Redis vs. MongoDB: Comparing In-Memory Databases with Percona Memory Engine

In this post, we’re comparing two of the most popular NoSQL databases: Redis (in-memory) and MongoDB (Percona memory storage engine).

Redis is a popular and very fast in-memory database structure store primarily used as a cache or a message broker. Being in-memory, it’s the data store of choice when response times trump everything else.

MongoDB is an on-disk document store that provides a JSON interface to data and has a very rich query language. Known for its speed, efficiency, and scalability, it’s currently the most popular NoSQL database used today. However, being an on-disk database, it can’t compare favorably to an in-memory database like Redis in terms of absolute performance. But, with the availability of the in-memory storage engines for MongoDB, a more direct comparison becomes feasible.

Percona Memory Engine for MongoDB

Starting in version 3.0, MongoDB provides an API to plug in the storage engine of your choice. A storage engine, from the MongoDB context, is the component of the database that’s responsible for managing how the data is stored, both in-memory and on-disk. MongoDB supports an in-memory storage engine, however, it’s currently limited to the Enterprise edition of the product. In 2016, Percona released an open source in-memory engine for the MongoDB Community Edition called the Percona Memory Engine for MongoDB. Like MonogDB’s in-memory engine, it’s also a variation of the WiredTiger storage engine, but with no persistence to disk.

With an in-memory MongoDB storage engine in place, we have a level playing field between Redis and MongoDB. So, why do we need to compare the two? Let’s look at the advantages of each of them as a caching solution.

Let’s look at Redis first.

Advantages of Redis as a Cache

  • A well-known caching solution that excels at it.
  • Redis isn’t a plain cache solution – it has an advanced data structures that provide many powerful ways to save and query data that can’t be achieved with a vanilla key-value cache.
  • Redis is fairly simple to setup, use and learn.
  • Redis provides persistence that you can opt to set up, so cache warming in the event of a crash is hassle free.

Disadvantages of Redis:

  • It doesn’t have inbuilt encryption on the wire.
  • No role-based account control (RBAC).
  • There isn’t a seamless, mature clustering solution.
  • Can be a pain to deploy in large-scale cloud deployments.

Advantages of MongoDB as a Cache

  •  MongoDB is a more traditional database with advanced data manipulation features (think aggregations and map-reduce) and a rich query language.
  • SSL, RBAC, and scale-out built in.
  • If you are already using MongoDB as your primary database, then your operational and development costs drop as you only have one database to learn and manage.

Look at this post from Peter Zaitsev on where the MongoDB in-memory engine might be a good fit.

Disadvantage of MongoDB:

  • With an in-memory engine, it offers no persistence until it’s deployed as a replica set with persistence configured on the read replica(s).

In this post, we’ll focus on quantifying the performance differences between Redis and MongoDB. A qualitative comparison and operational differences will be covered in subsequent posts.

Redis vs. In-Memory MongoDB


  • Redis performs considerably better for reads for all sorts of workloads, and better for writes as the workloads increase.
  • Even though MongoDB utilizes all the cores of the system, it gets CPU bound comparatively early.  While it still had compute available, it was better at writes than Redis.
  • Both databases are eventually compute-bound. Even though Redis is single-threaded, it (mostly) gets more done with running on one core than MongoDB does while saturating all the cores.
  • Redis, for non-trivial data sets, uses a lot more RAM compared to MongoDB to store the same amount of data.


We used YCSB to measure the performance, and have been using it to compare and benchmark the performance of MongoDB on various cloud providers and configurations in the past. We assume a basic understanding of YCSB workloads and features in the test rig description.

  • Database instance type:  AWS EC2 c4.xlarge featuring 4 cores, 7.5 GB memory, and enhanced networking to ensure we don’t have any network bottlenecks.
  • Client Machine:  AWS EC2 c4.xlarge in the same virtual private cloud (VPC) as the database servers.
  • Redis:  Version 3.2.8 with AOF and RDB turned off. Standalone.
  • MongoDB:  Percona Memory Engine based on MongoDB version 3.2.12. Standalone.
  • Network Throughput:  Measured via iperf as recommended by AWS:
    Test Complete. Summary Results:
    [ ID] Interval           Transfer     Bandwidth       Retr
    [  4]   0.00-60.00  sec  8.99 GBytes  1.29 Gbits/sec  146             sender
    [  4]   0.00-60.00  sec  8.99 GBytes  1.29 Gbits/sec                  receiver
  • Workload Details
    1. Insert Workload: 100 % Write – 2.5 million records
    2. Workload A: Update heavy workload – 50%/50% Reads/Writes – 25 million operations
    3. Workload B: Read mostly workload – 95%/5% Reads/Writes – 25 million operations
  • Client Load: Throughput and latency measured over incrementally increasing loads generated from the client. This was done by increasing the number of YCSB client load threads, starting at 8 and growing in multiples of 2


Workload B Performance

Since the primary use case for in-memory databases is cache, let’s look at Workload B first.

Here are the throughput/latency numbers from the 25 million operations workload and the ratio of reads/writes was 95%/5%. This would be a representative cache reading workload:


Note: Throughput is plotted against the primary axis (left), while latency is plotted against the secondary axis (right).

Observations during the Workload B run:

  • For MongoDB, CPU was saturated by 32 threads onwards. Greater than 300% usage with single-digit idle percentages.
  • For Redis, CPU utilization never crossed 95%. So, Redis was consistently performing better than MongoDB while running on a single thread, while MongoDB was saturating all the cores of the machine.
  • For Redis, at 128 threads, runs failed often with read-timeout exceptions.

Workload A Performance

Here are the throughput/latency numbers from the 25 million operations workload. The ratio of reads/writes was 50%/50%:


Note: Throughput is plotted against the primary axis (left), while latency is plotted against the secondary axis (right).

Observations during the Workload A run:

  • For MongoDB, CPU was saturated by 32 threads onwards. Greater than 300% usage with single-digit idle percentages.
  • For Redis, CPU utilization never crossed 95%.
  • For Redis, by 64 threads and above, runs failed often with read-timeout exceptions.

Insert Workload Performance

Finally, here are the throughput/latency numbers from the 2.5 million record insertion workload. The number of records was selected to ensure the total memory was used in the event Redis that did not exceed 80% (since Redis is the memory hog, see Appendix B).


Note: Throughput is plotted against the primary axis (left), while latency is plotted against the secondary axis (right).

Observations during the Insert Workload run:

  • For MongoDB, CPU was saturated by 32 threads onwards. Greater than 300% usage with single-digit idle percentages.
  • For Redis, CPU utilization never crossed 95%.


A: Single-Thread Performance

I had a strong urge to find this out – even though it’s not very useful in real-world conditions: who would be better when applying the same load to each of them from a single thread. That is, how would a single-threaded application perform?

B: Database Size

The default format of records inserted by YCSB are: each record is of 10 fields and each field is 100 bytes. Assuming each record to be around 1KB, the total expected size in memory would be upwards of 2.4GB. There was a stark contrast in the actual sizes as seen in the databases.


> db.usertable.count()
> db.usertable.findOne()
    "_id" : "user6284781860667377211",
    "field1" : BinData(0,"OUlxLllnPC0sJEovLTpyL18jNjk6ME8vKzF4Kzt2OUEzMSEwMkBvPytyODZ4Plk7KzRmK0FzOiYoNFU1O185KFB/IVF7LykmPkE9NF1pLDFoNih0KiIwJU89K0ElMSAgKCF+Lg=="),
    "field0" : BinData(0,"ODlwIzg0Ll5vK1s7NUV1O0htOVRnMk53JEd3KiwuOFN7Mj5oJ1FpM11nJ1hjK0BvOExhK1Y/LjEiJDByM14zPDtgPlcpKVYzI1kvKEc5PyY6OFs9PUMpLEltNEI/OUgzIFcnNQ=="),
    "field7" : BinData(0,"N155M1ZxPSh4O1B7IFUzJFNzNEB1OiAsM0J/NiMoIj9sP1Y1Kz9mKkJ/OiQsMSk2OCouKU1jOltrMj4iKEUzNCVqIV4lJC0qIFY3MUo9MFQrLUJrITdqOjJ6NVs9LVcjNExxIg=="),
    "field6" : BinData(0,"Njw6JVQnMyVmOiZyPFxrPz08IU1vO1JpIyZ0I1txPC9uN155Ij5iPi5oJSIsKVFhP0JxM1svMkphL0VlNzdsOlQxKUQnJF4xPkk9PUczNiF8MzdkNy9sLjg6NCNwIy1sKTw6MA=="),
    "field9" : BinData(0,"KDRqP1o3KzwgNUlzPjwgJEgtJC44PUUlPkknKU5pLzkuLEAtIlg9JFwpKzBqIzo2MCIoKTxgNU9tIz84OFB/MzJ4PjwoPCYyNj9mOjY+KU09JUk1I0l9O0s/IEUhNU05NShiNg=="),
    "field8" : BinData(0,"NDFiOj9mJyY6KTskO0A/OVg/NkchKEFtJUprIlJrPjYsKT98JyI8KFwzOEE7ICR4LUF9JkU1KyRkKikoK0g3MEMxKChsL10pKkAvPFRxLkxhOlotJFZlM0N/LiR4PjlqJ0FtOw=="),
    "field3" : BinData(0,"OSYoJTR+JEp9K00pKj0iITVuIzVqPkBpJFN9Myk4PDhqOjVuP1YhPSM2MFp/Kz14PTF4Mlk3PkhzKlx3L0xtKjkqPCY4JF0vIic6LEx7PVBzI0U9KEM1KDV4NiEuKFx5MiZyPw=="),
    "field2" : BinData(0,"Njd8LywkPlg9IFl7KlE5LV83ISskPVQpNDYgMEprOkprMy06LlotMUF5LDZ0IldzLl0tJVkjMTdgJkNxITFsNismLDxuIyYoNDgsLTc+OVpzKkBlMDtoLyBgLlctLCxsKzl+Mw=="),
    "field5" : BinData(0,"OCJiNlI1O0djK1BtIyc4LEQzNj9wPyQiPT8iNE1pODI2LShqNDg4JF1jNiZiNjZuNE5lNzA8OCAgMDp2OVkjNVU3MzIuJTgkNDp0IyVkJyk6IEEvKzVyK1s9PEAhKUJvPDxyOw=="),
    "field4" : BinData(0,"OFN1I0B7N1knNSR2LFp7PjUyPlJjP15jIUdlN0AhNEkhMC9+Lkd5P10jO1B3K10/I0orIUI1NzYuME81I0Y1NSYkMCxyI0w/LTc8PCEgJUZvMiQiIkIhPCF4LyN6K14rIUJlJg==")
> db.runCommand({ dbStats: 1, scale: 1 })
    "db" : "ycsb",
    "collections" : 1,
    "objects" : 2500000,
    "avgObjSize" : 1167.8795252,
    "dataSize" : 2919698813,
    "storageSize" : 2919698813,
    "numExtents" : 0,
    "indexes" : 1,
    "indexSize" : 76717901,
    "ok" : 1

So, the space taken is ~2.7GB which is pretty close to what we expected.


Let’s look at Redis now.

> info keyspace
# Keyspace
db0:keys=2500001,expires=0,avg_ttl=0> RANDOMKEY
"user3176318471616059981"> hgetall user3176318471616059981
 1) "field1"
 2) "#K/<No\"&l*M{,;f;]\x7f)Ss'+2<D}7^a8I/01&9.:)Q71T7,3r&\\y6:< Gk;6n*]-)*f>:p:O=?<:(;v/)0)Yw.W!8]+4B=8.z+*4!"
 3) "field2"
 4) "(9<9P5**d7<v((2-6*3Zg/.p4G=4Us;N+!C! I50>h=>p\"X9:Qo#C9:;z.Xs=Wy*H3/Fe&0`8)t.Ku0Q3)E#;Sy*C)[email protected]"
 5) "field5"
 6) "#1 %8x='l?5d38~&U!+/b./b;(6-:v!5h.Ou2R}./(*)4!8>\"B'!I)5U?0\" >Ro.Ru=849Im+Qm/Ai(;:$Z',]q:($%&(=3~5(~?"
 7) "field0"
 8) "+\"(1Pw.>*=807Jc?Y-5Nq#Aw=%*57r7!*=Tm!<j6%t3-45L5%Cs#/h;Mg:Vo690-/>-X}/X#.U) )f9-~;?p4;p*$< D-1_s!0p>"
 9) "field7"
10) ":]o/2p/3&(!b> |#:0>#0-9b>Pe6[}<Z{:S}9Uc*0<)?60]37'~'Jk-Li',x!;.5H'\"'|.!v4Y-!Hk=E\x7f2;8*9((-09*b#)x!Pg2"
11) "field3"
12) " C; ,f6Uq+^i Fi'8&0By\"^##Qg\":$+7$%Y;7Rs'\"d3Km'Es>.|33$ Vo*M%=\"<$&j%/<5]%\".h&Kc'5.46x5D35'0-3l:\"| !l;"
13) "field6"
14) "-5x6!22)j;O=?1&!:&.S=$;|//r'?d!W54(j!$:-H5.*n&Zc!0f;Vu2Cc?E{1)r?M'!Kg'-b<Dc*1d2M-9*d&(l?Uk5=8,>0.B#1"
15) "field9"
16) "(Xa&1t&Xq\"$((Ra/Q9&\": &>4Ua;Q=!T;(Vi2G+)Uu.+|:Ne;Ry3U\x7f!B\x7f>O7!Dc;V7?Eu7E9\"&<-Vi>7\"$Q%%A%1<2/V11: :^c+"
17) "field8"
18) "78(8L9.H#5N+.E5=2`<Wk+Pw?+j'Q=3\"$,Nk3O{+3p4K?0/ 5/r:W)5X}#;[email protected]\x7f\"+&#Ju+Z97#t:J9$'*(K).7&0/` 125O38O)0"
19) "field4"
20) "$F=)Ke5V15_)-'>=C-/Ka7<$;6r#_u F9)G/?;t& x?D%=Ba Zk+]) ($=I%3P3$<`>?*=*r9M1-Ye:S%%0,(Ns3,0'A\x7f&Y12A/5"> info memory
# Memory

At peak usage, Redis seems to be taking around 5.72G of memory i.e. twice as much memory as MongoDB takes. Now, this comparison may not be perfect because of the differences in the two databases, but this difference in memory usage is too large to ignore. YCSB inserts record in a hash in Redis, and an index is maintained in a sorted set. Since an individual entry is larger than 64, the hash is encoded normally and there is no savings in space. Redis performance comes at the prices of increased memory footprint.

This, in our opinion, can be an important data point in choosing between MongoDB and Redis – MongoDB might be preferable for users who care about reducing their memory costs.

C: Network Throughput

An in-memory database server is liable to be either compute-bound or network I/O-bound, so it was important throughout the entire set of these tests to ensure that we were never getting network-bound. Measuring network throughput while running application throughput tests adversely affects the overall throughput measurement. So, we ran subsequent network throughput measurements using iftop at the thread counts where the highest write throughputs were observed. This was found to be around 440 Mbps for both Redis and MongoDB at their respective peak throughput. Given our initial measurement of the maximum network bandwidth to be around 1.29 Gbps, we are certain that we never hit the network bounds. In fact, it only supports the inference that if Redis were multi-core, we might get much better numbers.