. last reporting time, date the device first came online, whether device is healthy, etc.
expensive information re-computed by scheduled jobs, e.g. average signal strength, average length of transmission, historical failure rates, etc.
These properties are all scalar values reflecting the most current data we have about a device. We have a separate way to store historical information.
The largest number of device instances we have to worry about will be around 100,000, so this is not a "big data" problem. In most cases a database will have 10,000 devices or less to worry about.
Writes to the data about an individual device happens infrequently-- typically every few hours. It's theoretically possible for a scheduled task, user-inputted configuration changes, and dynamic data to all make updates for the same device at the same time, but this seems very rare. Reads are more frequent: probably 10x per minute reads against at least one device in a database, and several times per hour for a full scan of some properties of all devices described in a database.
Deletes are relatively rare, in fact many cases we only "soft delete" devices so we can use them for historical reporting. New device inserts are more common, perhaps a few every day.
There are (at least) two obvious ways to store this data in our SQL database:
The current design of our application stores each of these families of information in separate tables, each with a clustered index on a Device ID primary key. One server application writes to one table each.
An alternate implementation that's been proposed is to use one large table, and create covering indexes as needed to accelerate queries for groups of properties (e.g. all static info, all reliability info, etc.) that are frequently queried together.
My question: is there a clearly superior option? If the answer is "it depends" then what are the circumstances which would make "one large table" or "multiple tables" better?
Answers should consider: performance, maintainability of DB itself, maintainability of code that reads/writes rows, and reliability in the face of unexpected behavior. Maintanability and reliability are probably a higher priority for us than performance, if we have to trade off.
Don't know about a clearly superior option, and I don't know about sql-server architecture. But I would go for the first option with separate tables for different families of data. Some advantages could be:
granting access to specific sets of data (may be desirable for future applications)
archiving different famalies of data at different rates
partial functionality of the application in the case of maintenance on a part (some tables available while another is restored)
indexing and partitioning/sharding can be performed on different attributes (static information could be partitioned on device id, logging information on date)
different families can be assigned to different cache areas (so the static data can remain in a more "static" cache, and more rapidly changing logging type data can be in another "rolling" cache area)
smaller rows pack more rows into a block which means fewer block pulls to scan a table for a specific attribute
less chance of row chaining if altering a table to add a row, easier to perform maintenance if you do
easier to understand the data when seprated into logical units (families)
I wouldn't consider table joining as a disadvantage when properly indexed. But more tables will mean more moving parts and the need for greater awareness/documentation on what is going on.
The first option is the recognized "standard" way to store such data in a relational database.
Although a good design would probably result in more tables. Relational databases software such as SQLServer were designed to store and retrieve data in multiple tables quickly and efficiently.
In addition such designs allow for great flexibility, both in terms of changing the database to store extra data, and, in allowing unexpected/unusual queries against the data stored.
The single table option sounds beguilingly simple to practitioners unfamiliar with Relational databases. In practice they perform very badly, are difficult to manage, and lead to a high number of deadlocks and timeouts.
They also lead to development paralysis. You cannot add a requested feature because it cannot be done without a total redesign of the "simple" database schema.
Related
I am still struggling with identifying how the concept of table distribution in azure sql data warehouse differs from concept of table partition in Sql server?
Definition of both seems to be achieving same results.
Azure DW has up to 60 computing nodes as part of it's MPP architecture. When you store a table on Azure DW you are storing it amongst those nodes. Your tables data is distributed across these nodes (using Hash distribution or Round Robin distribution depending on your needs). You can also choose to have your table (preferably a very small table) replicated across these nodes.
That is distribution. Each node has its own distinct records that only that node worries about when interacting with the data. It's a shared-nothing architecture.
Partitioning is completely divorced from this concept of distribution. When we partition a table we decide which rows belong into which partitions based on some scheme (like partitioning an order table by the order.create_date for instance). A chunk of records for each create_date then gets stored in its own table separate from any other create_date set of records (invisibly behind the scenes).
Partitioning is nice because you may find that you only want to select 10 days worth of orders from your table, so you only need to read against 10 smaller tables, instead of having to scan across years of order data to find the 10 days you are after.
Here's an example from the Microsoft website where horizontal partitioning is done on the name column with two "shards" based on the names alphabetical order:
Table distribution is a concept that is only available on MPP type RDBMSs like Azure DW or Teradata. It's easiest to think of it as a hardware concept that is somewhat divorced (to a degree) from the data. Azure gives you a lot of control here where other MPP databases base distribution on primary keys. Partitioning is available on nearly every RDBMS (MPP or not) and it's easiest to think of it as a storage/software concept that is defined by and dependent on the data in the table.
In the end, they do both work to solve the same problem. But... nearly every RDBMS concept (indexing, disk storage, optimization, partition, distribution, etc) are there to solve the same problem. Namely: "How do I get the exact data I need out as quickly as possible?" When you combine these concepts together to match your data retrieval needs you make your SQL requests CRAZY fast even against monstrously huge data.
Just for fun, allow me to explain it with an analogy.
Suppose there exists one massive book about all history of the world. It has the size of a 42 story building.
Now what if the librarian splits that book into 1 book per year. That makes it much easier to find all information you need for some specific years. Because you can just keep the other books on the shelves.
A small book is easier to carry too.
That's what table partitioning is about. (Reference: Data Partitioning in Azure)
Keeping chunks of data together, based on a key (or set of columns) that is usefull for the majority of the queries and has a nice average distribution.
This can reduce IO because only the relevant chunks need to be accessed.
Now what if the chief librarian unbinds that book. And sends sets of pages to many different libraries.
When we then need certain information, we ask each library to send us copies of the pages we need.
Even better, those librarians could already summarize the information of their pages and then just send only their summaries to one library that collects them for you.
That's what the table distribution is about. (Reference: Table Distribution Guidance in Azure)
To spread out the data over the different nodes.
Conceptually they are the same. The basic idea is that the data will be split across multiple stores. However, the implementation is radically different. Under the covers, Azure SQL Data Warehouse manages and maintains the 70 databases that each table you define is created within. You do nothing beyond define the keys. The distribution is taken care of. For partitioning, you have to define and maintain pretty much everything to get it to work. There's even more to it, but you get the core idea. These are different processes and mechanisms that are, at the macro level, arriving at a similar end point. However, the processes these things support are very different. The distribution assists in increased performance while partitioning is primarily a means of improved data management (rolling windows, etc.). These are very different things with different intents even as they are similar.
I'm currently designing an application where users can create/join groups, and then post content within a group. I'm trying to figure out how best to store this content in a RDBMS.
Option 1: Create a single table for all user content. One of the columns in this table will be the groupID, designating which group the content was posted in. Create an index using the groupID, to enable fast searching of content within a specific group. All content reads/writes will hit this single table.
Option 2: Whenever a user creates a new group, we dynamically create a new table. Something like group_content_{groupName}. All content reads/writes will be routed to the group-specific dynamically created table.
Pros for Option 1:
It's easier to search for content across multiple groups, using a single simple query, operating on a single table.
It's easier to construct simple cross-table queries, since the content table is static and well-defined.
It's easier to implement schema changes and changes to indexing/triggers etc, since there's only one table to maintain.
Pros for Option 2:
All reads and writes will be distributed across numerous tables, thus avoiding any bottlenecks that can result from a lot of traffic hitting a single table (though admittedly, all these tables are still in a single DB)
Each table will be much smaller in size, allowing for faster lookups, faster schema-changes, faster indexing, etc
If we want to shard the DB in future, the transition would be easier if all the data is already "sharded" across different tables.
What are the general recommendations between the above 2 options, from performance/development/maintenance perspectives?
One of the cardinal sins in computing is optimizing too early. It is the opinion of this DBA of 20+ years that you're overestimating the IO that's going to happen to these groups.. RDBMS's are very good at querying and writing this type of info within a standard set of tables. Worst case, you can partition them later. You'll have a lot more search capability and management ease with 1 set of tables instead of a set per user.
Imagine if the schema needs to change? do you really want to update hundreds or thousands of tables or write some long script to fix a mundane issue? Stick with a single set of tables and ignore sharding. Instead, think "maybe we'll partition the tables someday, if necessary"
It is a no-brainer. (1) is the way to go.
You list these as optimizations for the second method. All of these are misconceptions. See comments below:
All reads and writes will be distributed across numerous tables, thus
avoiding any bottlenecks that can result from a lot of traffic hitting
a single table (though admittedly, all these tables are still in a
single DB)
Reads and writes can just as easily be distributed within a table. The only issue would be write conflicts within a page. That is probably a pretty minor consideration, unless you are dealing with more than dozens of transactions per second.
Because of the next item (partially filled pages), you are actually much better off with a single table and pages that are mostly filled.
Each table will be much smaller in size, allowing for faster lookups,
faster schema-changes, faster indexing, etc
Smaller tables can be a performance disaster. Tables are stored on data pages. Each table is then a partially filled page. What you end up with is:
A lot of wasted space on disk.
A lot of wasted space in your page cache -- space that could be used to store records.
A lot of wasted I/O reading in partially filled pages.
If we want to shard the DB in future, the transition would be easier
if all the data is already "sharded" across different tables.
Postgres supports table partitioning, so you can store different parts of a table in different places. That should be sufficient for your purpose of spreading the I/O load.
Option 1: Performance=Normal Development=Easy Maintenance=Easy
Option 2: Performance=Fast Development=Complex Maintenance=Hard
I suggest to choose the Oprion1 and for the BIG table you can manage the performance with better indexes or cash indexes (for some DB) and the last thing is nothing help make the second Option 2, because development a maintenance time is fatal factor
I'm not a trained DBA, but perform some SQL tasks and have this question:
In SQL databases I've noticed the use archive tables that mimic another table with the exact same fields and which are used to accept rows from the original table when that data is deemed for archiving. Since I've seen examples where those tables reside in the same database and on the same drive, my assumption is that this was done to increase performance. Such tables didn't have more than a about 10 million rows in them...
Why would this be done instead of using a column to designate the status of the row, such as a boolean for an in/active flag?
At what point would this improve performance ?
What would be the best pattern to structure this correctly, given that the data may still need to be queried (or unioned with current data) ?
What else is there to say about this ?
The notion of archiving is a physical, not logical, one. Logically the archive table contains the exact same entity and ought to be the same table.
Physical concerns tend to be pragmatic. The overarching notion is that the "database is getting too (big/slow"). Archiving records makes it easier to do things like:
Optimize the index structure differently. Archive tables can have more indexes without affecting insert/update performance on the working table. In addition, the indexes can be rebuilt with full pages, while the working table will generally want to have pages that are 50% full and balanced.
Optimize storage media differently. You can put the archive table on slower/less expensive disk drives that maybe have more capacity.
Optimize backup strategies differently. Working tables may require hot backups or log shipping while archive tables can use snapshots.
Optimize replication differently, if you are using it. If an archive table is only updated once per day via nightly batch, you can use snapshot as opposed to transactional replication.
Different levels of access. Perhaps you want different security access levels for the archive table.
Lock contention. If you working table is very hot you'd rather have your MIS developers access the archive table where they are less likely to halt your operations when they run something and forget to specify dirty read semantics.
The best practice would not to use archive tables but to move the data from the OLTP database to an MIS database, data warehouse, or data marts with denormalized data. But some organizations will have trouble justifying the cost of an additional DB system (which aren't cheap). There are far fewer hurdles to adding an additional table to an existing DB.
I say this frequently, but...
Multiple tables of identical structure almost never makes sense.
A status flag is a much better idea. There are proper ways to increase performance (partitioning/indexing) without denormalizing data or otherwise creating redundancies. 10 million records is pretty small in the world of modern rdbms, so what you're seeing is the product of poor planning or misunderstanding of databases.
I have been hearing a lot about document oriented data stores like CouchDB. I understand the uses of BigTable like stores such as Cassandra. After reading this question, I was wondering what the conditions would be to merit using a document store?
Column-family stores such as Bigtable and Cassandra have very limited querying capabilities. The application is responsible for maintaining indexes in order to query a more complex data model.
Document databases allow you to query the content, not just the key. It will also manage the indexes for you, reducing the complexity of your application.
Domain-driven design evangelizes the use of aggregates and value objects. As Ayende points out, (complex) aggregates are very natural candidates to be stored as a single document, instead of normalizing them over multiple tables or column families. This will reduce the complexity of your persistence layer. There's also less chance that related data is scattered across multiple nodes, as all the data is contained in a single document.
If your application needs to store polymorphic objects, document databases are also a good candidate. Of course, this could also be stored in Cassandra, but you won't have as much querying capabilities. At least not out of the box.
Think of a document database as a luxurious sports car. It doesn't need a professional driver (read: complex application) to get you from A to B, it has features such as air conditioning and comfortable seats and it will lap the high-scalability track in an acceptable time. However, if you want to set a lap record on the high-scalability track, you will need a professional driver and a highly optimized car (e.g. Cassandra), which lacks features such as air conditioning.
Another feature of CouchDB is that you can create those aggregations, not as documents stored manually, but as views (which are derived from the stored data, and updated automatically.)
This is like power windows, heated seats, or the kicking stereo.
We're thinking about putting up a data warehouse system to load with web access logs that our web servers generate. The idea is to load the data in real-time.
To the user we want to present a line graph of the data and enable the user to drill down using the dimensions.
The question is how to balance and design the system so that ;
(1) the data can be fetched and presented to the user in real-time (<2 seconds),
(2) data can be aggregated on per-hour and per-day basis, and
(2) as large amount of data can still be stored in the warehouse, and
Our current data-rate is roughly ~10 accesses per second which gives us ~800k rows per day. My simple tests with MySQL and a simple star schema shows that my quires starts to take longer than 2 seconds when we have more than 8 million rows.
Is it possible it get real-time query performance from a "simple" data warehouse like this,
and still have it store a lot of data (it would be nice to be able to never throw away any data)
Are there ways to aggregate the data into higher resolution tables?
I got a feeling that this isn't really a new question (i've googled quite a lot though). Could maybe someone give points to data warehouse solutions like this? One that comes to mind is Splunk.
Maybe I'm grasping for too much.
UPDATE
My schema looks like this;
dimensions:
client (ip-address)
server
url
facts;
timestamp (in seconds)
bytes transmitted
Seth's answer above is a very reasonable answer and I feel confident that if you invest in the appropriate knowledge and hardware, it has a high chance of success.
Mozilla does a lot of web service analytics. We keep track of details on an hourly basis and we use a commercial DB product, Vertica. It would work very well for this approach but since it is a proprietary commercial product, it has a different set of associated costs.
Another technology that you might want to investigate would be MongoDB. It is a document store database that has a few features that make it potentially a great fit for this use case.
Namely, the capped collections (do a search for mongodb capped collections for more info)
And the fast increment operation for things like keeping track of page views, hits, etc.
http://blog.mongodb.org/post/171353301/using-mongodb-for-real-time-analytics
Doesn't sound like it would be a problem. MySQL is very fast.
For storing logging data, use MyISAM tables -- they're much faster and well suited for web server logs. (I think InnoDB is the default for new installations these days - foreign keys and all the other features of InnoDB aren't necessary for the log tables). You might also consider using merge tables - you can keep individual tables to a manageable size while still being able to access them all as one big table.
If you're still not able to keep up, then get yourself more memory, faster disks, a RAID, or a faster system, in that order.
Also: Never throwing away data is probably a bad idea. If each line is about 200 bytes long, you're talking about a minimum of 50 GB per year, just for the raw logging data. Multiply by at least two if you have indexes. Multiply again by (at least) two for backups.
You can keep it all if you want, but in my opinion you should consider storing the raw data for a few weeks and the aggregated data for a few years. For anything older, just store the reports. (That is, unless you are required by law to keep around. Even then, it probably won't be for more than 3-4 years).
Also, look into partitioning, especially if your queries mostly access latest data; you could -- for example -- set-up weekly partitions of ~5.5M rows.
If aggregating per-day and per hour, consider having date and time dimensions -- you did not list them so I assume you do not use them. The idea is not to have any functions in a query, like HOUR(myTimestamp) or DATE(myTimestamp). The date dimension should be partitioned the same way as fact tables.
With this in place, the query optimizer can use partition pruning, so the total size of tables does not influence the query response as before.
This has gotten to be a fairly common data warehousing application. I've run one for years that supported 20-100 million rows a day with 0.1 second response time (from database), over a second from web server. This isn't even on a huge server.
Your data volumes aren't too large, so I wouldn't think you'd need very expensive hardware. But I'd still go multi-core, 64-bit with a lot of memory.
But you will want to mostly hit aggregate data rather than detail data - especially for time-series graphing over days, months, etc. Aggregate data can be either periodically created on your database through an asynchronous process, or in cases like this is typically works best if your ETL process that transforms your data creates the aggregate data. Note that the aggregate is typically just a group-by of your fact table.
As others have said - partitioning is a good idea when accessing detail data. But this is less critical for the aggregate data. Also, reliance on pre-created dimensional values is much better than on functions or stored procs. Both of these are typical data warehousing strategies.
Regarding the database - if it were me I'd try Postgresql rather than MySQL. The reason is primarily optimizer maturity: postgresql can better handle the kinds of queries you're likely to run. MySQL is more likely to get confused on five-way joins, go bottom up when you run a subselect, etc. And if this application is worth a lot, then I'd consider a commercial database like db2, oracle, sql server. Then you'd get additional features like query parallelism, automatic query rewrite against aggregate tables, additional optimizer sophistication, etc.