🔥 Get a free month of Premium Access. Use code: FREEAUG at checkout

10 Client-side Storage Options and When to Use Them

    Craig Buckler
    Share

    Storing and manipulating data in the browser — also known as client-side storage — is useful when it’s not necessary or practical to send it to the web server.

    Situations for storing and manipulating data in the browser include:

    • retaining the state of a client-side application — such as the current screen, entered data, user preferences, etc.
    • utilities which access local data or files and have strict privacy requirements
    • progressive web apps (PWAs) which work offline

    Here are ten options for storing browser data:

    1. JavaScript variables
    2. DOM node storage
    3. Web Storage (localStorage and sessionStorage)
    4. IndexedDB
    5. Cache API (don’t use AppCache!)
    6. File System Access API
    7. File and Directory Entries API
    8. cookies
    9. window.name
    10. WebSQL

    This article investigates these ten different ways to store data in the browser, covering their limits, pros, cons, and the best uses of each technique.

    Before we browse the options, a quick note about data persistence …

    Data Persistence

    In general, data you store will either be:

    1. persistent: it remains until your code chooses to delete it, or
    2. volatile: it remains until the browser session ends, typically when the user closes the tab

    The reality is more nuanced.

    Persistent data can be blocked or deleted by the user, operating system, browser, or plugins at any point. The browser can decide to delete older or larger items as it approaches the capacity allocated to that storage type.

    Browsers also record page state. You can navigate away from a site and click back or close and re-open a tab; the page should look identical. Variables and data regarded as session-only are still available.

    1. JavaScript Variables

    metric comment
    capacity no strict limit but browser slowdowns or crashes could occur as you fill memory
    read/write speed the fastest option
    persistence poor: data is wiped by a browser refresh

    Storing state in JavaScript variables is the quickest and easiest option. I’m sure you don’t need an example, but …

    const
      a = 1,
      b = 'two',
      state = {
        msg:  'Hello',
        name: 'Craig'
      };

    Advantages:

    • easy to use
    • fast
    • no need for serialization or de-serialization

    Disadvantages:

    • fragile: refreshing or closing the tab wipes everything
    • third-party scripts can examine or overwrite global (window) values

    You’re already using variables. You could consider permanently storing variable state when the page unloads.

    2. DOM Node Storage

    metric comment
    capacity no strict limit but not ideal for lots of data
    read/write speed fast
    persistence poor: data can be wiped by other scripts or a refresh

    Most DOM elements, either on the page or in-memory, can store values in named attributes. It’s safer to use attribute names prefixed with data-:

    1. the attribute will never have associated HTML functionality
    2. you can access values via a dataset property rather than the longer .setAttribute() and .getAttribute() methods.

    Values are stored as strings so serialization and de-serialization may be required. For example:

    // locate <main> element
    const main = document.querySelector('main');
    
    // store values
    main.dataset.value1 = 1;
    main.dataset.state = JSON.stringify({ a:1, b:2 });
    
    // retreive values
    console.log( main.dataset.value1 ); // "1"
    console.log( JSON.parse(main.dataset.state).a ); // 1

    Advantages:

    • you can define values in JavaScript or HTML — such as <main data-value1="1">
    • useful for storing the state of a specific component
    • the DOM is fast! (contrary to popular opinion)

    Disadvantages:

    • fragile: refreshing or closing the tab wipes everything (unless a value is server-rendered into the HTML)
    • strings only: requires serialization and de-serialization
    • a larger DOM affects performance
    • third-party scripts can examine or overwrite values

    DOM node storage is slower than variables. Use it sparingly in situations where it’s practical to store a component’s state in HTML.

    3. Web Storage (localStorage and sessionStorage)

    metric comment
    capacity 5MB per domain
    read/write speed synchronous operation: can be slow
    persistence data remains until cleared

    Web Storage provides two similar APIs to define name/value pairs. Use:

    1. window.localStorage to store persistent data, and
    2. window.sessionStorage to retain session-only data while the browser tab remains open (but see Data Persistence)

    Store or update named items with .setItem():

    localStorage.setItem('value1', 123);
    localStorage.setItem('value2', 'abc');
    localStorage.setItem('state', JSON.stringify({ a:1, b:2, c:3 }));

    Retrieve them with .getItem():

    const state = JSON.parse( localStorage.getItem('state') );

    And delete them with .removeItem():

    localStorage.removeItem('state')

    Other properties and methods include:

    Changing any value raises a storage event in other browser tabs/windows connected to the same domain. Your application can respond accordingly:

    window.addEventListener('storage', s => {
    
      console.log(`item changed: ${ s.key }`);
      console.log(`from value  : ${ s.oldValue }`);
      console.log(`to new value: ${ s.newValue }`);
    
    });

    Advantages:

    • simple name/value pair API
    • session and persistent storage options
    • good browser support

    Disadvantages:

    • strings only: requires serialization and de-serialization
    • unstructured data with no transactions, indexing, or search
    • synchronous access will affect the performance of large datasets

    Web Storage is ideal for simpler, smaller, ad-hoc values. It’s less practical for storing large volumes of structured information, but you may be able to avoid performance issues by writing data when the page unloads.

    4. IndexedDB

    metric comment
    capacity depends on device. At least 1GB, but can be up to 60% of remaining disk space
    read/write speed fast
    persistence data remains until cleared

    IndexedDB offers a NoSQL-like low-level API for storing large volumes of data. The store can be indexed, updated using transactions, and searched using asynchronous methods.

    The IndexedDB API is complex and requires some event juggling. The following function opens a database connection when passed a name, version number, and optional upgrade function (called when the version number changes):

    // connect
    function dbConnect(dbName, version, upgrade) {
    
      return new Promise((resolve, reject) => {
    
        const request = indexedDB.open(dbName, version);
    
        request.onsuccess = e => {
          resolve(e.target.result);
        };
    
        request.onerror = e => {
          console.error(`indexedDB error: ${ e.target.errorCode }`);
        };
    
        request.onupgradeneeded = upgrade;
    
      });
    
    }

    The following code connects to a myDB database and initializes a todo object store (analogous to a SQL table or MongoDB collection). It then defines an auto-incrementing key named id:

    (async () => {
    
      const db = await dbConnect('myDB', 1.0, e => {
    
        db = e.target.result;
        const store = db.createObjectStore('todo', { keyPath: 'id', autoIncrement: true });
    
      })
    
    })();

    Once the db connection is ready, you can .add new data items in a transaction:

    db.transaction(['todo'], 'readwrite')
      .objectStore('todo')
      .add({ task: 'do something' })
      .onsuccess = () => console.log( 'added' );

    And you can retrieve values, such as the first item:

    db.transaction(['todo'], 'readonly')
      .objectStore('todo')
      .get(1)
      .onsuccess = data => console.log( data.target.result );
      // { id: 1, task: 'do something' }

    Advantages:

    • flexible data store with the largest space
    • robust transactions, indexing, and search options
    • good browser support

    Disadvantages:

    • a complex callback and event-based API

    IndexedDB is the best option for reliably storing large quantities of data, but you’ll want to reach for a wrapper library such as idb, Dexie.js, or JsStore.

    5. Cache API

    metric comment
    capacity depends on device, but Safari limits each domain to 50MB
    read/write speed fast
    persistence data remains until cleared or after two weeks in Safari

    The Cache API provides storage for HTTP request and response object pairs. You can create any number of named caches for storing any number of network data items.

    The API is typically used in service workers to cache network responses for progressive web apps. When a device disconnects from the network, cached assets can be re-served so a web app can function offline.

    The following code stores a network response in a cache named myCache:

    // cache name
    const cacheName = 'myCache';
    
    (async () => {
    
      // cache network response
      const stored = await cacheStore('/service.json') );
      console.log(stored ? 'stored OK' : 'store failed');
    
    })();
    
    // store request
    async function cacheStore( url ) {
    
      try {
    
        // open cache
        const cache = await caches.open( cacheName );
    
        // fetch and store response
        await cache.add( url );
        return true;
    
      }
      catch(err) {
        return undefined; // store failed
      }
    
    }

    A similar function can retrieve an item from the cache. In this example, it returns the response body text:

    (async () => {
    
      // fetch text from cached response
      const text = await cacheGet('/service.json') );
      console.log( text );
    
    })();
    
    async function cacheGet( url ) {
    
      try {
    
        const
    
          // open cache
          cache = await caches.open( cacheName ),
    
          // fetch stored response
          resp = await cache.match(url);
    
        // return body text
        return await resp.text();
    
      }
      catch(err) {
        return undefined; // cache get failed
      }
    
    }

    Advantages:

    • stores any network response
    • can improve web application performance
    • allows a web application to function offline
    • a modern Promise-based API

    Disadvantages:

    • not practical for storing application state
    • possibly less useful outside progressive web apps
    • Apple has not been kind to PWAs and the Cache API

    The Cache API is the best option for storing files and data retrieved from the network. You could probably use it to store application state, but it’s not designed for that purpose and there are better options.

    5.5 AppCache

    AppCache was the defunct predecessor to the Cache API. This isn’t the storage solution you’re looking for. Nothing to see here. Please move along.

    6. File System Access API

    metric comment
    capacity depends on remaining disk space
    read/write speed depends on file system
    persistence data remains until cleared

    The File System Access API allows a browser to read, write, modify, and delete files from your local file system. Browsers run in a sandboxed environment so the user must grant permission to a specific file or directory. This returns a FileSystemHandle so a web application can read or write data like a desktop app.

    The following function saves a Blob to a local file:

    async function save( blob ) {
    
      // create handle to a local file chosen by the user
      const handle = await window.showSaveFilePicker();
    
      // create writable stream
      const stream = await handle.createWritable();
    
      // write the data
      await stream.write(blob);
    
      // save and close the file
      await stream.close();
    }

    Advantages:

    • web apps can securely read and write to the local file system
    • less need to upload files or process data on a server
    • a great feature for progressive web apps

    Disadvantages:

    • minimal browser support (Chrome only)
    • the API may change

    This storage option excites me the most, but you’ll need to wait a couple of years before it becomes viable for production use.

    7. File and Directory Entries API

    metric comment
    capacity depends on remaining disk space
    read/write speed unknown
    persistence data remains until cleared

    The File and Directory Entries API provides a sandboxed file system available to a domain which can create, write, read, and delete directories and files.

    Advantages:

    • could have some interesting uses

    Disadvantages:

    • non-standard, incompatibilities between implementations, and behavior may change.

    MDN explicitly states: do not use this on production sites. Widespread support is several years away at best.

    8. Cookies

    metric comment
    capacity 80Kb per domain (20 cookies with up to 4Kb in each)
    read/write speed fast
    persistence good: data remains until it’s wiped or expires

    Cookies are domain-specific data. They have a reputation for tracking people, but they’re essential for any system which needs to maintain server state — such as logging on. Unlike other storage mechanisms, cookies are (usually) passed between the browser and server on every HTTP request and response. Both devices can examine, modify, and delete cookie data.

    document.cookie sets cookie values in client-side JavaScript. You must define a string with a name and value separated by an equals symbol (=). For Example:

    document.cookie = 'cookie1=123';
    document.cookie = 'anothercookie=abc';

    Values must not contain commas, semicolons, or whitespace, so encodeURIComponent() may be necessary:

    document.cookie = `hello=${ encodeURIComponent('Hello, everyone!') }`;

    Further cookie settings can be appended with semi-colon separators, including:

    • ;domain=: if not set, the cookie is only available on the current URL domain. Using ;path=mysite.com would permit it on any subdomain of mysite.com.
    • ;path=: if not set, the cookie is only available in the current path and child paths. Set ;path=/ to allow any path in the domain.
    • ;max-age=: the cookie expiry time in seconds — such as ;max-age=60.
    • ;expires=: a cookie expiry date — such as ;expires=Thu, 04 July 2021 10:34:38 UTC (use date.toUTCString() to format appropriately).
    • ;secure: the cookie will only be transmitted over HTTPS.
    • ;HTTPOnly: makes cookies inaccessible to client-side JavaScript.
    • ;samesite=: controls whether another domain can access a cookie. Set it to lax (the default, shares the cookie to the current domain), strict (stops the initial cookie being sent when following a link from another domain), or none (no restrictions).

    Example: set a state cookie which expires in 10 minutes and is available on any path in the current domain:

    const state = { a:1, b:2, c:3 };
    
    document.cookie = `state=${ encodeURIComponent(JSON.stringify(state)) }; path=/; max=age=600`;

    document.cookie returns a string containing every name and value pair separated by a semi-colon. For example:

    console.log( document.cookie );
    // "cookie1=123; anothercookie=abc; hello=Hello%2C%20everyone!; state=%7B%22a%22%3A1%2C%22b%22%3A2%2C%22c%22%3A3%7D"

    The function below parses the string and converts it to an object containing name-value pairs. For example:

    const
      cookie = cookieParser();
      state = cookie.state;
    
    console.log( state );
    // { a:1, b:2, c:3 }
    
    
    // parse cookie values
    function cookieParser() {
    
      const nameValue = {};
    
      document.cookie
        .split('; ')
        .map(nv => {
    
          nv = nv.split('=');
          if (nv[0]) {
    
            let v = decodeURIComponent( nv[1] || '' );
    
            try { v = JSON.parse(v); }
            catch(e){}
    
            nameValue[ nv[0] ] = v;
    
          }
    
        })
    
      return nameValue;
    
    }

    Advantages:

    • a reliable way to retain state between the client and server
    • limited to a domain and, optionally, a path
    • automatic expiry control with max-age (seconds) or Expires (date)
    • used in the current session by default (set an expiry date to persist the data beyond page refreshes and tab closing)

    Disadvantages:

    • cookies are often blocked by browsers and plugins (they’re generally converted to session cookies so sites continue to work)
    • clunky JavaScript implementation (it’s best to create your own cookie handler or opt for a library such as js-cookie)
    • strings only (requires serialization and de-serialization)
    • limited storage space
    • cookies can be examined by third-party scripts unless you restrict access
    • blamed for privacy invasion (regional legislation may require you to show a warning for non-essential cookies)
    • cookie data is appended to every HTTP request and response which can affect performance (storing 50Kb of cookie data, then requesting ten 1 byte files, would incur one megabyte of bandwidth)

    Avoid cookies unless there’s no viable alternative.

    9. window.name

    metric comment
    capacity varies, but several megabytes should be possible
    read/write speed fast
    persistence session data remains until the tab is closed

    The window.name property sets and gets the name of the window’s browsing context. You can set a single string value which persists between browser refreshes or linking elsewhere and clicking back. For example:

    let state = { a:1, b:2, c:3 };
    window.name = JSON.stringify( state );

    Examine the value using:

    state = JSON.parse( window.name );
    console.log( state.b );
    // 2

    Advantages:

    • easy to use
    • can be used for session-only data

    The disadvantages:

    • strings only: requires serialization and de-serialization
    • pages in other domains can read, modify, or delete the data (never use it for sensitive information)

    window.name was never designed for data storage. It’s a hack and there are better options.

    10. WebSQL

    metric comment
    capacity 5MB per domain
    read/write speed sluggish
    persistence data remains until cleared

    WebSQL was an effort to bring SQL-like database storage to the browser. Example code:

    // create DB (name, version, description, size in bytes)
    const db = openDatabase('todo', '1.0', 'my to-do list', 1024 * 1024);
    
    // create table and insert first item
    db.transaction( t => {
    
      t.executeSql('CREATE TABLE task (id unique, name)');
      t.executeSql('INSERT INTO task (id,name) VALUES (1, "wash cat")');
    
    });
    
    // output array of all items
    db.transaction( t => {
    
      t.executeSql(
        "SELECT * FROM task",
        [],
        (t, results) => { console.log(results.rows); }
      );
    
    });

    Chrome and some editions of Safari support the technology, but it was opposed by Mozilla and Microsoft in favor of IndexedDB.

    Advantages:

    • designed for robust client-side data storage and access
    • familiar SQL syntax often used by server-side developers

    Disadvantages:

    • limited and buggy browser support
    • inconsistent SQL syntax across browsers
    • asynchronous but clunky callback-based API
    • poor performance

    Do not use WebSQL! It hasn’t been a viable option since the specification was deprecated in 2010.

    Scrutinizing Storage

    The Storage API can examine space available to Web Storage, IndexedDB, and the Cache API. All browsers except Safari and IE support the Promise-based API which offers an .estimate() method to calculate the quota (space available to the domain) and usage (space already used). For example:

    (async () => {
    
      if (!navigator.storage) return;
    
      const storage = await navigator.storage.estimate();
    
      console.log(`bytes allocated  : ${ storage.quota }`);
      console.log(`bytes in use     : ${ storage.usage }`);
    
      const pcUsed = Math.round((storage.usage / storage.quota) * 100);
      console.log(`storage used     : ${ pcUsed }%`);
    
      const mbRemain = Math.floor((storage.quota - storage.usage) / 1024 / 1024);
      console.log(`storage remaining: ${ mbRemain } MB`);
    
    })();

    Two further asynchronous methods are available:

    • .persist(): returns true if the site has permission to store persistent data, and
    • .persisted(): returns true if the site has already stored persistent data

    The Application panel in browser developer tools (named Storage in Firefox) allows you to view, modify, and clear localStorage, sessionStorage, IndexedDB, WebSQL, cookies, and cache storage.

    You can also examine cookie data sent in the HTTP request and response headers by clicking any item in the developer tools’ Network panel.

    Storage Smorgasbord

    None of these storage solutions is perfect, and you’ll need to adopt several in a complex web application. That means learning more APIs. But having a choice in each situation is a good thing — assuming you can choose the appropriate option, of course!